Systems and methods for generating a 3D colour image of an object

Abstract

An imaging system for generating a three-dimensional colour image of an object, 305, the system comprising one or more light sources, 301, configured to illuminate the object with light of different wavelengths and a monochrome camera, 303, comprising an array of pixels arranged to receive light reflected from the object when the object is illuminated by the one or more light sources. The system is configured to capture a first plurality of images on the monochrome camera when illuminating the object with a first set of wavelengths of light and to further capture a second plurality of images, wherein for each one of the second plurality of images, the object is illuminated with a different respective set of wavelengths of light. The system further comprising a processor configured to generate, based on the first plurality of captured images, a 3D image of the object and determine, based on the second plurality of images, a colour value for each point in the 3D image.

Description

FIELD Embodiments described herein relate to systems and methods for generating a 3D colour image of an object. BACKGROUND Three-dimensional surface imaging (3D surface imaging) is a fast growing field of technology. The term “3D surface imaging” as used herein can be understood to refer to the process of generating a 3D representation of the surface(s) of an object by capturing spatial information in all three dimensions - in other words, by capturing depth information in addition to the two-dimensional spatial information present in a conventional image or photograph. This 3D representation can be visually displayed as a “3D image” on a screen, for example. A number of different techniques can be used to obtain the data required to generate a 3D image of an object’s surface. These techniques include, but are not limited to, structured light illumination, time of flight imaging, holographic techniques, stereo systems (both active and passive) and laser line triangulation. In each case, the data may be captured in the form of a “point cloud”, in which intensity values are recorded for different points in three-dimensional space, with each point in the cloud having its own set of (x, y, z) coordinates and an associated intensity value I. The point cloud may in turn be rendered to produce the 3D image of the object, which may be viewed from multiple angles. When rendering the 3D image, it is often desirable to supplement the image with colour information; the colour information is particularly useful when using 3D imaging for object recognition, for example. In order to do so, it is necessary to capture measurements of colour during the imaging process. Typically, such measurements are obtained by using a colour camera for the imaging process. A colour camera comprises an array of pixels with different coloured filters placed over each pixel. An example of such a camera is shown in Figure 1A. Here, the colour filters are arranged in a Bayer layout, which can be further understood with reference to Figure 1B. The array can be divided into cells of two-by-two pixels, with Figure 1B showing one of the cells in close-up. Within each cell, two of the pixels, arranged on the diagonal, are provided with green emission filters. Of the remaining two pixels, one is provided with a red filter and the other is provided with a blue pixel. The number of green pixels is, therefore, double the number of red pixels or blue pixels; this is owing to the human eye having a higher sensitivity to green light. The red, blue and green filters will only allow light of certain wavebands to pass through and reach the photodiode in the underlying pixel. Figure 2 shows how the quantum efficiency of each pixel will typically vary depending on the colour filter overlaying that pixel. The quantum efficiency provides a measure of the number of photoelectrons produced by a pixel as a fraction of the total number of photons of all colours incident on the pixel. The curve 201 shows how the quantum efficiency varies for pixels having a blue colour filter. For these pixels, the quantum efficiency is greatest for light at a wavelength of 450nm, and falls off to zero for light having wavelengths of 550nm to 750nm, before rising slightly again towards 800nm. The curve 202 shows how the quantum efficiency varies for pixels having a green colour filter. In this case, the quantum efficiency rises steeply from 460nm to a peak at around 530nm, before falling again to a minimum at 650nm and then rising to a slightly higher level at 725nm from where is plateaus. The curve 203 shows how the quantum efficiency varies for pixels having a red colour filter. Here, the quantum efficiency is close to zero for wavelengths below 560nm, before rising steeply to a peak at around 600nm and then falling more gradually towards higher wavelengths. As a final comparison, the curve 204 shown how the quantum efficiency of pixels varies in a monochrome camera having the same array of pixels but with the Bayer filter no longer being present. As might be expected, the curve 204 forms an envelope around the peaks in the respective curves 201, 202, 203, with the quantum efficiency being noticeably higher for all wavelengths between 400nm and 800nm. It can be seen from the above that the use of the Bayer filter to distinguish colour in the image poses a number of issues. To begin with, the spatial layout of the green, red and blue filters means that the spatial resolution in each colour channel is effectively halved compared to that of a monochrome camera; this is because photons of different wavelengths will only be detected by those pixels whose colour filter coincides with their wavelength. Although various demosaicking algorithms can be used to reconstruct a full resolution image from the captured data, such algorithms can result in artifacts appearing in the image. For example, aliasing may occur in the event that the image contains a set of highly repetitive features. The loss in resolution notwithstanding, a further concern is that the overall light signal is greatly reduced, since only those photons whose wavelengths match the colour filter of the pixel on which they fall will be captured. This loss of signal can be particularly detrimental in the case of 3D imaging, where signal-to-noise requirements are typically much higher than for 2D imaging. Using a monochrome camera instead of a colour camera for 3D imaging can provide a higher signal-to-noise when reconstructing the 3D image of an object, as photons of all colours can be detected at each individual camera pixel. However, this then comes at the cost of sacrificing the colour information available when using a Bayer filter. It is desirable, therefore, to develop further means for capturing 3D images of objects with colour information included in the image. SUMMARY According to a first aspect of the present invention, there is provided an imaging system for generating a 3D colour image of an object, the system comprising: one or more light sources configured to illuminate the object with light of different wavelengths; and a monochrome camera comprising an array of pixels arranged to receive light reflected from the object when the object is illuminated by the one or more light sources, the system being configured to capture a first plurality of images on the monochrome camera when illuminating the object with a first set of wavelengths of light; the system being further configured to capture a second plurality of images, wherein for each one of the second plurality of images, the object is illuminated with a different respective set of wavelengths of light; the system further comprising a processor configured to: generate, based on the first plurality of captured images, a 3D image of the object; and determine, based on the second plurality of images, a colour value for each point in the 3D image. The system may comprise one or more emission filters, wherein when illuminating the object with a respective set of wavelengths of light, the one or more emission filters are arranged to allow the wavelengths of light in the respective set of wavelengths to pass through towards the monochrome camera. The one or more emission filters may comprise a single emission filter having a plurality of pass bands, each pass band coinciding with a respective set of wavelengths used to illuminate the object. The one or more emission filters may comprise a plurality of emission filters insertable into the light path between the object and the camera. For each image captured by the system, the emission filter selected for insertion into the light path may be one whose pass band coincides with the set of wavelengths being used to illuminate the object at the time of capturing the image. The one or more emission filters may be arranged such that when illuminating the object with light in the first set of wavelengths, the light is able to pass through the filter towards each one of the pixels in the array. The imaging system may be configured to capture the second plurality of images on the monochrome camera. For each one of the first plurality of images, the system may be configured to illuminate the object with a spatially varied 2D illumination pattern, the pattern being altered for each one of the first plurality of images. The processor may be configured to generate the 3D image of the object by determining, for each pixel in the array, a variation in intensity of light incident on the pixel across the first plurality of images. The one or more light sources may be configured to sequentially illuminate the object with light in a blue spectral band, green spectral band, and red spectral band. The one or more light sources may comprise one or more light emitting diodes, LEDs. According to a second aspect of the present invention, there is provided a method of generating a 3D colour image of an object, the method comprising: capturing, on a monochrome camera, a first plurality of images of the object when illuminating the object with light having a first set of wavelengths; capturing a second plurality of images, wherein for each one of the second plurality of images, the object is illuminated with a different set of wavelengths; generating, based on the first plurality of captured images, a 3D image of the object; and determining, based on the signal received in each one of the second plurality of images, a colour value for each point in the 3D image. BRIEF DESCRIPTION OF DRAWINGS Embodiments of the invention will now be described by way of example with reference to the accompanying drawings in which: Figure 1 shows an example of a pixel array in a colour camera having a Bayer filter; Figure 2 shows how the quantum efficiency varies as a function of wavelength for the different filters employed in the Bayer filter of Figure 1; Figure 3A shows an imaging system according to an embodiment; Figure 3B shows a simplified diagram of the geometry of the imaging system of Figure 3A; Figure 4 shows an imaging system according to an embodiment; Figure 5 shows an example of a pixel array in a monochrome camera; Figure 6 shows an imaging system according to an embodiment; Figure 7 shows pass bands in the multi-bandpass filter of Figure 6, together with typical emission spectra for red, green and blue LEDs; Figure 8 shows an image of a scene comprising a group of fluorescent marker pens of different colours, together with a colour palette of different coloured squares; Figure 9A shows an intensity image captured when illuminating the scene of Figure 8 with light in the red portion of the spectrum and detecting the reflected light with UV and IR wavelengths filtered out; Figure 9B shows the intensity image captured when illuminating the scene of Figure 8 with light in the red portion of the spectrum and detecting the reflected light through an emission filter according to an embodiment; Figure 9C shows the percentage difference between the intensity in the image of Figure 9A and the intensity in the image of Figure 9B, relative to the summed intensity in the two images. Figure 10A shows an intensity image captured when illuminating the scene of Figure 8 with light in the green portion of the spectrum and detecting the reflected light with UV and IR wavelengths filtered out; Figure 10B shows the intensity image captured when illuminating the scene of Figure 8 with light in the green portion of the spectrum and detecting the reflected light through an emission filter according to an embodiment; Figure 10C shows the percentage difference between the intensity in the image of Figure 10A and the intensity in the image of Figure 10B, relative to the summed intensity in the two images. Figure 11A shows an intensity image captured when illuminating the scene of Figure 8 with light in the blue portion of the spectrum and detecting the reflected light with UV and IR wavelengths filtered out; Figure 11B shows the intensity image captured when illuminating the scene of Figure 8 with light in the blue portion of the spectrum and detecting the reflected light through an emission filter according to an embodiment; Figure 11C shows the percentage difference between the intensity in the image of Figure 11A and the intensity in the image of Figure 11B, relative to the summed intensity in the two images. Figure 12 shows the same image as Figure 8, with a region of interest highlighted around the five marker pens. Figure 13A shows the normalized colour coordinates for each column of pixels in the region of interest highlighted in Figure 12, as obtained by combining the images shown in Figures 9A, 10A and 11A and after white balancing the colour channels; Figure 13B shows the normalized colour coordinates for each column of pixels in the region of interest highlighted in Figure 12, as obtained by combining the images shown in Figures 9B, 10B and 11B and after white balancing the colour channels. Figure 14A shows an example 2D colour reconstruction of an object captured using an imaging system according to an embodiment described herein. Figure 14B shows an example 2D colour reconstruction of an object as obtained when using a conventional imaging system with a colour camera including a Bayer filter. Figure 15A shows an image template comprising an array of repeating lines and different colours; Figure 15B shows an image of the template of Figure 15A as captured using a conventional imaging system employing a colour camera with Bayer filter. Figure 15C shows an image of the same template as Figure 15B, but captured using an imaging system according to an embodiment. DETAILED DESCRIPTION Embodiments described herein provide an imaging system capable of capturing and rendering 3D colour images of objects. A “3D colour image” can be understood to refer to a 3D colour representation of the surface(s) of an object, allowing the object to be viewed from different angles I perspectives. The surfaces may be generated from a point cloud comprising a collection of points in space, each one defining a point on the surface of the object and having its own set of (x, y, z) coordinates and an associated intensity value. Thus, the 3D image will include depth information in addition to the two-dimensional spatial information present in a conventional image or photograph. The colour information may be overlaid on the 3D image once rendered. Figure 3A shows an imaging system according to an embodiment. The system comprises a projector 301 and a camera 303. In this example, the imaging system is used to capture a 3D image of an object 305 using structured light illumination. The projector 301 is used to project a spatially varied 2D illumination pattern onto the object and the camera 303 used to acquire a 2D image of the illuminated object. As shown in Figure 3A, the illumination pattern may comprises a series of light and dark fringes 307, 309. Figure 3B shows a simplified diagram of the system geometry. The camera and projector are located a distance B apart. A point on the object that lies a distance D away from the camera and projector is located at an angle 9C from the camera and 9P from the projector 303. Owing to the angle between the camera and the projector, any variations in the surface topology of the object will cause the pattern of light and dark fringes, as detected by the camera, to become distorted. By corollary, the distortion in the pattern will encode information about the 3D surface of the object, and can be used to deduce its surface topology. The 3D information can be recovered by capturing sequential images in which the object is illuminated with different patterns of light, and comparing the measured intensity for each pixel across the sequence of images. In some embodiments, a phase shifting technique may be used to obtain the 3D information. Phase shifting is a well-known technique in which a sequence of sinusoidally modulated intensity patterns is projected onto the object, with each pattern being phase shifted with respect to the previous one. A 2D image of the illuminated object is captured each time the intensity pattern is changed. Variations in the surface topology of the object will give rise to a change in the phase of the intensity pattern as seen by the camera at different points across the surface. By comparing the intensities of light in the same pixel across the sequence of 2D images, it is possible to compute the phase at each point, and in turn use this to obtain depth information about the object. The data is output as a 2D array, in which each element maps to a respective one of the pixels of the camera, and defines the 3D spatial coordinates of a point as seen in that pixel. It will be appreciated that other techniques, besides phase shifting, may also be used to recover the 3D spatial information; for example, in some embodiments, a Graycoding technique may be used, or a combination of Gray-coding and phase shifting. The precise algorithms used to decode the 3D spatial information from the sequence of 2D images will vary depending on the specific illumination patterns and the way in which those patterns are varied across the sequence of images; further information on algorithms for recovering the depth information using these and other techniques is available in the publication “Structured light projection for accurate 3D shape determination” (0. Skotheim and F. Couweleers, ICEM12 - 12th International Conference on Experimental Mechanics 29 August - 2 September, 2004, Politecnico di Bari, Italy). In each case, the 3D spatial information in the object is computed by considering the variation in intensities at each point on the object as the illumination pattern changes and the points are exposed to light and dark regions of the pattern. Further details of approaches using a combination of Gray-coding and phase shifting can be found, for example, in an article by Giovanna Sansoni, Matteo Carocci and Roberto Rodella, entitled “Three-dimensional vision based on a combination of Graycode and phase-shift light protection: analysis and compensation of the systematic errors” - Applied Optics, 38, 6565-6573, 1999, and in US Patent No. US11763518B2, the entire contents of which are incorporated herein by reference. It will be clear that regardless of precisely which algorithm is used for the structured illumination, in order to compute the 3D spatial information with high accuracy, it is desirable to measure the variation in intensity at each point on the object with maximal signal to noise. It follows that it is desirable to capture as many of the photons reflected by the object as possible. At the same time, it is desirable to recover colour information from the object so as to add colour to the 3D rendering of the object. Figure 4 shows the imaging system of Figure 3 in more detail. In the present embodiment, the projector 301 includes a plurality of light sources, including a red LED 401, blue LED 403 and green LED 405. The light emitted by the LEDs is coupled into a common light path within the projector and directed via one or more lens and mirrors to a Digital Micromirror device DMD (not shown in Figure 3). The DMD comprises an array of pixel elements, each of which comprises a respective mirror. By adjusting the position of each mirror, it is possible to generate different patterns of illumination, such as the pattern of light and dark fringes seen in Figure 3. The pattern of light generated by the DMD is directed through one or more further lens and mirrors before being projected from the output 407 of the projector towards the object. The camera 303 itself comprises a monochrome sensor 409, such as a CCD or CMOS chip having an array of pixel elements. A lens 411 is used to focus light reflected by the object onto the sensor, and a processor 412 is used to process the images captured on the sensor 409. Using a monochrome camera means that light incident on any one of the pixels in the sensor array can be detected, regardless of its wavelength. This can be further understood with reference to Figure 5A, which shows a schematic of the pixel array in the camera 303, and Figure 5B, which shows a sample two-by-two cell within the pixel array. In comparison to the array shown in Figure 1, which includes the Bayer filter for colour discrimination, each one of the four pixel elements in the two-by-two cell of Figure 5B is responsive to light of blue (B), green (G) and red (R) wavelengths. Since each one of the elements in the sensor 409 is responsive to light in the blue, green and red parts of the spectrum, it is possible to obtain a high-resolution, high signal-to-noise image even when illuminating the object with a single wavelength or narrow band of wavelengths. For example, if illuminating the object with light from the blue LED alone, the (blue) light reflected from the object will be collected in each one of the pixel elements; this contrasts with the case of Figure 2 in which a Bayer colour filter is used in the camera, and where only a quarter of the pixel elements in the array will be responsive to the blue light reflected by the object. As discussed above, the projector 301 is configured to project a sequence of 2D illumination patterns on the object, with the camera 303 being used to capture an image of the object each time the illumination pattern changes. The captured images may be processed either in real-time or offline to produce a 3D reconstruction of the object as described in the above-cited publications. The 2D illumination patterns may be generated using light of a single wavelength or narrow band of wavelengths. For example, the 2D illumination patterns may be generated using light from a single one of the LED light sources, such as the blue LED. Although the light captured on the monochrome camera when projecting the different illumination patterns onto the object can be used to reconstruct a 3D image of the object, this signal alone will not allow for colour to be incorporated in the 3D image. In order to obtain the necessary colour information, a second set of images is acquired with the object being illuminated sequentially with light in the blue, red and green spectral bands. For example, the object may be illuminated sequentially by the blue LED, the green LED and the red LED. Here, rather than illuminating the object with a particular pattern of light and dark fringes, the whole field is illuminated uniformly by each colour in turn, with a respective “blue” image, “green image” and “red image” being captured on the monochrome camera (it will be appreciated that the blue, green and red images may be captured in any order, and indeed, these three sequential images may be captured either before or after the images used in the 3D reconstruction of the object are captured). When illuminating the object with the different wavelengths of light, the intensity of the signal detected on the camera sensor will depend on how reflective each part of the object is to light of the wavelength(s) in question. Thus, each one of the three images will encode information concerning the colour of the object. The three images can in turn be combined by the processor using known algorithms to determine the colour of each point on the object surface. The colour information can in turn be incorporated in the 3D image to render a color 3D image of the object. By means of the process described above, it is possible to obtain a 3D colour image of the object without the sacrifice seen in spatial resolution and photon count when using a Bayer filter to provide colour information. Since fewer photons are “wasted” when capturing the images required to perform the 3D reconstruction of the object, the signal-noise-ratio in the captured images will be higher, in turn allowing for a more accurate 3D rendering of the object to be achieved. It is further noted that the brightest modern LEDs operate in the blue part of the spectrum, hence it may be desirable to illuminate the object with light in that part of the spectrum. As discussed above, in conventional imaging systems that incorporate a Bayer filter, typically only one in four pixels will be covered by a blue filter, meaning that a significant amount of light will be lost when illuminating the object with blue light. In contrast, in the embodiments described herein, it is possible to image the object with blue light with no loss in spatial resolution, thereby making it possible to take full advantage of the higher power LEDs available in the blue part of the spectrum. Moreover, in contrast to conventional imaging systems that rely on a Bayer filter to provide colour information and which present issues in terms of spatial binning of pixels, embodiments described herein facilitate the use of binning of pixels (e.g. 2x2 or 4x4 binning) to even further increase sensitivity and / or speed of the image acquisition. Figure 6 shows an example of an imaging system 600 according to a further embodiment. Features that are the same as in the embodiment of Figure 4 are labelled with the same reference numerals. In addition to the components shown in Figure 4, the camera 303 of the imaging system 600 includes an emission filter 413 that is placed in the light path between the camera lens and the monochrome camera. The emission filter 413 is configured to filter light of certain wavelengths, whilst allowing light in the same spectral band as that being used to illuminate the object to pass through. Figure 7 shows the transmission profile of the emission filter 413 according to one embodiment. In this case, the emission filter is a multi-bandpass filter, having three pass bands 701, 703, 705 that coincide with the emission wavelengths of the blue, green and red LEDs in the projector. For purpose of explanation, typical emission spectra 707, 709, 711 of the respective blue, green and red LEDs are also shown in Figure 7. When illuminating the object with the blue LED, the emission filter will allow the blue reflected light to pass through, whilst blocking much of the ambient light at longer wavelengths. Similarly, when illuminating the object with the green LED, the emission filter will allow the green reflected light to pass through. Likewise, when illuminating the object with the red LED, the emission filter will allow the red reflected light to pass through. Since the object is only illuminated with light in one spectral band at a time, the emission filter 413 will not prevent the light that is being projected onto the object and reflected back to the camera from reaching the sensor array, nor will it limit the spatial resolution of the colour images obtained when illuminating the object with light in the different spectral bands. The emission filter 413 can help to further increase the signal-to-noise ratio obtained in the blue, green and red images, by blocking ambient light from reaching the camera sensor. In addition, the emission filter 413 can help to mitigate against spectral crosstalk seen in cases where the object being imaged is fluorescent. A fluorescent object will absorb photons at a shorter (e.g. blue) wavelength, and within the scale of picoseconds or nanoseconds, release some of the absorbed energy in the form of light at a longer wavelength. Since the photons emitted by the object will be at a longer wavelength than the “true” colour of the object (i.e. the colour of light reflected by the object), the fluorescence photons can create a false perception of the object’s colour when the separate colour images (blue, green and red) are processed and combined into a single colour image. As discussed above, the pass-bands in the emission filter 413 are chosen to coincide with the wavelength(s) of light used to illuminate the object. In each case, therefore, the pass bands will permit the passage of light reflected from the object, whilst substantially or entirely blocking the passage of any fluorescent photons emitted by the object. It will be appreciated that in the case of the multi-bandpass filter shown in Figure 7, the green and red passbands 703, 705 may still allow a small amount of fluorescence light to pass through when illuminating the object with light in the blue region of the spectrum. The extent to which this happens can be largely avoided by ensuring that the emission bands of the light sources in the projector are as narrow as possible, and that the passbands in the filter 413 are similarly narrow and coincide as much as possible with those emission bands. Alternatively, the multi-bandpass filter of Figure 7 may be replaced by a filter wheel, having separate red, green and blue filters. The filter wheel may be rotated such that when illuminating the object with the blue light source, the blue filter is positioned in the light path between the camera lens and the sensor array. Then, when switching to illuminate the object with the green light source, the filter wheel may be rotated such that the green filter is now positioned in the light path between the camera lens and sensor array; likewise, when illuminating the object with the red light source, the filter wheel may be rotated such that the red filter is now positioned in the light path between the camera lens and sensor array. The filter wheel can allow for use of filters with a single pass band, further reducing the noise seen in each colour image due to ambient light and / or fluorescence. The multibandpass filter, meanwhile, provides a single filter that can be fixed in place, without imposing any mechanical switching requirements on the imaging system. The multibandpass filter can, therefore, offer an advantage where fast image acquisitions on the order of a few milliseconds are required, and where the reduction in noise offered by the single passband filters in the filter wheel is offset by the longer switching time needed to rotate the wheel between each image capture. The impact of using the emission filter 413 can be demonstrated with reference to Figures 8 to 11. Figure 8 shows an image of a scene comprising a group of fluorescent marker pens of different colours, together with a colour palette of different coloured squares. Here, the numbers on the x and y axes denote column and row numbers, respectively. Figure 9A shows an intensity image captured when illuminating the scene with light in the red portion of the spectrum and detecting the reflected light with UV and IR wavelengths filtered out. Figure 9B shows the intensity image captured when illuminating the scene with light in the red portion of the spectrum and detecting the reflected light through the emission filter 413 having the pass bands shown in Figure 7. Figure 9C shows, for each pixel in the image, the percentage difference between the intensity in the image of Figure 9A and the intensity in the image of Figure 9B, relative to the summed intensity in the two images. It can be seen that, apart from a small number of edge effects towards the edge of the table on which the marker pens are positioned, the difference between the two images in Figures 9A and 9B is minimal (the edge effects arise from a slight variation in camera position between applying and removing the emission filter 413). Figure 10A shows an intensity image captured when illuminating the scene with light in the green portion of the spectrum and detecting the reflected light with UV and IR wavelengths filtered out. Figure 10B shows the intensity image captured when illuminating the scene with light in the green portion of the spectrum and detecting the reflected light through the emission filter 413 having the pass bands shown in Figure 7. Figure 10C shows, for each pixel in the image, the percentage difference between the intensity in the image of Figure 10A and the intensity in the image of Figure 10B, relative to the summed intensity in the two images. As in Figure 9C, it can be seen that, apart from a small number of edge effects towards the edge of the table on which the marker pens are positioned, the difference between the two images in Figures 10A and 10B is minimal. Figure 11A shows an intensity image captured when illuminating the scene with light in the blue portion of the spectrum and detecting the reflected light with UV and IR wavelengths filtered out. Figure 11B shows the intensity image captured when illuminating the scene with light in the blue portion of the spectrum and detecting the reflected light through the emission filter 413 having the pass bands shown in Figure 7. Figure 11C shows, for each pixel in the image, the percentage difference between the intensity in the image of Figure 11A and the intensity in the image of Figure 11B, relative to the summed intensity in the two images. In contrast to Figures 9C and 10C, Figure 11C shows that several of the fluorescent pens have different intensity values in Figures 11A and 11B. The difference in intensity between the two images 11A and 11B can be traced to spectral bleed-through caused by the fluorescence of the marker pens. In particular, whereas the emission filter 413 will block fluorescent light emitted by the marker pens, no such blocking occurs in the image captured in Figure 11 A; thus, the image captured in Figure 11A will include not only blue light reflected by the marker pens, but also green and yellow fluorescence emitted as a result of illuminating the pens with light in the blue portion of the spectrum. The results seen in Figures 9 to 11 are further explored in Figures 12 and 13. Figure 12 shows the same image as Figure 8, but with a region of interest 1201 highlighted around the five marker pens. Figure 13A shows the normalized colour coordinates for each column of pixels in the region of interest 1201, as obtained by combining the images shown in Figures 9A, 10A and 11A and after white balancing the colour channels; that is, Figure 13A shows, for each column, the relative red, green and blue light intensities as determined when the scene is illuminated by red, green and blue light in turn and imaged through a filter that blocks UV and IR wavelengths only. Figure 13B shows the normalized colour coordinates for each column of pixels in the region of interest 1201, as obtained by combining the images shown in Figures 9B, 10B and 11B and after white balancing the colour channels; that is, Figure 13B shows, for each column, the relative red, green and blue light intensities as determined when the scene is illuminated by red, green and blue light in turn and imaged through the emission filter 413. It can be seen that the results for the blue marker pen located in columns 825 - 925 of the image are broadly similar in both Figure 13A and 13B. However, in each one of the four other marker pens, the magnitude of the blue contribution is greater in Figure 13A than in Figure 13B; the heightened blue contribution in those regions of the image is an artefact arising from the fact that when illuminating the scene with blue light, the fluorescence seen at longer wavelength bands (green, yellow, red) is erroneously interpreted as being blue light reflected from the objects in the scene. That is, even though it is only the marker pen in columns 850 - 900 that is truly blue in colour, the other marker pens are erroneously perceived as reflecting blue light, because the filter used in Figure 11A does not distinguish between blue light that is being reflected by the marker pens and fluorescence emanating from those pens at longer wavelengths. The artefact is particularly prevalent in the case of the green marker pen in columns 675 -700, where the green contribution in Figure 13A is significantly diminished compared to that in Figure 13B, owing to the need to accommodate an (apparent) larger contribution from blue reflected light. The same artefact is also very noticeable for the yellow marker pen in columns 500 - 575, where the (apparent) larger contribution from the blue light in Figure 13A results in an overall grey appearance of the marker pen; this can be compared with the case in Figure 13B where the blue contribution in columns 500 - 575 is much reduced and the true yellow colour of the marker pen is clearer in light of the higher relative contributions from the red and green components. Accordingly, through use of the emission filter 413, it is possible to resolve the colours of fluorescent objects in the image scene with greater accuracy, even when illuminating the scene with light in the blue region of the spectrum. Figure 14A shows an example 2D colour reconstruction of an object captured using an imaging system according to an embodiment described herein. By way of comparison, Figure 14B shows an example 2D colour reconstruction of an object as obtained when using a conventional imaging system with a colour camera including a Bayer filter. Here, the lines 1401 indicate alternate pixel rows in the image and demonstrate that the barcode can be resolved on the scale of individual pixels. Comparing Figure 14A to Figure 14B, it can be seen that the embodiments described herein can achieve a higher spatial resolution, with a much clearer distinction in the individual bars in the barcode. Figures 15A, 15B and 15C show a further example of how by using a monochrome camera to capture sequential blue, green and red colour images of an object, and combining the information in those images, it is possible to obtain a single colour image of higher spatial resolution than is possible when using a colour camera with Bayer filter. In more detail, Figure 15A shows an image template comprising an array of repeating lines and different colours. Figure 15B shows an image of the template as captured using a conventional imaging system employing a colour camera with Bayer filter. Figure 15C shows an image of the same template captured using an imaging system according to an embodiment. It can be seen that in the case of the conventional algorithm, the demosaicking applied to the images captured on the colour camera results in an image of lower spatial resolution, as compared to the case when using an imaging system according to the embodiments described herein. More specifically, the low frequency ripples occurring in columns 115 to 175 are Moire (or aliasing) type artefacts related to the demosaicking algorithm. The monochrome camera as used in embodiments described herein captures all high frequency stripes except for the sub-pixel stripes in columns 0-115 and rows 15-70. The conventional system meanwhile fails to capture all striping in the image accurately. In the embodiments described above, each image in the second set of images is captured by illuminating the scene or object with a respective set of one or more wavelengths, with each set of wavelengths being confined to one of the blue, green and red regions of the spectrum. However, it will be appreciated that the wavelengths used to illuminate the object when capturing each image need not coincide entirely with one or other of the blue, green and red parts of the spectrum. For example, the first image in the second set of images may be captured by illuminating the object or scene with a set of wavelengths spanning both the blue and green regions of the spectrum, whilst the second image in the second set of images may be captured by illuminating the object or scene with a different set of wavelengths spanning the green and red regions of the spectrum. It will be further noted that the respective sets of wavelength may overlap to some degree; that is, the set of wavelengths used to capture the first image in the second set of images may include some wavelengths of light that are also present in the illumination used for capturing the next image in the second set of images. Provided that there is at least some variation between the wavelengths used to capture each one of the images in the second set of images, and an appropriate calibration is performed beforehand, it will be possible to determine the colour contribution in each pixel of the image, even where some illumination wavelengths are used for more than one image in the second set of images. Implementations of the subject matter and the operations described in this specification can be realized in digital electronic circuitry, or in computer software, firmware, or hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Implementations of the subject matter described in this specification can be realized using one or more computer programs, i.e., one or more modules of computer program instructions, encoded on computer storage medium for execution by, or to control the operation of, data processing apparatus. Alternatively or in addition, the program instructions can be encoded on an artificially generated propagated signal, eg., a machine-generated electrical, optical, or electromagnetic signal that is generated to encode information for transmission to suitable receiver apparatus for execution by a data processing apparatus. A computer storage medium can be, or be included in, a computer-readable storage device, a computer-readable storage substrate, a random or serial access memory array or device, or a combination of one or more of them. Moreover, while a computer storage medium is not a propagated signal, a computer storage medium can be a source or destination of computer program instructions encoded in an artificially generated propagated signal. The computer storage medium can also be, or be included in, one or more separate physical components or media (e.g., multiple CDs, disks, or other storage devices). While certain embodiments have been described, these embodiments have been presented by way of example only and are not intended to limit the scope of the invention. Indeed, the novel methods, devices and systems described herein may be embodied in a variety of forms; furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit and scope of the invention. Furthermore, it will be understood that features disclosed in relation to one embodiment can be combined with features disclosed in relation to another embodiment. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the spirit and scope of the invention.

Claims

1. An imaging system for generating a 3D colour image of an object, the system comprising:one or more light sources configured to illuminate the object with light of different wavelengths; anda monochrome camera comprising an array of pixels arranged to receive light reflected from the object when the object is illuminated by the one or more light sources, the system being configured to capture a first plurality of images on the monochrome camera when illuminating the object with a first set of wavelengths of light;the system being further configured to capture a second plurality of images, wherein for each one of the second plurality of images, the object is illuminated with a different respective set of wavelengths of light;the system further comprising a processor configured to:generate, based on the first plurality of captured images, a 3D image of the object; anddetermine, based on the second plurality of images, a colour value for each point in the 3D image.

2. An imaging system according to claim 1, further comprising one or more emission filters, wherein when illuminating the object with a respective set of wavelengths of light, the one or more emission filters are arranged to allow the wavelengths of light in the respective set of wavelengths to pass through towards the monochrome camera.

3. An imaging system according to claim 2, wherein the one or more emission filters comprise a single emission filter having a plurality of pass bands, each pass band coinciding with a respective set of wavelengths used to illuminate the object.

4. An imaging system according to claim 2, wherein the one or more emission filters comprise a plurality of emission filters insertable into the light path between the object and the camera,wherein for each image captured by the system, the emission filter selected for insertion into the light path is one whose pass band coincides with the set of wavelengths being used to illuminate the object at the time of capturing the image.

5. An imaging system according to any one of claims 2 to 4, wherein the one or moreemission filters are arranged such that when illuminating the object with light in the first set of wavelengths, the light is able to pass through the filter towards each one of the pixels in the array.

6. An imaging system according to any one of the preceding claims, wherein the imaging system is configured to capture the second plurality of images on the monochrome camera.

7. An imaging system according to any one of the preceding claims, wherein for each one of the first plurality of images, the system is configured to illuminate the object with a spatially varied 2D illumination pattern, the pattern being altered for each one of the first plurality of images;wherein the processor is configured to generate the 3D image of the object by determining, for each pixel in the array, a variation in intensity of light incident on the pixel across the first plurality of images.

8. An imaging system according to any one of the preceding claims, wherein the one or more light sources are configured to sequentially illuminate the object with light in a blue spectral band, green spectral band, and red spectral band.

9. An imaging system according to any one of the preceding claims, wherein the one or more light sources comprises one or more light emitting diodes, LEDs.

10. A method of generating a 3D colour image of an object, the method comprising: capturing, on a monochrome camera, a first plurality of images of the object when illuminating the object with light having a first set of wavelengths;capturing a second plurality of images, wherein for each one of the second plurality of images, the object is illuminated with a different set of wavelengths;generating, based on the first plurality of captured images, a 3D image of the object; anddetermining, based on the signal received in each one of the second plurality of images, a colour value for each point in the 3D image.20

Images