WO2009081168A1 - Remote classification of substances - Google Patents

Abstract

The present invention provides a method and apparatus for the remote classification of substances; the said method comprising the steps of : emitting at least one pulse of incident radiation from a source of electromagnetic radiation towards a target or target area; monitoring the said target area for fluorescent radiation emitted from the from the target or target area in response to said at least one pulse of incident radiation; recording the emission response of the target or target area in response to said at least one pulse of incident radiation; comparing the emission response of the target or target area with reference emission response data for known substances/materials; identifying a target substance or substance in said target area in accordance with said comparison wherein the emission response of the target area or target area is recorded for at least two filtered wavebands and step (d) is repeated for each of the said wavebands. Preferably the method comprises the step of normalising the recorded values for each waveband with respect to the total number of photons detected for all said wavebands to provide non-dimensional data for comparison with corresponding reference data.

Description

REMOTE CLASSIFICATION OF SUBSTANCES

The present invention relates to a method and apparatus for the remote classification of substances. In particular, although not exclusively, the invention relates to a method and apparatus for the remote classification of controlled substances, for example, explosives and/or narcotics.

In certain situations it is useful to be able to remotely detect and classify substances. For instance, at airports where large numbers of people are present and items of luggage are handled it is useful to be able to determine the presence of explosives, or other substances such as narcotics. The presence may be indicated by such things as smears on the outside of luggage, clothes and other objects. Further, it is useful if the classification of the substances could be made without the luggage or object having to be handled manually since such handling is laborious.

US-A 1-2004/0051867 discusses a system for remotely detecting and identifying controlled substances on an object. This system uses a laser for illuminating at least part of the object and particular detectors for the detection of luminescence arising from the incident laser energy. The. method for detecting and identifying the substances involves second harmonics and Raman scattering analysis.

The present invention provides an alternative method and apparatus for the remote classification of substances.

The apparatus and method are able to determine the presence and classify a substance in circumstances where the quantity of substance is extremely small and may be regarded as a trace amount. According to an aspect of the present invention there is provided a method for ' the remote classification of substances; the said method comprising the steps of: a) emitting incident radiation from a source of electromagnetic radiation towards a target or target area; b) monitoring the said target area for fluorescent radiation emitted from the target or target area in response to said incident radiation; c) recording the emission response of the target or target area in response to said incident radiation; d) comparing the emission response of the target or target area with reference emission response data for known substances/materials; e) identifying a target substance or substance in said target in accordance with said comparison wherein the emission response of the target area or target area is recorded for at least two^ filtered wavebands and step d) is repeated for each of the said wavebands.

In preferred embodiments, step c) comprises the step of recording substantially the whole fluorescence emission response of the target or target area for each respective waveband.

Preferably the recorded fluorescence emission response of the target or target area for each respective waveband coσesponds to the number of detected photons recorded in the respective waveband during the decay period or period of continuous illumination of said target or target area by said incident radiation, and wherein the method further comprises the step of normalising the recorded values for each waveband with respect to the total number of photons detected for all said wavebands during said period to provide non-dimensional data for comparison with corresponding reference data in step e).

Step e) may comprise the step of determining a cross-distance measurement for each of the said normalised recorded values for the said respective wavebands with corresponding reference emission data for said known substances, and identifying which of the said known substances is present by identifying the best match with said reference data.

The said cross-distance measurement comparison may comprise a least squares method, utilising the root mean squared value of the differences in the respective measured and reference emission response values.

The said measured emission response values for the respective wavebands may be selectively weighted prior to said comparison with said reference values.

The fluorescence emission responses for the respective filtered wavebands may be recorded either simultaneously by separate recoding means or sequentially by means of selective filters sequentially applied to the recoding means.

The said filtered wavebands may comprise at least UV filtered wavebands. Preferably the said filtered wavebands are selected from the range of 300 to 450 nm. The said wavebands are substantially equal, for example between 20nm and 30 nm.

At least one of the said wavebands my be in the visible spectrum for the simultaneous visual display of the target or target area with the detected substance so that the spatial position of an identified substance or material on the target or in the target area can be superimposed and shown on a visual display in use. The incident radiation may comprise an emission frequency capable of generating a fluorescence emission response in target substances to be identified, and wherein the incident radiation is preferably laser generated UV radiation, preferably laser generated and more preferably having a wavelength in the region of 266nm.

In preferred embodiments the method further comprises the step of measuring the distance of the target or target area from the electromagnetic radiation source emitter and/or the distance of the target area from the fluorescence emission recording means.

The method may include the step of determining the temporal location and/or magnitude of the peak intensity of the resultant fluorescent radiation for each combination of incident and resultant wavelength.

An intensity value for background fluorescent radiation may be determined.

The determined intensity values may be compared to the background fluorescent radiation, the results of said comparison aiding the classification of substances in the ^• target area. The substances are classified into narcotics or explosives. In preferred embodiments the substances are classified into types of explosives such as CA, PE4, P¹TN, RDX, SmA, SmH for example.

The step of recording the resultant fluorescent radiation is effected by means of a digital camera or the like, whereby fluorescence decay data is recorded on a pixel by pixel basis by said digital camera or like device. The aforementioned method may further comprise the steps of capturing and displaying a visual image of the target or target area and indicating on the displayed image the location of any classified substances.

The step of emitting incidental radiation may comprise the step of emitting at least one pulse of electromagnetic radiation or continuously illuminating the target or target area with electromagnetic radiation.

The step of continuously illuminating the said target or target area may comprise the step of emitting electromagnetic radiation from a narrow band light source or LED.

The step of recording the emission response of the target or target area may comprise the step of recording substantially the whole fluorescence emission decay response of the target or the target area.

In one aspect, the invention provides a method for the remote classification of substances in a target area comprising the following steps: a) providing a source of electromagnetic radiation and emitting at least one incident pulse of electromagnetic radiation, having a first incident wavelength, in the direction of the target area; b) providing a filter and capture means and filtering the resultant fluorescent radiation from the target area to exclude all radiation except that having at least one resultant wavelength, and capturing said filtered radiation at selected points in time relative to said incident pulse; c) analysing the fluorescent radiation captured at each said selected point in time to determine its intensity; d) repeating steps a) to c) with the repeated incident pulse having a different incident wavelength and the resultant fluorescent radiation being filtered to exclude all radiation except that having either the same resultant wavelength as in step b), or a different resultant wavelength; and/or repeating simultaneously steps b) and c) the resultant fluorescence being filtered to exclude all radiation except that having a resultant wavelength different from that in step b); e) determining differences between aspects of the captured radiation for two or more combinations of incident and resultant wavelengths; f) ' comparing the determined differences with differences between aspects of known fluorescent radiation for the same combinations of incident and resultant wavelengths; and g) analysing the results of the comparison to classify substances within the target area.

When the resultant fluorescent radiation is discussed as being filtered to exclude all radiation except that having at least one resultant wavelength it should be understood that this also includes the possibility that the filter excludes all radiation except that having at least one range of resultant wavelengths. For example, the range may extend between 1 nm and several hundred nm either side of the centre wavelength.

The substances may be classified broadly in that they may be identified, for example, as narcotics or explosives. Alternatively, or additionally, the substances may be classified more specifically in that they may be identified, for example, as plastic based explosives or gun-powder based explosives. It also possible that the substances may be classified even more specifically in that they may be identified, for example, as a particular plastic based explosives, such as C4, or gun-powder based explosives, such as safety fuse.

The method of the invention is predicated on the knowledge that the intensity of fluorescent radiation, resultant from an incident pulse of suitable electromagnetic radiation, will reduce over time. The total lifetime of fluorescence is typically less than 1 μs. This reduction of intensity will follow an exponential decay curve which may be represented by the following equation : l = x.e-^{γ t} [l]

where I is the intensity, X is an intensity parameter, e is the base of natural logarithms, γ is the decay time parameter, and t is time. An example of the intensity of resultant fluorescent radiation measured over time following an incident pulse of electromagnetic radiation is shown in Figure 2.

The characteristics of the decay curve for each substance that fluoresces when stimulated by an incident pulse will be substantially unique. This will be based upon the type and inherent properties of the substance. These inherent properties are fixed in that the present method does not alter them. It will also be based upon factors such as the strength or intensity of the incident pulse, its wavelength, the distance from the source of the pulse to the substance, the wavelength of the filtered resultant fluorescent radiation of interest, the quantity of substance and environmental factors such as temperature, humidity and the scattering properties of the intermediate medium through which the radiation passes. To certain extents these variable factors may all be affected by the method.

The term "intensity" may be taken to approximate, be the same as, or be proportional to the number of photons collected in a given period, or the amplitude of the resultant fluorescent radiation, and accordingly all references to "intensity" herein should be understood Io also mean number of photons or amplitude as necessary and appropriate.

Accordingly, it is an aim of the invention to determine characteristics of the decay curves. However, because each decay curve is also affected by the variable factors, as well as by the fixed factors, it is not possible to merely compare an intensity value as determined at a relative point in time with a known intensity value for the same relative point in time and identify or classify a substance on that basis. Instead, there are various other ways in which a substance may be classified based on collected intensity values.

In one embodiment, the method determines an intensity value for a relative point in time for received fluorescent radiation having a particular resultant wavelength having been caused by an incidence pulse having a particular incidence wavelength. The method may then determine an intensity value for the same relative point in time for received fluorescent radiation having a particular resultant wavelength having been caused by an incidence pulse having a different incidence wavelength, or an intensity value for the same relative point in time for received fluorescent radiation having a different resultant wavelength having been caused by an incidence pulse having the same incidence wavelength.

These two determined intensity values may be different from one another. However, they may have a relationship with one another in that they both may be determined from the same substance. The relationship may be represented by their proportionality with respect to each other.

The method contemplates, in its simplest form, that intensity values may be known for particular substances. These known intensity values are for received fluorescent radiation produced by electromagnetic pulses wherein the incident and resultant wavelengths are the same as used to produce the determined values discussed above. Furthermore, the known values are for the same relative points in time. The proportionality of these known intensity values may then be compared to the proportionality of the determined values.

If the proportionalities are the same or substantially similar then the substance in question may be classified or identified. More particularly, if the comparison is within a pre-defined range the substance may be classified. The pre-defined range may be +/- 5% of the difference between aspects of known received fluorescent radiation. Other ranges are contemplated and may be +/- 3%.

The confidence in the veracity of the comparison (s) may be improved in one or more of several ways. The simplest way to improve the confidence is to repeat the method of emitting an incidence pulse and capturing the received fluorescent radiation such that an average of the intensity values may be obtained. However, other more sophisticated methods, of which some may rely on a comparison of the differences between aspects of captured received fluorescent radiation and differences between known aspects of known substances, are also possible as described below.

In step b) of the method the selected points in time are relative to the incident pulse. However, they could also be relative to the peak intensity, and may be in the range 5ns to 150ns following the peak intensity. The temporal position of the peak intensity may be determined from the temporal position of the incident pulse if the distances of the source of electro-magnetic radiation to the target area and of the target area to the capture means are known since the radiation will travel at the speed of light and the approximate relative temporal position of the peak intensity to the incident pulse has been found to be substantially similar for many substances of interest. The method may therefore include the step of measuring the distance of the target area from the electromagnetic radiation emitter and/or the distance of the target area from the capture means. This step may not be required if the distances are fixed, for instance the target is known to be a certain distance from the emitter and receiver.

The method may also include the step of determining the temporal location and/or magnitude of the peak intensity of the resultant fluorescent radiation for each combination of incident and resultant wavelength.

By knowing the distances involved, and because it is known that intensity is indirectly proportional to the square of the distance, it is possible to apply scaling factors to determined intensity values, if necessary, for more direct comparison with known values which may have been determined in situations with differing distances.

One way of improving the confidence of the classification of substances is to increase the combination of incident and resultant wavelengths and the number of time periods over which the received fluorescent radiation is captured for each combination. In one embodiment, the method may be repeated as necessary to determine at least one intensity value at each selected point in time for each different combination of incident and resultant wavelengths from a possible at least two different incident wavelengths and at least four different resultant wavelengths. The incident wavelengths may be 266nm and 355nm and the resultant wavelengths may be 340nm, 370nm, 387nm and 406nm. As discussed above, the resultant fluorescent radiation may be filtered so that a range of wavelengths is achieved. For instance, one filter may provide a centre wavelength of 340nm with a range of +/- 13nm either side (i.e. 327nm to 353nm), another filter may provide a centre wavelength of 370nm with a range of +/- 18nm either side (i.e. 352nm to 388nm), another filter may provide a centre wavelength of 387nm with a range of +/- 6nm either side (i.e.381 nm to 393nm), and another filter may provide a centre wavelength of 406nm with a range of +/- 8nm either side (i.e. 398nm to 414nm) .

As already discussed, more than one intensity value may be determined for each selected point in time and the values processed to provide a single intensity value with a greater degree of confidence in the accuracy thereof. This processing may take several different or a combination of forms. For instance, a simple averaging may be used. Or, an algorithm to apply a best-fit curve to the values may be used. This latter example is possible because the shape of the curve (exponential decay) is known.

Other possibilities are the use of a least squares error estimation which may be applied to all values or only to particular values such as those in the upper half of the decay curve.

If a best-fit curve is applied then it may be defined by the parameters (with reference to the equation above) X and γ. The parameters may be determined by plotting the natural logarithm of the intensity vales against the time values. This will produce a substantially straight line, the gradient of which will be equivalent to γ and the point at which the line crosses the y (intensity) axis will be equivalent to X. If the line is not straight then it is because the determined values are not accurate enough. This may be addressed by further sampling (capturing of intensity values) and/or further processing of the data. The act of plotting the data is not necessarily physically undertaken since it is contemplated that in one embodiment the method is carried out by a computer. If the parameters are defined then it is these that may form the aspects of step e) in the method.

If a best-fit curve (or other type of curve, including actual) is applied to the intensity values determined at certain points in time then it is possible to derive intensity values at other points in time from the curve and for use in steps e) and f) (the difference determination step and comparison step). This means that the resultant fluorescent radiation may be captured at times relative the incident pulse other than those possibly dictated by the known aspects (e.g. intensity values) of known substances.

Another aspect which may be used to aid the classification of substances is the maximum peak intensity value. This value will vary for different combinations of incident and received fluorescent radiation and accordingly the differences therebetween can be compared with that of the difference between known maximum peak intensities for known substances. These maximum peak intensities may be normalised relative to a datum if necessary.

The temporal location and magnitude of the peak intensity of the received fluorescent radiation may be determined by continuously capturing (or sampling) the received fluorescent radiation from the target after one incident pulse of electromagnetic radiation. Each sample may be analysed to determine the intensity of the radiation. Since the intensity of fluorescent radiation is known to increase to a peak and then decay over time the peak may readily be determined. The temporal position of this peak relative to the pulse of incident radiation may also be readily determined.

Another possible method for determining the temporal location and magnitude of the peak (or maximum intensity) of the received fluorescent radiation is to emit a plurality of pulses, or possible even a continuous stream, of incident pulses of radiation, each separated by a suitable delay to allow any fluorescence to be received without interference by the incident pulses. Any received fluorescent radiation from the target may be captured at a suitable interval after each incident pulse and analysed to determine the intensity of any fluorescence. The interval may be defined by any number of equal or non-equal increments. For example, the interval may be defined by a number of increments having a duration of 750ps. The interval between each pulse and the capture of the received fluorescent radiation may be increased incrementally for each subsequent capture so that a set of results is obtained from which the maximum intensity and its temporal location relative to the incident pulse which produced it may be found.

Yet another aspect which may be used to aid the classification of substances is the lack of received fluorescent radiation above the background intensity value. This background intensity value may be determined by operating the capture means for a pre-defined period without any incident pulses having been emitted. The above equation [1 ] may be written as

I = X.Θ-^{γ t} + C [2]

where C is the background intensity value. This background parameter does not affect the determination of γ. Nor, since it is the differences between aspects of known received fluorescent radiation which are compared, will it affect the comparing step in the method. However, if the method requires absolute values of intensity for comparison purposes then C may need to be taken account of to normalise the X values.

In some embodiments it may be desirable to vary the duration over which the resultant fluorescent radiation is captured between capture events. This may be due to an insufficient number of photons being emitted, or being received, by the capture means to enable meaningful analysis. Accordingly, if the duration is increased a greater number of photons may be captured. However, this may mean that the resultant difference between intensity values may not be meaningfully compared to that of known substances. To compensate for this the intensity value(s) may be corrected or normalised. This may be by the application of a suitable scaling factor. It is contemplated that the duration for capturing received fluorescent radiation may be less than one second. The duration may be between lps and 100ns. Preferably it may be between 5ns and 35ns. However-, other durations are contemplated.

In one embodiment, the incident pulse and received fluorescent radiation is in the ultraviolet range. The method may be restricted to a particular wavelength of incident radiation. This may be effected by a particular type of laser emitting device which can produce a laser pulse at one of two wavelengths. In one embodiment, these wavelengths may be 266nm and 355nm. Alternatively, or additionally, other laser emitting devices can be included to allow for laser pulses having other wavelengths. Filtering of the resultant fluorescent radiation may take place by means of suitable filter means. In one embodiment this could be an electronically controlled cassette having several different filters. Possible filters are 340nm, 370nm, 387nm and 406nm although other wavelengths are contemplated.

The known characteristics of fluorescence associated with substances may be in the form of data which may be held in a database. The data may include a range of intensity values (in the form of "look-up tables") for various combinations of incident and received fluorescent radiation wavelengths for each type of substance. These values may be converted or convertible into logarithmic values. Alternatively, or additionally, these values may be normalised such that they can usefully be compared with the values determined by the relevant methods described herein. Additionally, or alternatively, the data may be/have been processed so that rather than a multitude of intensity values each substance is represented, for each combination of incident and resultant wavelengths, by a formula equivalent to [1] or [2] above such that the X and γ parameters are known. In this way the respective intensity value for a known substance may be calculated on-demand for any particular value of t to be used in the comparison step of the method.

The step of capturing the resultant fluorescent radiation may be effected by means of a digital camera which may capture an image of the target area over a wide range of wavelengths including infra-red, ultra-violet and visible light radiation. The camera may be an image intensifying camera. Steps c) to g) of the method may be followed for the captured resultant fluorescent radiation, on a pixel by pixel basis.

The images may be defined by an array of pixels forming a CCD (Charge Coupled Device) in the digital camera. The method may include the step of analysing the received fluorescent radiation received from the target area pixel by pixel, and of classifying any substance determined on a pixel by pixel basis.

The image capture means may capture an image of the target which is subsequently split and filtered to provide separate images of the target each having a different wavelength or range of wavelengths. This means that intensity values for different resultant wavelengths may be determined substantially simultaneously. Furthermore, an image in the visible range and an image in the ultraviolet range could be provided from a single captured event, or sample.

The visual image of the target area may be displayed on appropriate means, for instance a VDU monitor. The location of any determined substance may also be visually indicated on this displayed image by appropriate means such as by highlighting one or more pixels. For example, if the image is of a suitcase on a conveyor belt, part of the handle may be highlighted in the visible image if that is the location of a particular classified substance.

The steps of the method may be repeated, each repetition using a wavelength different from, or the same as, the wavelength used previously. The repetition of all or part of the method steps may occur substantially simultaneously. This may be effected, for example, by using more than one pulse of electromagnetic radiation possibly directed to different target areas, more than one receiving means, and/or by processing and analysing the same or different wavelengths of received radiation. Repetition of the method may be used to average results from the same target which may preferably reduce the risk of false identifications or classifications. The repetition may aid the comparison of results for the same pixel, if the target is stationary, or for the same point on the target by means of different pixels if the target is moving.

The steps of the method do not necessarily have to be performed in the exact order listed above. For example, the steps of filtering the resultant fluorescent radiation and then capturing it before analysing it to determine its intensity may be reversed in that the resultant fluorescent radiation is first captured, then filtered and only then analysed.

In another aspect, the invention provides apparatus for the remote identification of substances in a target area comprising : a source of electromagnetic radiation for emitting at least one incident pulse of electromagnetic radiation, having one of a range of pre-defined incident wavelengths, in the direction of the target area; filtering means for filtering the resultant fluorescent radiation from the target area to exclude all radiation except that having one of a range of pre-defined resultant wavelengths; capturing means for capturing said resultant fluorescent radiation in at least two time periods each at a predetermined and different time delay after said incident pulse; analysis means for determining the intensity of the resultant fluorescent radiation captured in each said time period; processing means for determining differences between aspects of the captured resultant fluorescent radiation for two or more combinations of incident and resultant wavelengths; comparison means for comparing the determined differences with differences between aspects of known fluorescent radiation for the same combinations of incident and resultant wavelengths; and analysis means for analysing the results of the comparison to classify substances within the target area.

The apparatus may be operated in accordance with any of the method steps described herein.

The apparatus may include distance measuring means for determining the distance between the source of electromagnetic radiation and the target area, and/or the distance between the target area and the capturing means.

The aspects of known intensity fluorescent radiation values for known substances may be held in a database. The database may form part of the apparatus or may be separate therefrom but in communication with the apparatus. The capturing means may be a camera for capturing an image of the target area as defined by an array of pixels. The apparatus may be capable of analysing the resultant fluorescent radiation received from the target area pixel by pixel and may be capable of classifying a substance on a pixel by pixel basis. The apparatus may include an image display means for displaying an image of the target area and for indicating on the image the location of a classified substance. Furthermore, the apparatus may include means for communicating to an interested party the classification of a substance in the target area.

The analysis means, processing means, comparison means and analysis means may, in one embodiment, all be comprised in a CPU, such as a computer.

The source of electromagnetic radiation may be an ultraviolet laser. The wavelength of the ultraviolet radiation may lie in the range 150nm to 400nm. More particularly, the wavelength may be switched between 266 and 355nm. The laser may be controlled by control means such that it provides an illuminating pulse having a duration short enough so that it is possible to detect the subsequent fluorescent emission from the target area without the illuminating pulse still being present and thus causing interference. The laser may be controlled such that the pulse has a duration lying in the range of between approximately 1 ps and 4ns. More preferably the duration of the illuminating ultraviolet radiation pulse may be approximately 1 ns.

The illuminating laser pulse may be controlled so that it provides a substantially dispersed pulse of radiation. The dispersal of the radiation may be proportional to the distance from the source to the target. Accordingly, if the distance is greater than a certain threshold then the dispersal may be such that the intensity of the radiation is too weak to promote fluorescence. Preferably, the distance of the laser to the target is less than 10 metres. More preferably, the distance is approximately 1.5 metres. If the intensity of the incident pulse is increased then the effective distance over which it will still promote fluorescence may be greater than 10 metres. However, if the intensity is greater than a certain threshold there is a risk of eye damage to bystanders.

The apparatus may include more than one point of delivery of laser illumination. Each point may be supplied with light from the same or a different source. In other words, there may also be more than one source of electromagnetic radiation provided. Each point may deliver laser light having the same or a different wavelength from the other point(s). This may be effected by the suitable arrangement of well-known equipment to filter the source light.

To prolong the life of the electromagnetic radiation and associated equipment, the laser beam may be produced only as required. The control of the production of the laser beam may be automatic, semi-automatic, or manually operated. For instance, if automatic, a beam-break could be employed to operate the apparatus such that a laser beam is produced. As an example, in a typical airport situation the beam break could be provided across the conveyor belt on which luggage is carried so that it detects when a piece of luggage is in the correct area for being illuminated by the laser light. Another example is a beam break being used to trigger an illumination of a person or group of people for detection of controlled substances associated with that person or group of people.

Additionally or alternatively, the laser source could be operated so that it produces continuous pulses of incident light.

The receiving/capture means may comprise a CCD camera which may receive radiation at various wavelengths in the electromagnetic spectrum. In particular, the receiving means is able to receive ultraviolet radiation at least. Further, in one embodiment the receiving means is also capable of receiving electromagnetic radiation in the visible range (approximately 400nm to 700nm).

Means for controlling the capture of images by the receiving means may be included. For instance, a shutter may be provided which is variable between an open receiving position and a closed non-receiving position. When the shutter is open an image may be captured. The time period during which the shutter is open may be known as the sampling period. This sampling period may be set at a predetermined value lying between l ps and 100ns. More preferably it may lie between 5ns and 35ns.

The camera may comprise an array of pixels. For instance, the array may be 640 by 480 pixels in size. At a distance of 1.5 metres from the camera to the target each pixel may view an area of the target having a dimension of approximately 0.3mm in diameter. If the distance between the target and the camera lens is increased then the area of the target viewed by each pixel will increase. The converse is also true. However, it is possible to use a camera having an array with a greater or lesser number of pixels which would increase or decrease respectively the definition of the viewed target.

The receiving means are connected to analysis means, which may in one embodiment be a computer, such that the radiation received by the receiving means is analysed by the analysis means.

The apparatus may include control means for controlling the various parts of the apparatus. For instance, the control means may control the source of electromagnetic radiation. This may allow the intensity, duration and other aspects of the incident pulse of electromagnetic radiation to be varied. Further, the control means and analysis means may both be provided by the same computer.

The apparatus includes processor means for processing the intensity values to improve the confidence of the classification of substances.

In one embodiment, the analysis means is capable of analysing the fluorescent radiation received from the target area pixel by pixel and is capable of determining a substance on a pixel by pixel basis. In other words, it is possible for the intensity of the fluorescent radiation to be analysed on a pixel by pixel basis.

In this manner, each pixel may be repeatedly analysed to identify if there is any received fluorescent radiation and if so may determine the intensity and temporal position of the peak of that radiation. It is also possible for any pixel which has been identified as having received fluorescent radiation to be further analysed at more than one particular point in time to determine the intensity of the received fluorescent radiation at each combination of incident and resultant wavelengths so that their difference(s) may be compared with that of known substances held in the database.

In this way, the location of a determined substance may be identified by a pixel or group of pixels in an image recorded by the camera. If the camera has also captured an image of the target area in the visual range of the electromagnetic spectrum then it may be possible for the location of the determined substance to be indicated on this visible image.

In this connection, the apparatus may include means for splitting the captured image, via well known image splitting means, into a plurality of identical images. The apparatus may also include means for filtering the images. For example, one split image may be filtered to exclude all but ultraviolet radiation for capturing via a CCD camera as discussed above. Another split image may be filtered to exclude all but the visual image and this may be directed to recording means and/or image display means. An electronically controlled rotating cassette having a range of different filters may be used to filter the received radiation prior to it reaching the camera. The filters may filter the received resultant radiation in the range 320nm to 420nm. The filters may be band-pass filters which provide a filtered radiation having a range of wavlengths. Once a substance has been classified by the analysis of the captured ultraviolet radiation, its location may be "overlain" onto the visual image so that an operator may be able to visually locate the substance on the target area.

The image display means may be a VDU monitor. This monitor may also be used for displaying the results of x-ray analysis of the target. The three types of display (visual, x-ray, ultraviolet) could be shown simultaneously or non-simultaneously on the same monitor.

The apparatus may include means for communicating to an interested party the specific identification of a substance identified in the target area. As examples, this may take the form of an indicator light, a pre-defined message, or a sound.

The apparatus may include more than one camera each having an associated filter which may filter the received fluorescent radiation at the same and/or different wavelength from the others. This may allow for faster identification of substances.

In another embodiment, the resultant fluorescence is analysed but rather than an image of the location of any classified substance of interest being overlain onto a visual image of the target the apparatus may provide an indication, such as a light, sound or other determinable signal, that a substance of interest has been classified.

Embodiments of the present invention will now be more particularly described, by way of example only, with reference to the accompanying drawings, in which:

Figure 1 is a perspective view of an arrangement of apparatus according to one embodiment of the present invention;

Figure 2 is a typical graphical example (line graph of intensity versus time) of laser induced received fluorescent radiation; Figure 3 is an example of captured intensity values for two combinations of incident and resultant wavelengths;

Figure 4 is an example of a over-sampling resultant fluorescent radiation for averaging purposes;

Figure 5 is typical graphical example (line graph of the natural logarithm of intensity versus time) of laser induced received fluorescent radiation; and

Figure 6 is an example of how line graphs of values for the natural logarithm of intensity versus values for time can be used to determine parameters in the equation describing intensity decay over time.

Figure 7 is a schematic representation of an alternative system for implementing the methods of the present invention.

In Figure 1 apparatus 10 for the remote identification of substances is schematically shown. The apparatus 10 comprises an ultraviolet laser 40 shown emitting a dispersed ultraviolet laser beam 50 towards the target, which in this example is a box 20 located on a conveyor belt 30. However, the target may be a person or group of people and/or other objects. The apparatus 10 also includes a camera 60 which is shown positioned facing the target area in such a way as to receive any resultant fluorescent radiation arising from the incident laser beam. The camera may be an image intensifying camera. Although the camera 60 and ultraviolet laser 40 are shown immediately adjacent one another they may be located in separate and spaced positions. The apparatus also includes a processor 80 to which the laser 40 and camera 60 are connected by connection means 70. These means 70 may be wired or wireless. The apparatus also includes an image display means 100 to which the processor 80 is connected via communication means 90 which again may be wired or wireless. The image display means 100 shows an image 1 10 which includes the conveyor belt 30 and target box 20. A substance has been identified and is indicated by a highlighted area 140 on the image of the box 20. A keypad 105 is also connected to the apparatus for interaction therewith.

The processor 80 includes control means for controlling the operation of the various components of the apparatus 10. For example, the processor 80 may control the laser 40 and the camera 60. It may also control analysis means and identification means (not shown), which may form part of the processor 80 or may be separate. It may control the sampling period and the delay between the incident pulse and the time when samples (or capture) of the resultant fluorescent radiation occur.

In Figure 2, a relaxation profile of a known substances is illustrated. The profile is plotted on a line graph having intensity as the ordinate (y) axis and time as the abscissa (x) axis. The relaxation profile is obtained by measuring or determining the fluorescent radiation over time resultant from an electromagnetic pulse such as from a UV laser. The profile takes the form of a rise in intensity, which temporally follows the incident pulse, to a maximum, before decaying in an exponential decay curve. The rise time of the fluorescent radiation may be governed by the laser pulse width (in time). The resultant fluorescent radiation intensity may be determined by the convolution of the laser pulse and the fluorescent decay function. The temporal position of the peak intensity relative to the incident pulse (not shown) may be reliant on the distance between the source of the pulse and the target and the distance between the target and the means of capture of the resultant fluorescent radiation. The magnitude of the intensity values are reliant on various factors such as the quantity of material present at the target area. The curve is for a particular wavelength, the resultant fluorescent radiation having been filtered to produce this selected wavelength.

In its simplest form the method may be partly embodied by the capture of two intensity values for each of two combinations of incident and resultant radiation wavelengths, where the resultant radiation is as determined from resultant fluorescent radiation derived from a target substance, the substance having been excited by a pulse of suitable incident radiation. Figure 3 shows such a situation. Two exponential decay curves A,B are shown each having a different combination of incident and resultant radiation wavelengths. These curves A,B are sampled, or have radiation captured, at two points in time ti, h. The resultant intensity values captured are represented by Un, Ut2, Iβn and Iβt2. For the purpose of this example, the points in time ti, h are the same for both curves A and B. The proportions of Iβn to Un and of Iβt2 to Ut2 are determined. These proportions are one example of a difference of the aspects of the captured radiation. These proportions are then compared to proportions of known intensity values for the same points in time ti, t2 (relative to a point such as the temporal location of the peak intensity) and for the same combination of incident and resultant radiation wavelengths for various substances. If the two sets of proportions are within a tolerance of each other then the substance in question may be classified in accordance with the known substance. If this comparison is effected by a processor and database (i.e. a computer) these comparisons can be made relatively quickly. Because it is the proportions that are being prepared it does not matter that the quantities of substance or the distances involved may be different between the known and determined values. To improve the confidence of the intensity results oversαmpling may be effected whereby several values for the intensity are determined at the same and/or substantially the same and/or different points in time for each combination of incident and resultant wavelengths so that a range of values for each curve is determined as shown in Figure 4. The values may then be averaged or otherwise processed to provide a single intensity value for each time increment such that a curve may be plotted. A best-fit curve may be plotted.

In these examples, the capture means or camera may be controlled to capture received fluorescent radiation over any selected time duration. Although, the graphs imply that these are singular points in time the capture of the radiation will last, in reality, for predetermined periods such as several picoseconds. Processing, such as by normalising, of the results can correct for differing time durations to allow meaningful analysis of the results.

Since the decay curves are exponential if the natural logarithm of the intensity values are plotted against time the result is a straight line graph as shown in Figure 5. This Figure also shows the background radiation C. The straight line graph may be obtained by plotting the logarithm of the raw intensity data captured by the apparatus according to the method. Alternatively, the logarithm of intensity values which are the result of processing of the raw data (such as by the use of best-fit curves or simple averaging) may be plotted. It is expected that in some case the use of processed values may result in lines which are more straight then the use of raw data vales.

Once the straight lines are plotted, the parameters describing the form of the lines may be obtained. These parameters may also be used to describe the exponential decay curves by finding the inverse logarithm of the parameters. The parameters in question are X and γ which are taken from the equation I = X.e^{~γ t} where I is the intensity, X is an intensity parameter, e is the base of natural logarithms, γ is the decay time parameter, and t is time.

The natural logarithm of X is the point at which the straight line crosses the y-axis. The natural logarithm of γ is the gradient of the straight line. These parameters may be determined for each line which represents the resultant fluorescent radiation for a unique combination of incident and resultant wavelengths for each target.

In Figure 6, four straight lines D, E, F and G are shown. These are examples of the natural logarithms of intensity plotted against their respective time values (or of the intensity and time values plotted against one another on a semi- logarithmic scale). Each has a gradient and each crosses the y-axis (the line being extended to cross the y-axis if necessary). The point at which the lines cross the y-axis are equivalent to the X values (XD, XE, XF, XG). The gradient of each line D, E, F and G is equivalent to the decay time parameter γo YE YF YG for each combination of incident and resultant wavelengths. Furthermore, intensity values Im, \m. toi, \m, bti, bt2, Uti and Ut2 may be determined for times ti and h for each combination of incident and resultant wavelengths. These may be calculated from the equation In(I) = ln(x) -γ.t since X, γ, t are known.

It may be seen that once the parameters for each line or curve are determined, by whichever method, differences in their aspects (whether intensity values, parameters, time values, etc) may be compared with a database which includes the differences, or the same type of parameters for determining, the differences, of known substances. The use of the term "plotting", in the foregoing, with regard to the lines and curves may be taken to mean the metaphorical or virtual plotting of such lines and curves since, in one embodiment, the method may be carried out by a computer which may not necessarily actually or physically plot any graphs.

Although natural logarithms have been used in the above examples and description the invention should not be interpreted to be limited thereto as the use of other bases are contemplated. A system for implementing the above described embodiment of the present invention is shown in Figure 7. The system comprises: i. Dual short pulse (2ns) diffused beam Nd;YAG UV Lasers (266 &

355nm) (102), ii. Dual nanosecond gated Image lntensifer cameras with UV optics and selectable optical UV band pass filters ( 104) , iii. CCTV camera for Image overlay and evidential recording ( 106) , iv. Industrial FPGA processor ( 108) with picosecond accuracy to control system timing, v. P. C (1 10) providing vision processing, detection algorithms, Operator Interface (GUI) and remote access via TCP/IP.

The system (100) operates as follows: The target (person, object etc.) is illuminated by the short pulse (2ns) Diffused Beam Lasers ( 102) . After each laser pulse the resulting fluorescence image is captured by the nanosecond gated Image intensified cameras (104) and converted into a digital Image. This digital image is analysed in real time both in the spectral and temporal domains by the analysis software, which provides an automatic indication of detection of a suspect substance trace. To further aid identification the black and white Image of the target from the context camera is overlaid to show the "area of Interest" on the target. The system is capable of operating at up to 30 Frames (Images) per second and only a few frames are necessary for a detection.

The lasers, camera's imaging optics and spectral filters are mounted in a head assembly. This assembly may be tripod mounted depending on the operation requirement. The head assembly is connected via umbilical cords to the laser power supply's/control computer cabinet.

It is important to realise there are two distinct elements to the decay process from a measurement point of view, namely the mean level of the instantaneous emission of photons is given by an exponential decay function: exp(-k*t) and the number of photons measured in any given time slot are subject to a statistical distribution. This distinction, particularly the statistical variation, is often overlooked. In the embodiments previously described the following considerations apply.

In the system previously described the number of photons that are available in a given time period are caputured, ie the camera gate is opened at time ti and closed at time h

Fluorescence decays like other decay phenomena. The number of photons captured in any given time period (after excitation) follows Poisson statistics. With a reasonable number, say >10 photons captured, this distribution is approximate to a Gaussian distribution.

One interesting and useful property of the Poisson distribution is that its Variance equals its mean value. It will be understood that when describing the statistics etc it is important to differentiate between mean values and instantaneous values. The instantaneous mean value s of the decay process can be described by an exponential function:

s = exp(-0.693t/ β) where β is the half life

(-0.693 = In (0.5)) so that when t = β the level has dropped to a half.

What is measured (in terms of mean values) is the integral of this function between the two camera gate opening/closing times. In any one case the measured value will be a sample from the Gaussian distribution whose mean and variance are given by this integral.

These are important considerations to the methodology of the present invention. By considering mean values only, it is possible to make false assumptions of what will and what will not work. If it is possible only to capture a small number of photons, the variance will be very high and it may not be possible to distinguish one substance from another. In order to reduce the standard deviation relative to the mean values it is necessary to capture as many photons as possible. This can be achieved by moving the detector (camera) closer to the target (1/r² effect), a larger detector area (aperture squared - but not necessarily a smaller lens f no.), more excitation power or by integration over many samples. The update rate is normally constrained by the maximum repetition rate of the laser / camera system typically 50-100 Hz. Hence, it can be understood that if the process has to operate in say 1 second, for example in a body or baggage scanner, there may be severe constraints placed on the system compared with a scenario where one could take say 60 seconds or more, when examining a static object. In addition, it will be understood that the number ot photons that determine the statistics is not the same number of photons hitting the front of the camera system, but the number hitting the first stage detector. In the case of a UV camera on its own it is possible to get Quantum Efficiencies (QE) of the order of 50% or above, as determined by the camera sensor, normally a CCD or a CMOS sensor. Sensor noise is not an issue in the present system because the exposure time is very, very short compared with that used on astronomical telescopes, for example, readout noise may need to be taken into account.

In embodiments of the system described above, which preferably comprises a Multi- Channel Plate (MCP) to give accurate camera gate opening times of the order of a few ns, the QE is low, peaking at 15% and dropping off to less than 10% for some wavelengths. The QE in this case is determined by the first cathode in the MCP subsystem.

When the camera or cameras are calibrated, the calibration relates the energy level (photons at a given wavelength) at the front face of the camera, to the readout value obtained for a given camera shutter time. The following considerations therefore apply: 1. If the camera lens is changed the calibration must be redone or an allowance made for the difference in aperture / stop setting of the lens 2. As noted above it is the quantum efficiency, or the number of photons triggering the first detector, that determine the statistics.

There are in effect two parameters that can be measured, directly or indirectly. The peak amplitude of the fluorescence at the start of the decay process (ie t = 0) and its half life β. It will be understood that any one material / substance can be made up of more than one fluorescing material and hence the measured fluorescence can have a mixture of these values. For practical purposes the system measures the parameters of one approximate average value.

It is important therefore to recognise the potential errors that can be introduced by the equipment itself, when trying to estimate mean values.

The laser and its diffuser will typically have a variation of around ± 3%. This can be eliminated / reduced by measuring the laser power on each shot. The laser power can be measured by a feedback system fitted to the laser or the system may allow for the inclusion of any pulse to pulse variation in the diffuser. Lasers operating in multimode produce more power but the different modes within the laser tube can present a variation in intensity across the emitting face of the laser tube, which is present at the target. The optical diffuser helps to eliminate this latter effect.

The measurement of power at the target can be obtained by placing a known material of sufficient area near the target such that by integrating over this reference material the laser power can be estimated.

The camera / receive subsystem is also subject to a number of errors.

These are caused by jitter in the timing system, partly in the overall control subsystem which controls the laser firing and the camera trigger. The camera subsystem itself has jitter on its gate opening and closing times.

Together these amount to a scatter of around ± 0.75 to 1 ns in effective gate opening time with respect to the laser firing. Taking ±750ps as a typical error and recognising that short half-life material eg cotton where β is around 2ns and a long half-life material eg masking tape where β is around 15ns, jitter can impose a significant error on any amplitude measurement, especially if the gate opening time is 20ns or thereabouts, as typically it may be.

A computer model of the mean values expected in any gate opening times has shown that for short half life material the amplitude measurement may be subject to errors as large as ±30% and for long half-life materials around ±3% (typically the same as the error in the laser before correction).

Timing errors can imply that there is little point in trying to measure half-life to any degree of accuracy, particularly for short half-life materials

However, what can be measured is the total energy in any filter band by integrating from before the peak response is obtained to after the fluorescence has decayed. In practice this can achieve the degree of accuracy required' for the camera gating for this mode of operation without the need for a MCP in front of the camera.

A camera fitted with an MCP can additionally be used to divide long half-life substances from short half-life substances without obtaining an explicit measurement of half-life. This can further improve substance identification by providing an additional level of discrimination.

The present invention therefore contemplates other embodiments whereby the apparatus and the method of operation are modified compared with the system previously described. As previously mentioned, the laser may be switchable, electronically, between two or more excitation wavelengths, 266nm and 355nm as previously proposed. Jn one embodiment separate lasers may be provided for faster operation but in practice only one laser may be used, for example 266nm. The present inventors have found .that substantially all the data generated by the longer wavelength laser is included in the data generated by the shorter wavelength laser. The band of interest is typically 320nm to 450nm and therefore the shorter wavelength laser is more suitable for this waveband without detriment to the longer wavelength returns.

In preferred embodiments an electronically controlled rotating cassette of filters is provided in front of each camera. It is of course possible to provide a plurality of cameras each having a different respective filter if faster operation is required. As previously mentioned in order to obtain spectral information it is not necessary to use an image-intensified camera, since accurate gate timings are not required. In previous described embodiments the image intensifier is in practice only used to accurately control the camera gate open time, it does not provide any direct image

• intensification. The present embodiment contemplates the use of a UV sensitive camera. A UV camera on its own does not provide very short gate opening times, that is to say the opening times are in micro-seconds rather than nano-seconds, the later being achievable with an image intensified camera. However, the present embodiment is concerned with capturing the total fluorescence in a given filter band during the relevant time period of interest where the UV camera shutter may be left open.

As implied above a simplified process can be used which reduces the cost of the equipment. This is based on using one or more cameras to capture the spectrum of the total fluorescence as described above. This may be augmented, where necessary by using an image intensified camera, ie with a fast shutter capability, to identify whether the fluorescence lasts for a short or a relatively long period of time. This may or may not use a selection of filters, in the simplest case only one filter is used to constrain the incoming fluorescence to the band of interest (~ 300 to ~450nm). One or more time bands may be used. This approach leads to a reduction in variation in the identification achieved when there is only a low number of photons in each return. If the band of interest is extended to ~450nm it will be necessary for any background illumination of the scene to be such that there is minimum background light energy in that band of interest. This may be achieved by using standard fluorescent light fittings fitted with a suitable filter on the tube.

In this embodiment a simplified classification method is used. Substances are separated by the use of a least square (or similar) distance measurement between the relative spectrum level in each of the filter bands used in front of the camera (s).

The spectra are normalised by the total photon count across the total band on interest. This is achieved by summing the outputs from all filters used. This normalisation removes any variation due to either a variation in the amount of substance or the power level of the illumination at each pixel of interest in the image.

If an image intensified camera is used to determine long or short fluorescing times, this can be added as a weighted input to the least squared classification process.

In either case superior separation may be achieved by applying a series of weights to the various parameters (filter bands / fluorescing time). An exemplary system constructed in accordance with the modified embodiment of the invention is described below.

The system comprises a 266nm laser, or similar, equipped with a high grade diffuser. The laser operates at the highest possible repetition rate (50 to 100Hz) compatible with the rest of the system and the maximum power available, compatible with the application.

As previously mentioned it has been found that a 355nm laser provides no additional information to that provided by the 266nm laser in terms of spectral content of the return signal. The spectral content available with the 266nm laser (from ~300nm upwards) has been found to give a sufficient spectral range for discrimination.

A means of spreading or diffusing the laser beam to cover the area of interest is required. Ground glass screens can have too great a variation across the field of interest. Although it may be possible to eliminate most of this by processing, it is possible to provide diffusers which are designed for a specific area of coverage such as square or rectangular etc.

The sensitivity of the system and the error rate are dependent on obtaining the maximum number of return photons. From an illumination point of view this number, in any given time, is proportional to the product of the illuminating power density at the target and the repetition rate.

The maximum power will be determined by the particular application and may be restricted, for example when use involves exposure to human skin or eyes. The detection system comprises two or more cameras, typically greater than 480 640 pixels.

The cameras are sensitive from the UV to the yellow visible spectrum (~290nm to 500+nm). Sensitivity is normally defined in terms of the 'Quantum Efficiency' ie the % of photons at the face of the system which are captured by the detector}

The cameras are equipped with rotating filter wheels when the number of filters required exceeds the number of cameras used. The filters are synchronised with both the laser firing sequence and with each other. A timing system ensures that a filter is fully in front of the optical capture system when the laser fires. The filters on each camera are to be capable of synchronisation such that the filter sequence (fl , f2, f3 etc) is identical on each camera.

A two camera system may be preferred as a minimum cost system, however, systems with more than two cameras are also envisaged.

The filters on the first camera consist of one visible region filter, although not necessarily required the laser pulse can be skipped when this filter is in use, and the remaining filters are identical wideband filters covering the range of the spectral filters on camera two.

The filters on the second camera consist of one visible region filter. The remaining filters comprise a range of filters covering the spectral area of interest. A possible range would be 7 or 8 adjacent filters, 20nm wide, with a starting centre wavelength of 315nm. The processing system (see below) is based on comparing the total photon response to the photon response in a given band. The two camera system enables this to happen for each laser shot at the same time eliminating any variation in total laser power (-3%) or across-scene variation caused by the laser beam diffusion system. However, the dual system does not eliminate any variation caused by the statistics of the photon decay.

The two visible region filters allow for the production of a visual image to help identify where any detected explosive or other substance is found in terms of the visual scene. These visual images are also used to align the camera systems, which may be mechanical or by means of a two dimensional transformation in the processing, or most probably by a combination of techniques. Use of a visible region filter also eliminates a dedicated visual camera.

Each camera is equipped with a UV transparent optical system.

The optical system may be either lens or mirror based. The capture aperture is the important parameter in terms of the number of photons captured. The system is such that the target image area is captured by the sensor (eg CCD) in the camera.

Visible alignment markers may also be necessary in order to map pixels from one camera to another.

The area in which the system is to be used is preferably illuminated with a light source that has no significant radiation below - 450nm. This cut off point is determined by the upper filter band used for spectral decomposition of the return photon spectra. Suitable background illumination can be achieved either by means of a specific lamp source (eg sodium) or by applying a filter to a typical fluorescent lamp. Suitable fluorescent lamps suppress all UV radiation and are then overlaid with a filter to give a high-pass (in nm) effect.

The detection and classification system is based on a comparison of the spectral properties of each pixel in the scene against a data base of known substances.

In outline the processing consists of obtaining the spectral pattern from each filter in the system for each relevant pixel in the image. This consists of dividing the total broad-band photon count (Y) into the photon count for each filter [X₁); these counts may be obtained from one laser shot or are the results of many laser shots. The error in classification will reduce as the number of photons received increases. The statistics of Xι/Y where X and Y are Poisson distributed can be readily obtained. The method of comparison may be any of the traditional nearest neighbour methods, for example the least square distance method.

Although aspects of the present invention have been described with reference to the embodiments disclosed herein, it will be readily apparent to the skilled person that the present invention is not limited to those precise embodiments but contemplates arrangements with various modifications and changes which may be effected by the skilled person without further exercise of inventive skill and effort. For example, it is to be understood that the excitation source does not necessarily have to be a laser. The present invention contemplates embodiments where the source of electromagnetic radiation comprises an LED or the like or a continuous narrowband light source, for example a mercury/xenon lamp, with or without a filter provided on the light source. In this respect it will be understood by the skilled person that the period over which the emission response is recorded will not necessarily be the decay period previously referred to, although it may be the same, less or more probably greater than the decay period as determined by the particular requirements of the specific application. It will be understood that it does not necessarily matter if the target or target area is continuously illuminated as it is the relative spectral power in the different wave bands that is of interest and is what is measured in methods of the aforementioned aspect of the present invention.

In addition, it is to be understood that the detecting means may not necessarily be a camera, at least in the conventional sense of the word "camera", that is to say the detecting means may be a CCD which may be a one dimensional array, for example a 1 x 256 pixel device or a two dimensional device such as a 4 x 512 device, or possibly a line of light detectors, possible if the subjects of the examination are moving relative to the detector, for example a bag on a conveyor belt. In this respect it is possible to use the relative movement to produce a synthetic image of the target by taking many readings (scans) from say a vertically mounted sensor as a bag moves horizontally.

Claims

1. A method for the remote classification of substances; the said method comprising the steps of: a) emitting incident radiation from a source of electromagnetic radiation towards a target or target area; b) monitoring the said target area for fluorescent radiation emitted from the from the target or target area in response to said incident radiation; c) recording the emission response of the target or target area in response to said incident radiation; d) comparing the emission response of the target or target area with reference emission response data for known substances/materials; e) identifying a target substance or substance in said target in accordance with said comparison wherein the emission response of the target or target area is recorded for at least two filtered wavebands and step d) is repeated for each of the said wavebands.

2. A method as claimed in Claim 1 wherein step c) comprises the step of recording substantially the whole fluorescence emission response of the target or target area for each respective waveband.

3. A method as claimed in Claim 2 wherein the recorded fluorescence emission response of the target or target area for each respective waveband corresponds to the number of detected photons recorded in the respective waveband during the decay period or period of continuous illumination of said target or target area by said incident radiation, and wherein the method further comprises the step of normalising the recorded values for each waveband with respect to the total number of photons detected for all said wavebands during said period to provide non-dimensional data for comparison with corresponding reference data in step e).

4. A method as claimed in any preceding claim wherein step e) comprises the step of determining a cross-distance measurement for each of the said normalised recorded values for the said respective wavebands with corresponding reference emission data for said known substances, and identifying which of the said known substances is present by identifying the best match with said reference data.

5. A method as claimed in Claim 4 wherein the said cross-distance measurement comparison comprises a least squares method, utilising the root mean squared value of the differences in the respective measured and reference emission response values.

6. A method as claimed in Claim 5 wherein the said measured emission response values for the respective wavebands are selectively weighted prior to said comparison with said reference values.

7. A method as claimed in claim 2 or Claim 3 wherein the fluorescence emission responses for the respective filtered wavebands are recorded either simultaneously by separate recoding means or sequentially by means of selective filters sequentially applied to the recoding means.

8. A method as claimed in any preceding claim wherein the said filtered wavebands comprise at least UV filtered wavebands.

9. A method as claimed in any preceding claim wherein the said filtered wavebands are selected from the range of 300 to 450 nm.

10. A method as claimed in any preceding claim wherein the said wavebands are substantially equal, for example between 20nm and 30 nm.

1 1. A method as claimed in Claim 8, Claim 9 or Claim 10 wherein at least one of the said wavebands is in the visible spectrum for the simultaneous visual display of the target or target area with the detected substance so that the spatial position of an identified substance or material on the target or in the target area can be superimposed and shown on a visual display in use.

12. A method as claimed in any preceding claim wherein the incident radiation comprises an emission frequency capable of generating a fluorescence emission response in target substances to be identified, and wherein the incident radiation is preferably laser generated UV radiation, preferably laser generated and more preferably having a wavelength in the region of 266nm.

13. A method as claimed in any preceding claim further comprising the step of measuring the distance of the target or target area from the electromagnetic radiation source emitter and/or the distance of the target area from the fluorescence emission recording means.

14. The method of any preceding claim, including the step of determining the temporal location and/or magnitude of the peak intensity of the resultant fluorescent radiation for each combination of incident and resultant wavelength.

15. The method of any preceding claim, wherein an intensity value for background fluorescent radiation is determined.

16. The method of any preceding claim wherein determined intensity values are compared to the background fluorescent radiation, the results of said comparison aiding the classification of substances in the target area.

17. The method of any preceding claim, wherein substances are classified into narcotics or explosives.

18. The method of Claim 15, wherein substances classified as explosives are also classified into types of explosives.

19. The method of any preceding claim wherein the step of recording the resultant fluorescent radiation is effected by means of a digital camera or the like, whereby fluorescence decay data is recorded on a pixel by pixel basis by said digital camera or like device.

20. The method of any preceding claim further comprising the steps of capturing and displaying a visual image of the target or target area and indicating on the displayed image the location of any classified substances.

21. The method as claimed in any preceding claims wherein the said step of emitting incidental radiation comprises the step of emitting at least one pulse of electromagnetic radiation or continuously illuminating the target or target area with electromagnetic radiation.

22. The method as claimed in Claim 21 wherein the step of continuously illuminating the said target or target area comprises the step of emitting electromagnetic radiation from a narrow band light source or LED.

23. The method of any preceding claims wherein the step of recording the emission response of the target or target area comprises the step of recording substantially the whole fluorescence emission decay response of the target or the target area.

24. Apparatus for the remote classification of substances comprising: electromagnetic radiation emitting means for emitting incident radiation from a source of electromagnetic radiation towards a target or target area; monitoring means for monitoring the said target area for fluorescent radiation, emitted from the from the target or target area in response to said incident radiation; recording means for recording the emission response of the target or target area in response to said incident radiation; comparison means for comparing the emission response of the target or target area with reference emission response data for known substances/materials; identification means for identifying a target substance or substance in said target in accordance with said comparison wherein the emission response of the target or target area is recorded for at least two filtered wavebands and step d) is repeated for each of the said wavebands.