Electromagnetic Radiation Measuring Device

Abstract

An electromagnetic radiation measurement device is disclosed that includes an electromagnetic radiation sensor and a reconfigurable filter assembly. The reconfigurable filter assembly has a series of elements, each element of the series filtering a different wavelength band for measurement by the sensor. The wavelength bands of adjacent elements of the series overlap, thereby defining differential sub-bands. The device also includes a processor configured to derive measurements for the sub-bands utilizing sensor measurements for the series of different elements.

Description

[Technical field]

[0001] The present invention relates to electromagnetic radiation measuring devices and related methods. [Background technology]

[0002] Hyperspectral imaging can be used to image a scene and extract spectral details from the image. Traditionally, this can be done using a series of optical bandpass filters to selectively tune the wavelength bands being measured. Alternatively, point-by-point or line-by-line spectrometry techniques may be used, known as "whisk broom" or "push broom" scanners. There are also snapshot imagers that acquire the entire data cube in one measurement. However, these techniques are optically complex and costly, have combined uncertainties and measurement errors in the spatial and spectral dimensions, and typically have long scene acquisition times.

[0003] Conventional optical bandpass filter based hyperspectral imaging (HSI) works as follows: light from a scene contains photons of a mixture of wavelengths (spectral information), and an optical bandpass filter placed in front of the scene transmits a range of wavelengths depending on the bandwidth of the filter. The filter is changed to transmit a different wavelength band. Conventional tunable filters, including thin film dielectric filters, acousto-optic tunable filters and liquid crystal tunable filters, have a bandwidth of over 3 nm, which determines the spectral resolution achieved by this method.

[0004] It is also possible to use a grating monochromator as the spectrally dispersive system from which the spectral resolution is obtained, which can achieve high spectral resolution but at the expense of a significant loss of signal to noise ratio. Summary of the Invention [Problem to be solved by the invention]

[0005] Aspects of the present invention seek to provide improved electromagnetic radiation measurement devices and methods. [Means for solving the problem]

[0006] According to an aspect of the present invention there is provided an electromagnetic radiation measuring device as claimed in claim 1.

[0007] According to an aspect of the present invention there is provided an electromagnetic radiation measuring device comprising: An electromagnetic radiation sensor; a reconfigurable filter assembly having a series of configurations, each configuration in the series enabling transmission of a different wavelength band through the filter assembly to a sensor, the wavelength bands of adjacent configurations in the series overlapping to define differential sub-bands; a processor configured to derive measurements for the sub-bands using sensor measurements for a series of different configurations; Includes.

[0008] Optional features of the above aspects are set out in the dependent claims and / or detailed below.

[0009] Utilizing filter assembly configurations with overlapping wavelength bands allows preferred embodiments to extract a higher level of spectral detail than can be obtained with separate configurations. In particular, preferred embodiments allow measurements to be determined for differential sub-bands that are potentially much narrower than the wavelength bands possible for any individual configuration. This can provide high spectral resolution that can be useful, for example, in spectroscopy or imaging to detect a number of specific conditions that give rise to specific narrowband signals or signatures.

[0010] As will be understood, when the wavelength band of a structure is discussed herein, this refers to the wavelength band of electromagnetic radiation that the structure filters incident electromagnetic radiation, for example, by allowing that wavelength band to pass through the filter assembly, and by preventing wavelengths outside the wavelength band from passing through the filter assembly.

[0011] In a preferred embodiment, the wavelength bands of a series of adjacent structures overlap, thereby defining differential sub-bands within the wavelength bands outside the overlap region.

[0012] In a preferred embodiment, the wavelength bands of adjacent structures partially overlap each other.

[0013] In a preferred embodiment, measurements for the sub-bands are derived by processing sensor measurements associated with each of a series of components, taking into account overlap between the wavelength bands of the different components.

[0014] It should be noted that the terms "series" and "adjacent" above refer to wavelength intervals where adjacent configurations are configurations that have overlapping wavelength bands. These configurations may or may not be structurally adjacent within the filter assembly. Furthermore, stepping through the series of configurations involves stepping from one configuration to an adjacent configuration and monotonically increasing or decreasing the center wavelength of the wavelength band for which the filter assembly filters incident electromagnetic radiation. However, stepping through the series in sequence is not required in all embodiments when making measurements, and measurements can be made out of sequence in some embodiments.

[0015] In some embodiments, for each of a series of a plurality of constructs, In the wavelength band of the structure, a first derivative sub-band is additionally included and a second derivative sub-band is optionally excluded compared to the wavelength band of an adjacent structure; The processor is configured to derive a measurement value for the first differential sub-band.

[0016] The plurality of constituents can include all constituents, or all constituents except one or more reference constituents, each of the one or more reference constituents having a wavelength band that is associated only with a substantially spectrally invariant background signal and / or is substantially zero.

[0017] In some embodiments, the processor is configured to derive the measurement for the first derivative sub-band using a process that includes determining a difference between a sensor measurement associated with a wavelength band of the structure and a sensor measurement associated with a wavelength band of an adjacent structure, and optionally adjusting for a measurement associated with the second derivative sub-band.

[0018] In some embodiments, the reconfigurable filter assembly is configured such that sensor measurements associated with wavelength bands of the reference structures in the series are associated substantially only with a spectrally invariant background signal and / or are substantially zero.

[0019] In some embodiments, the processor is configured to input sensor measurements associated with the set of structures into a fitting process to determine measurements for sub-bands that match the sensor measurements, the fitting process taking into account details of the wavelength bands associated with the set of structures.

[0020] In a preferred embodiment, the reconfigurable filter assembly is configured to perform a measurement cycle that includes employing all configurations in a series of configurations, not necessarily in sequential order.

[0021] In some embodiments, the reconfigurable filter assembly is configured to cycle through a series of configurations.

[0022] In some embodiments, the reconfigurable filter assembly is continuously variable to form a series of configurations.

[0023] In this manner, the size of the differential sub-bands is variable and can be selected according to user criteria, enabling the device to provide a selectively adjustable level of spectral detail.

[0024] In some embodiments, the reconfigurable filter assembly is reconfigurable to change the optical interference effects of one or more filter media to form a series of configurations.

[0025] In some embodiments, the reconfigurable filter assembly is rotatable and / or slidable to change the optical interference effects of one or more filter media.

[0026] In some embodiments, the processor is configured to utilize sensor measurements associated with a first characteristic of the electromagnetic radiation for a set of different structures to derive measurements for a sub-band associated with the first characteristic.

[0027] In some embodiments, the first characteristic is selected from the group consisting of brightness, time dependence, polarization, spatial distribution, and phase.

[0028] In some embodiments, the processor may also be configured to utilize sensor measurements for the set of different configurations associated with a second characteristic of the electromagnetic radiation to derive measurements for the sub-band associated with the second characteristic, which may be selected from the same set as the first characteristic.

[0029] It should be noted that the second feature may be related to the dimension in which the measurements related to the first feature are derived for the sub-band. For example, the first feature may be time-dependent and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the time-dependent spatial distribution of the electromagnetic radiation. Similarly, the first feature may be polarization and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the spatial distribution of the polarization. In other embodiments, the first feature may be luminance and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the spatial distribution of the luminance. In other embodiments, the first feature may be phase and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the spatial distribution of the phase.

[0030] It is also noted that the processor may be configured to derive measurements for the sub-bands related to a third or further feature, as well as the first and second features. The third and / or further feature may be related to the dimension in which the measurements related to the first and second features are derived for the sub-bands. For example, the first feature may be polarization, the second feature may be time-dependent, and the third feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-bands related to the time-dependent spatial distribution of polarization.

[0031] In some embodiments, the device includes an electromagnetic radiation modulator configured to modulate electromagnetic radiation to obtain pulses of electromagnetic radiation incident on the target area, and a filter arrangement ( For each filter configuration, the sensor is configured to sense the time attenuation of electromagnetic radiation reflected, scattered or emitted from a target area and filtered by the filter assembly.

[0032] In some embodiments, the processor is configured to derive a measurement for the sub-band for each of the plurality of pixels.

[0033] In some embodiments the device is an imaging device and the sensor is an imaging sensor.

[0034] In some embodiments, the processor may be configured to perform the processing described in any of the above paragraphs for each pixel of the plurality of pixels.

[0035] According to an aspect of the present invention, there is provided a method as claimed in claim 15.

[0036] According to an aspect of the present invention, there is provided a method for measuring electromagnetic radiation, the method comprising: modifying elements of a reconfigurable filter assembly having a series of elements, each element of the series allowing a different wavelength band to be transmitted through the filter assembly, the wavelength bands of adjacent elements of the series overlapping, thereby defining differential sub-bands; measuring with a sensor the electromagnetic radiation that passes through the reconfigurable filter assembly for each of the series of configurations; deriving measurements for the sub-bands utilizing the sensor measurements; Includes.

[0037] Optional features of the above method aspects are set out in the dependent claims and / or detailed below.

[0038] In some embodiments, for each of a series of a plurality of constructs, In the wavelength band of the structure, a first derivative sub-band is additionally included and optionally a second derivative sub-band is excluded compared to the wavelength band of an adjacent structure; The method includes deriving a measurement value for a first differential sub-band.

[0039] In some embodiments, the method includes deriving a measurement for the first derivative sub-band using a process that includes determining a difference between a sensor measurement associated with a wavelength band of the structure and a sensor measurement associated with a wavelength band of an adjacent structure, and optionally adjusting for a measurement associated with a second derivative sub-band.

[0040] In some embodiments, the method includes inputting sensor measurements associated with the set of structures into a fitting process to determine measurements for sub-bands that match the sensor measurements, the fitting process incorporating details of wavelength bands associated with the set of structures.

[0041] In some embodiments, the method includes stepping sequentially through a series of constructs.

[0042] In some embodiments, the method includes measuring with a sensor a first characteristic of the electromagnetic radiation filtered by the filter assembly for a series of different configurations, and using the measurements of the first characteristic to derive measurements for a sub-band associated with the first characteristic.

[0043] In some embodiments, the first characteristic is selected from the group consisting of brightness, time dependence, polarization, spatial distribution, and phase.

[0044] In some embodiments, the method also includes measuring with a sensor a first characteristic of the electromagnetic radiation filtered by the filter assembly for a series of different configurations, and utilizing the measurement of the second characteristic to derive a measurement for a sub-band associated with the second characteristic. The second characteristic may be selected from the same group as the first characteristic. The first and second characteristics, and possibly further characteristics, may be as discussed above with respect to the system.

[0045] In some embodiments, the method includes modulating radiation incident on the target area into pulses; Measuring the electromagnetic radiation filtered by the reconfigurable filter assembly with a sensor includes measuring with the sensor the time attenuation of the electromagnetic radiation reflected, scattered or emitted from the target area and filtered by the filter assembly.

[0046] In some embodiments, the method includes deriving a measurement value for the sub-band for each of a plurality of pixels.

[0047] According to an aspect of the invention, there is provided an executable program configured to perform the above-mentioned method when executed on the above-mentioned device.

[0048] According to an aspect of the present invention, a computer readable medium encoding a program is provided.

[0049] The processor of the device can be programmed by a program.

[0050] Embodiments of the invention will now be described in detail, by way of example only, with reference to the accompanying drawings, in which: [Brief description of the drawings]

[0051] [Figure 1(a)] FIG. 1 illustrates conventional filter-based hyperspectral imaging. [Figure 1(b)] FIG. 1 illustrates conventional filter-based hyperspectral imaging. [Figure 1(c)] FIG. 1 illustrates conventional filter-based hyperspectral imaging. [Figure 1(d)] FIG. 1 illustrates conventional filter-based hyperspectral imaging.

[0052] [Diagram 2] FIG. 2 is a schematic diagram of a device according to an embodiment of the present invention.

[0053] [Figure 2A] FIG. 3 is another schematic diagram of the device of FIG. 2.

[0054] [Figure 3(a)] FIG. 3 illustrates differential sub-bands used in the embodiment of FIG. 2. [Figure 3(b)] FIG. 3 illustrates differential sub-bands used in the embodiment of FIG. 2. [Figure 3(c)] FIG. 3 illustrates differential sub-bands used in the embodiment of FIG. 2.

[0055] [Figure 4] FIG. 3 is a diagram illustrating an analysis method of the embodiment of FIG. 2. [Diagram 5] FIG. 3 is a diagram illustrating an analysis method of the embodiment of FIG. 2. [Figure 6] FIG. 3 is a diagram illustrating an analysis method of the embodiment of FIG. 2. [Figure 7] FIG. 3 is a diagram illustrating an analysis method of the embodiment of FIG. 2.

[0056] [Figure 8(a)] FIG. 2 shows real and ideal transmission functions. [Figure 8(b)] FIG. 2 shows real and ideal transmission functions.

[0057] [Figure 9] FIG. 3 illustrates a reconfigurable filter assembly of the embodiment of FIG. 2.

[0058] [Figure 10(e)] FIG. 1 illustrates an embodiment of the present invention. [Figure 10(f)] FIG. 1 illustrates an embodiment of the present invention. [Figure 10(g)] FIG. 1 illustrates an embodiment of the present invention. [Figure 10(h)] FIG. 1 illustrates an embodiment of the present invention.

[0059] [Figure 11]FIG. 1 is a schematic diagram of a time multiplexing embodiment.

[0060] [Figure 12] FIG. 13 is a schematic diagram of another embodiment that can be used for time multiplexing. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0061] FIG. 1 shows a conventional system for hyperspectral imaging that uses conventional filters.

[0062] Referring to FIG. 1, simply put, conventional filters sample the spectral region in a limited way and cannot faithfully reproduce all features.

[0063] The working principle of hyperspectral imaging based on conventional optical bandpass filters is shown in Figure 1(a)-(d). Light from a scene contains photons of a mixture of wavelengths (spectral information), and an optical bandpass filter placed in front of this scene transmits a range of wavelengths depending on the bandwidth of the filter. To transmit different wavelength bands, the filter itself must be replaced or tuned to transmit separate passbands.

[0064] Conventional tunable filters, including thin-film dielectric filters, acousto-optic tunable filters, and liquid crystal tunable filters, have bandwidths greater than 3 nm, which determines the spectral resolution achieved with this scheme. The spectral resolution of these systems is fixed by design and cannot be altered to accommodate measurements to scenes containing spectra finer than the resolution. The spectral measurement (c) with this type of filter, which is the output as a function of wavelength of detected radiation, results in a smoothed spectrum (d) because the spectral averaging is performed over a broad wavelength band limited by the design of the filter.

[0065] In contrast, as described below, in embodiments of the present invention, a reconfigurable filter assembly having a series of elements with overlapping wavelength bands is used, and measurements at differential sub-bands defined by the overlapping wavelength bands are derived to obtain a higher level of spectral detail than can be obtained by separate elements. In particular, in preferred embodiments, measurements at differential sub-bands that are potentially much narrower than the wavelength bands possible for any individual element can be determined. This can provide high spectral resolution, which can be useful, for example, in spectroscopy or imaging, to detect a number of specific conditions that give rise to specific narrowband signals or signatures.

[0066] Thus, embodiments of the present invention can provide hyperspectral imaging using differential reconstruction algorithms that can sample the spectral range without resolution limitations and faithfully reproduce spectral features without compromising signal-to-noise.

[0067] In particular, an embodiment of the present invention will be described with reference to FIG.

[0068] 2 shows an electromagnetic radiation measurement device 10 including an electromagnetic radiation sensor (the "sensor" in the claims) 12, a reconfigurable filter assembly 14, and a processor 16. In this embodiment, the device 10 is an imaging device and the electromagnetic radiation sensor 12 is an imaging sensor (spatially resolved detector), in this example a complementary metal oxide semiconductor (CMOS) sensor configured to detect electromagnetic radiation at each of a plurality of pixels, although other sensors, e.g., CCD, EMCCD, may be used in other embodiments.

[0069] Another schematic diagram of device 10 is shown in FIG. 2A.

[0070] As can be seen, device 10, in this embodiment, includes a continuous light source 50 (not shown in FIG. 2) that illuminates the target area. However, in some embodiments, the continuous light source may be ambient lighting in target area 30, and not necessarily specific lighting from device 10.

[0071] In this embodiment, the device 10 includes a collimating lens 32 (not shown in FIG. 2A ) between the target area 30 and the reconfigurable filter assembly 14. The target area 30 is of course not part of the device, but refers to an area that, in use, includes a scene, sample, or object that the device is configured to measure or image. The scene, sample, or object to be measured / imaged is generally one or more features that absorb and then re-emit incident light, or reflect or scatter incident light. The device 10 also includes an imaging lens 34 (not shown in FIG. 2A ) between the reconfigurable filter assembly 14 and the sensor 12. The lenses 32, 34 are configured such that light from the target area 30 passes along an optical axis 40 (not shown in FIG. 2A ) through the collimating lens 32, the reconfigurable filter assembly 14, and the imaging lens 34 to be incident on the sensor 12 at a suitable focus. However, as will be appreciated by those skilled in the art, various optical component configurations other than those shown in FIG. 2 may be implemented to focus light from target area 30 that has passed through reconfigurable filter assembly 14 onto sensor 12 to be filtered in accordance with the teachings of the present disclosure.

[0072] The reconfigurable filter assembly has a series of components, each component in the series allowing transmission of a different wavelength band through the filter assembly 14 to the sensor 12 for measurement by the sensor 12, and thus filtered up to that wavelength band. The wavelength bands of adjacent components in the series overlap, thereby defining differential sub-bands.

[0073] In this embodiment, the processor 16 is configured to control the reconfigurable filter device 14 and the sensor 12, as well as to receive data therefrom, for coordinating sensor measurements with the configuration of the reconfigurable filter device 14. The processor may be programmed with a suitable executable program to perform the functions described herein.

[0074] The processor 16 is configured to derive sub-band measurements for each of a plurality of pixels using the sensor measurements for a series of different configurations.

[0075] It should be noted that in this context, the terms "series" and "adjacent" refer to wavelength intervals where adjacent features are features that have overlapping wavelength bands. These features may or may not be structurally adjacent within the filter assembly 14. In other words, if one were to step through a series of features, one would step from one feature to an adjacent feature, monotonically increasing or decreasing the center wavelength of the wavelength band that the filter assembly 14 transmits. However, as will be explained below, it is not actually necessary in all embodiments to step through a series when making measurements.

[0076] FIG. 3 shows a schematic of how overlapping wavelength bands create differential sub-bands.

[0077] FIG. 3(a) shows band n 20 and band (n+1) 21. n can take any value such that band (n+1) 22 represents the wavelength band of any set of structures, and band n 20 represents the wavelength band of an adjacent structure with a lower center wavelength. Note that in FIG. 3, the wavelength increases in the left to right direction such that the boundary wavelengths of band 20 and band 22 are slightly offset from one another. The vertical direction of FIG. 3 is not significant.

[0078] In Fig. 3(b), band 20 and band 22 are each shown with a shaded area representing a derivative sub-band. In particular, band 22 includes an additional first derivative sub-band 24 and excludes a second derivative sub-band 26 compared to band 20, so that wavelength bands 20 and 22 overlap with each other. Fig. 3(c) shows the first and second derivative sub-bands and an overlap area 28, which represents wavelengths included in both band 20 and band 22. As can be seen, the derivative sub-bands 24, 26 of adjacent structures are outside the overlap area of ​​these structures.

[0079] In this embodiment, the second derivative sub-band 26 is excluded from band 22, although this is not necessary in all embodiments. In some embodiments, the wavelength bands can continually increase as the series progresses, with each band adding to the first derivative sub-band but not excluding any wavelengths.

[0080] Returning to FIG. 3, as discussed below, in this embodiment, for each pixel, measurements for the differential sub-bands are derived by processing sensor measurements associated with each of a set of components, taking into account the overlap between the wavelength bands of the different components.

[0081] In this embodiment, reconfigurable filter assembly 14 is configured to perform a measurement cycle that includes employing all of the configurations in the series. In particular, reconfigurable filter assembly 14 is configured to step sequentially through the series of configurations, although in other embodiments, the measurement cycle may employ other orders of configurations.

[0082] For each structure, the electromagnetic radiation sensor 12 is configured to obtain sensor measurements for each pixel and to pass the sensor measurements to the processor 16. The processor 16 is configured to utilize these sensor measurements to derive measurements for each differential sub-band of the plurality of pixels.

[0083] In this embodiment, the processor 16 is configured to derive measurements for the differential sub-bands using the analysis methods described below.

[0084] In this embodiment, the reconfigurable filter assembly 14 is configured such that the sensor measurements associated with the wavelength band of the first arrangement in the series are considered to be associated with a substantially spectrally invariant background signal, which in this example is substantially zero. In other words, the initial band serves as a background sampling band, and the associated first filter arrangement can be considered as a reference arrangement. As a result, there is a substantial absence of readings associated with the first arrangement, allowing the differential sub-bands introduced in the later arrangements in the series to be analytically determined.

[0085] In other embodiments, the reference construct (having a wavelength band considered to be associated with a substantially spectrally invariant signal) need not be the first construct, but can be any other construct in the series, which means that a certain offset will be present in the measured spectrum, but this offset may be acceptable depending on the application.

[0086] In some embodiments, there may be multiple reference constructs.

[0087] In this embodiment, all of the derivative sub-bands are of equal size, and the second derivative sub-band 26 of every structure in the series that does not overlap with the wavelength band of a reference structure is configured to correspond to the first derivative sub-band 24 of the previous structure in the series.

[0088] 3, for each of a series of a plurality of structures (in this embodiment, the plurality of structures is all structures except the reference structure) and for each pixel, the processor 16 is configured to derive a measurement for the first derivative sub-band using a process that includes determining the difference between a sensor measurement associated with a wavelength band of the structure (in this case, band 22) and a sensor measurement associated with a wavelength band of an adjacent structure (in this case, wavelength band 20), and, where appropriate, adjusted for a measurement associated with the second derivative sub-band 26. As will be apparent to those skilled in the art, taking the difference between band 22 and band 20 results in a value for the first derivative sub-band 24 reduced by the value of the second derivative sub-band 26. Thus, where appropriate, the calculation is adjusted for, in this case by adding, the measurement of the second derivative sub-band 26.

[0089] If the second derivative sub-band 26 is associated with a part of the wavelength band of the reference construct, its value is known, since the reference construct was associated only with a spectrally invariant background signal (zero in this example). If the second derivative sub-band 26 was not part of the wavelength band of the reference construct, its value would have been calculated beforehand as the first derivative sub-band of the previous construct in the series.

[0090] In other words, the first derivative subband 24 is the difference between Image 2 and Image 1, where Image 1 samples the background (i.e., approximately 0), this band represents the area under the spectral curve and the wavelength values ​​are obtained from the inverse Riemann sum.

[0091] This process is described in more detail below with reference to Figures 4-7.

[0092] FIG. 4 shows a number of differential sub-bands A1 to A6. A1 to A5 form a wavelength band I1 of a first configuration (reference configuration) series, in which the background light is sampled and in this embodiment gives a zero reading. A2 to A6 form a wavelength band I2 of a second configuration series. In other words I1=A1+A2+A3+A4+A5 I2=A2+A3+A4+A5+A6 Since A1=0, I2-I1=A6-A1=A6

[0093] FIG. 5 shows A6 in more detail. A6 is the first derivative sub-band of the second configuration, and A1 is the second derivative sub-band of the second configuration. As can be seen, A6 has a reading of interest that is not just a background signal. The spectral intensity S(λ) of A6 c ) can be calculated as follows:

[0094]

number

[0095] It turns out that knowing the intensity of any "strip" allows us to approximate its spectral intensity by a simple division.

[0096] Turning now to the differential sub-bands that do not overlap with the wavelength bands of the reference structures, and referring to FIG. 7, A6-A 10 form a wavelength band I6 of a sixth series of components, and A7 to A 11 form the wavelength band I7 of the seventh member of the series.

[0097] A 11 The spectral brightness S(λ c ) can be calculated using similar logic as for A6, except that the second derivative subband (in this case A6) is not zero. However, for the second construct in the series, its spectral intensity is derived as the first derivative subband.

[0098] Therefore, for Figure 7:

[0099]

number

[0100] The reconstruction relies on the difference between two adjacent structures, e.g. I7-I6, and by taking this difference, the overlap region between I7 and I6 is removed, leaving the leading sub-band from I7 (first derivative sub-band 24) with the negative and positive contributions from the trailing sub-band from I6 (second derivative sub-band 26). The reconstruction has two parts. In the first part of the reconstruction, any and all negative sub-bands (second derivative sub-band 26) are fixed to a background value (effectively zero in this embodiment) since these sub-bands are part of the wavelength band of the reference structure. This allows the leading sub-band (first derivative sub-band 24) to be uniquely determined. In the second part of the reconstruction, the negative contributions are no longer fixed to zero or background, but are respectively equal to the previously determined leading sub-band (first derivative sub-band 24). Now the negative contributions are added back into the subtraction so that the new leading sub-band can be uniquely determined.

[0101] Therefore, the spectral intensity S of the first derivative subband of any component i i (λ) can be calculated as follows: where index i refers to the i-th wavelength point used in the reconstruction.

[0102]

number

[0103] In this way, the spectral intensity may be determined for differential sub-bands potentially much narrower than the wavelength band transmitted by any one of the filter configurations, making it possible to determine much more desirable spectral detail than could be measured directly using the separate filters.

[0104] It can be seen that in this embodiment of the analytical model, the derivation of measurements for the differential sub-bands is performed in order of increasing wavelength, regardless of the order in which the reconfigurable filter assembly cycles through the different configurations (although it should be understood that in other embodiments the derivation can be performed in order of decreasing wavelength, in which case the direction of increasing wavelength in FIG. 3 would be reversed, such that the first or “leading” differential sub-band is at the low wavelength end of the wavelength band, and the second or “trailing” differential sub-band is at the high wavelength end of the wavelength band).

[0105] Furthermore, if the reference structure is not the first structure in the series, then for structures having a wavelength band of wavelengths larger than that of the reference structure, the derivation of the measurements for the differential sub-bands may be performed in order of increasing wavelength starting from the reference structure, whereas for structures having a wavelength band of wavelengths smaller than that of the reference structure, the derivation of the measurements for the differential sub-bands may be performed in order of decreasing wavelength starting from the reference structure.

[0106] In another embodiment of the present invention, which is the same as the embodiment described above except as described below, measurements for the differential sub-bands are derived using numerical methods rather than analytical methods.

[0107] An embodiment of the numerical method is described below.

[0108] Using numerical methods, the processor 16 is configured to input sensor measurements associated with the set of structures into a fitting process to determine measurements for the derivative sub-bands that match the sensor measurements. The fitting process incorporates details of the wavelength bands associated with each structure in the set of structures, such that for each structure in the set, measurements for the derivative sub-bands can be determined that provide the best match with the actual sensor measurements associated with the wavelength band.

[0109] Referring to Fig. 8, although real optical filters do not have ideal rectangular transmission bands, numerical fitting methods can still be applied to the reconstruction. Fig. 8(a) shows an ideal rectangular transmission band, while Fig. 8(b) shows an example of a real transmission band.

[0110] The reconfigurable filter assembly 14, as cycled through a series of elements, can be thought of as applying a matrix A corresponding to a set of i distinct transmission filters (a series of i different elements, where i is the number of elements in the series) to a measured vector x(λ), which corresponds to the optical spectrum (intensity as a function of wavelength) to be detected.

[0111] The output is a vector I(i) corresponding to the set of recorded light intensities for a sequence of i measurements, as measured by sensor 12 .

[0112] In other words, for each pixel, the luminance is measured for each of a series of transmission bands corresponding to different configurations.

[0113] This can be expressed as the following matrix equation:

[0114]

number

[0115] In other words, each of the structures has the following transmission function:

[0116]

number

[0117]

number

[0118]

number

[0119] Since the transmission spectrum (A) of a series of filter configurations is known and the output (in other words the sensor measurement for each of the configurations) (I) is measured, a fitting algorithm can be applied to approximate the measurand (x) by finding the measurand (x) that best fits the output (I) of the matrix equation above.

[0120] Standard fitting algorithms and the selection of starting vectors are suitable for this purpose.

[0121] In this way, the spectral intensity may be determined for differential sub-bands potentially much narrower than the wavelength band transmitted by any one of the filter configurations, making it possible to determine much more desirable spectral detail than could be measured directly using the separate filters.

[0122] It should be understood that the analytical and numerical methods described above are performed for each of a number of pixels of the sensor 12 to derive a differential sub-band spectral intensity for each.

[0123] Referring to details of the reconfigurable filter device embodiment of Figure 2, the reconfigurable filter device 14 is shown in detail in Figure 9. In this embodiment, the reconfigurable filter assembly comprises a plurality of filter media 36. The filter media 36 are configured such that the transmission function is dependent on the optical interference effects of the filter media, such that changing the angle of the filter media relative to the optical axis 40 changes the transmission function, thereby changing the transmitted wavelength band.

[0124] In this embodiment, as seen in Figure 9, the reconfigurable filter assembly 14 includes a first filter media 36a, a second filter media 36b, and a third filter media 36c having different principal transmission functions, although in other embodiments more or fewer filter media may be used. As seen in Figure 9, the first, second, and third filter media are disposed at angles to one another and are all mounted on a rotational axis 38 to allow the filter assembly 14, and in particular the filter media 36, ​​to rotate relative to an optical axis 40.

[0125] As the axis 38 rotates, the angle of each of the filter media 36a, 36b, 36c relative to the optical axis 40 changes, thereby changing the wavelength bands transmitted by the filter assembly 14. Thus, a series of different configurations may be formed with different rotation angles of the filter assembly 14. In other words, rotation of the axis 38, and thus the reconfigurable filter assembly 14, changes the optical interference effects in each of the filter media 36a, 36b, 36c, thereby forming a series of different configurations. The reconfigurable filter assembly 14 is then reconfigurable by rotation of the axis 38 to form a series of configurations.

[0126] Reconfigurable filter assembly 14 is mounted via axis 38 to a rotational stage / mount 42, which is mounted to a rotational base 44 configured to rotate the rotational stage / mount. Processor 16 is configured to control the movement of rotational stage 42, thereby rotating reconfigurable filter assembly 14 so that it adopts different configurations in a series of configurations.

[0127] As can be seen, this allows the reconfigurable filter assembly 14 to be continuously tunable to form a series of configurations. The size of the derivative sub-bands is variable and may be selected according to user criteria to enable the device to provide a selectively adjustable level of spectral detail. In this embodiment, the transmission bands may be precisely tuned in 1 nm steps to enable the size of the derivative sub-bands to be as small as about 1 nm or less.

[0128] As will be appreciated by those skilled in the art, because the filter media in this embodiment is a dichroic type filter media, the transmission function depends on the optical interference effects of the filter media, which change as the filter assembly 14 is rotated.

[0129] Reference is now made to Figures 10(e)-10(h). As mentioned above, the tunability of the filter device comes from changing the angle of incidence by rotating the filter around its own axis. This changes the wavelength dependence of the interference effects of the components, which in turn changes the transmitted wavelength band (f). While the absolute bandwidth of the filter at any angle is large (g), the relative differential bandwidth between the two filter positions can be made arbitrarily narrow. It is this arbitrarily narrow relative differential bandwidth over which the entire spectrum can be scanned and its features reconstructed with high fidelity (h).

[0130] However, while a particular filter arrangement has been described, in other embodiments, the reconfigurable filter assembly 14 can be replaced with any type of tunable filter assembly, provided that the filter assembly can employ multiple different configurations that allow filtering for different wavelength bands that overlap to define differential sub-bands as discussed in this disclosure. For example, in some embodiments, the reconfigurable filter assembly and / or the individual filter media can be slidable to change the optical interference effects of one or more filter media in addition to or instead of being rotatable. Furthermore, while the filter assemblies described in detail herein generally use transmission filters where light passes through the filter, some embodiments can use reflective filters where light is reflected and filtered. Indeed, any filter can be used, regardless of the spectral filtering mechanism, provided that it filters in accordance with the teachings disclosed herein. For example, in the analysis methods discussed with reference to Figures 4-7, the filter assembly configuration preferably provides a rectangular bandpass filter with step edges on both sides, or an edgepass filter with a step edge on one side.

[0131] 2 and 2A, an exemplary method of use of the device will now be described.

[0132] When it is desired to make a measurement of target area 30, sensor 12 is directed towards target area 30 via reconfigurable filter assembly 14 such that light coming from target area 30 passes through, is filtered by, and is incident on sensor 12. Reconfigurable filter assembly 14 is operated to perform a measurement cycle whereby the configuration of the reconfigurable filter assembly is changed to employ all configurations in the series of configurations as described above.

[0133] For each configuration employed, the sensor 12 performs a sensor measurement, thereby measuring the electromagnetic radiation that has been passed through and filtered by the reconfigurable filter assembly, in this embodiment, each sensor measurement is performed for each of a plurality of pixels of the multi-pixel image.

[0134] The sensor measurements are passed to a processor 16 which uses the analytical or numerical methods described above to derive, for each pixel, measurements for each of the differential sub-bands encompassed by the wavelength band of the set of elements.

[0135] In embodiments of the present invention, it is possible to use any monochrome camera sensor for sensor 12 (e.g., any existing CMOS or CCD sensor without a Bayer matrix) since the described embodiment turns it into a hyperspectral camera. The embodiments of the present invention can be thought of as changing a 2D (two spatial dimensions) camera into a 3D (two spatial + one spectral dimension) camera. In embodiments, spatial and spectral information can be obtained from a scene simultaneously.

[0136] Embodiments can provide tunable spectral decomposition and can be implemented as hyperspectral cameras as well as multispectral cameras.

[0137] It should be noted that the above embodiment does not rely on a dispersive element but uses a single high transmission optical element. This optical element is an optical bandpass filter, whose transmission band can be precisely tuned in 1 nm steps. The resolution issue can be addressed without sacrificing signal.

[0138] Embodiments of the present invention may provide the following advantages: Spectral resolution of 1nm (tunable) or better -New non-dispersive spectrum acquisition Low cost and low power consumption Small physical signature (e.g. mass) -Compatible with most existing imaging systems Simultaneous high optical throughput and high spectral and spatial resolution

[0139] The embodiments may be applied to: Quality inspection and control for manufacturing Enabling HSI on small satellites (e.g. CubeSats) Complementing data collection with panchromatic imagers on large satellites and airborne systems -UAV-based environmental monitoring Land use and crop health inspections Healthcare and medical imaging (ophthalmology, blood oxygenation, tissue diagnosis, etc.) Food processing monitoring and analysis Chemical and biological imaging ·Artwork fraud detection Forensic Analysis Security High-precision data input for AI algorithms Camera-based monitoring devices for open surgery Infrared gas detection

[0140] In the above embodiments, a spectrally fixed continuous light source is used. However, in some embodiments, a spectrally reconfigurable light source can be used. If the light source is spectrally reconfigurable, a spatio-spectral datacube can be obtained for each possible state of the light source without changing the reconstruction algorithm. The sequence is: 1. Change the light source configuration, 2. Pass through the filter assembly as discussed above, 3. Reconstruct the datacube, 4. Repeat.

[0141] It is also possible to use pulsed light, rather than continuous light, for example from a pulsed light source or by using a modulator, to increase the signal to noise ratio, as is known.

[0142] In the above-described embodiments, the device is an imaging device, the sensor is an imaging sensor, and the processor is configured to derive measurements related to the luminance of the differential sub-band for each of the plurality of pixels. In other words, the above-described method for deriving measurements of the differential sub-bands is performed for each of the plurality of pixels. However, this is not necessary in all embodiments. In some embodiments, the sensor / detector can be a single pixel detector. For example, in some embodiments, the device can operate like a spectrometer where only a single measurement is derived for each of the differential sub-bands.

[0143] In some embodiments, the teachings herein can be applied to any measurement system where the sensors / detectors perform some kind of multiplexed measurement of light. In the embodiment of Figure 2, spatial multiplexing is performed. However, in other embodiments, other multiplexing can be performed, including but not limited to spatial, temporal, and polarized (with an in-line polarized analyzer).

[0144] For example, the sensor 12 may measure a first characteristic of the light filtered by the filter assembly 14, and the processor 16 may be configured to utilize sensor measurements related to the first characteristic of the light for a series of different configurations to derive measurements related to the first characteristic for the sub-band. The first characteristic may be selected from the group consisting of brightness, time dependence, polarization, spatial distribution, and phase. The sensor 12 may also measure additional characteristics of the light filtered by the filter assembly 14, in some embodiments, selected from the same group as the first characteristic, and the processor 16 may be configured to utilize sensor measurements related to the additional characteristics of the light to derive measurements related to the additional characteristics for the sub-band.

[0145] It should be noted that further features may be related to the dimension in which the measurements related to the first feature are derived for the sub-band. For example, the first feature may be time-dependent and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the time-dependent spatial distribution of light. Similarly, the first feature may be polarization and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the spatial distribution of polarization. In other embodiments, the first feature may be luminance and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the spatial distribution of luminance. In other embodiments, the first feature may be phase and the second feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the spatial distribution of phase. In other embodiments, the first feature may be polarization, the second feature may be time-dependent, and the third feature may be a spatial distribution, whereby the processor is configured to derive measurements for the sub-band related to the time-dependent spatial distribution of polarization.

[0146] Please refer to Figure 11. This figure is a schematic diagram of a time multiplexing embodiment. This embodiment is the same as that of Figure 2, except that the sensor 12' is a time-resolved detector and a pulsed light source 50' is used to illuminate the target area 30.

[0147] The time-resolved detector 12' is configured to collect the intensity decay of all photons without separating them into spectral channels.

[0148] The pulsed light source 50' is a form of light modulator in that it is configured to modulate light to obtain pulses of light incident on the target area 30. The pulsed light source 50' is configured such that the pulses of light and the measurement of the sensor 12' are timed interdependently to measure the time decay of light reflected, scattered or emitted from the target area and filtered by the filter assembly, for example using heterodyne or homodyne detection in known manner.

[0149] The target area 30 may include an object or sample that absorbs and then re-emits incident light, or that reflects or scatters incident light.

[0150] For each element of the filter assembly, a set of decay curves is acquired by detector 12' (a 2D array of digital numbers containing the relative brightness and each time bin, each time bin representing the period over which the light is measured and having a length corresponding to the time resolution of the system). In other words, the acquired data is the transient response for each filter element.

[0151] The set of all attenuation curves for each and every filter structure is passed to a processor 16 which first reshapes the data to respond for each time bin for each filter structure, and then derives measurements for the sub-bands using the analysis algorithms disclosed herein. The processor then generates a spectrum for each time bin.

[0152] The data can then be analyzed as spectra per time bin or as decay curves per spectral channel.

[0153] Figure 12 is a schematic diagram of another embodiment that can be used for time multiplexing. The embodiment of Figure 12 is the same as the embodiment of Figure 11, except that a continuous light source is used in the embodiment of Figure 12 and an optical modulator is provided in the form of an optical chopper 52. The optical chopper is frequency phase locked with a lock-in amplifier 54 and configured to modulate the light from the continuous light source to form pulses of light that are incident on the target area 30, which allows measurement of the time decay of light reflected, scattered or emitted from the target area 30 and filtered by the filter assembly, for example using heterodyne or homodyne detection as discussed above.

[0154] In this embodiment, the detector / sensor 12 does not need to be time resolved and can be the same as in the embodiment of FIG.

[0155] The analysis is performed similarly to the embodiment of FIG.

[0156] The embodiments of Figures 11 and 12 have been described assuming a spectrally fixed light source. If a spectrally reconfigurable light source is used, the sequence is: 1. change the light source configuration, 2. change the filter configuration, 3. measure the signal in an appropriate time window to detect the time decay as discussed above, 4. repeat 2 and 3 until all filter configurations have been passed, 5. repeat 1 to 4 until all light source configurations have been passed.

[0157] In the embodiments of Figures 11 and 12, time-resolved detection is performed for each of the multiple pixels. In other words, the processor is configured to derive measurements for sub-bands related to the time-dependent spatial distribution of the light. However, as discussed above, in other embodiments, when the spatial distribution characteristics are not desired, the time-resolved detection can be performed using a single pixel sensor / detector and / or the time-dependent characteristics of the light can be combined with other characteristics of the light in the measurement.

[0158] In the above embodiments, the electromagnetic radiation is preferably light, although in embodiments it may encompass other regions of the electromagnetic spectrum in other embodiments.

[0159] All optional and preferred features and modifications of the described embodiments and the dependent claims can be used in all aspects of the invention taught herein. Moreover, the individual features of the dependent claims, and all optional and preferred features and modifications of the described embodiments are combinable and interchangeable with each other.

[0160] The disclosure of UK Patent No. 2118924.6, from which this application claims priority, is hereby incorporated by reference in its entirety.

Claims

1. An electromagnetic radiation measuring device, A sensor that measures electromagnetic radiation, A reconfigurable filter assembly having a series of components, each of which filters to a different wavelength band for measurement by the sensor, and the wavelength bands of adjacent components overlap, thereby defining a differential subband. A processor configured to derive measured values ​​for the differential subband using sensor measurements for the aforementioned series of different components, Electromagnetic radiation measuring devices, including those mentioned above.

2. For each of the aforementioned series of multiple components, In the wavelength band of the aforementioned component, compared to the wavelength band of an adjacent component, a first differential subband is added and a second differential subband is optionally excluded. The device according to claim 1, wherein the processor is configured to derive a measurement value for the first differential subband.

3. The device according to claim 2, wherein the processor is configured to derive a measurement for the first differential subband using a process that includes determining the difference between a sensor measurement related to the wavelength band of the component and a sensor measurement related to the wavelength band of the adjacent component, and optionally adjusting to a measurement related to the second differential subband.

4. The device according to claim 3, wherein the reconfigurable filter assembly is configured such that sensor measurements relating to the wavelength band of the reference configuration in the series relate substantially only to and / or substantially to a spectrally invariant background signal.

5. The device according to any one of claims 1 to 4, wherein the processor is configured to input sensor measurements associated with the series of components into a fitting process to determine measurements for the differential subband that fit the sensor measurements, and the fitting process incorporates details of the wavelength band associated with the series of components.

6. The device according to any one of claims 1 to 4, wherein the reconfigurable filter assembly is configured to proceed sequentially through the series of components.

7. The device according to any one of claims 1 to 4, wherein the reconfigurable filter assembly is continuously variable to form the series of components.

8. The device according to any one of claims 1 to 4, wherein the reconfigurable filter assembly is reconfigurable to form the series of components by changing the optical interference effect of one or more filter media.

9. The device according to claim 8, wherein the reconfigurable filter assembly is rotatable and / or slidable to change the optical interference effect of the one or more filter media.

10. The device according to any one of claims 1 to 4, wherein the processor is configured to derive a measurement for the differential subband relating to the first feature, using sensor measurements relating to a first feature of electromagnetic radiation for a series of different components.

11. The device according to claim 10, wherein the first feature is selected from the group consisting of luminance, time dependence, polarization, spatial distribution, and phase.

12. A device according to any one of claims 1 to 4, comprising an electromagnetic radiation modulator configured to modulate electromagnetic radiation to obtain pulses of electromagnetic radiation incident on a target region, wherein for each filter configuration, the sensor is configured to sense the time decay of electromagnetic radiation reflected, scattered or emitted from the target region and filtered by the filter assembly.

13. The device according to any one of claims 1 to 4, wherein the processor is configured to derive a measurement of the differential subband with respect to each of the plurality of pixels.

14. The device according to claim 13, wherein the device is an imaging device and the sensor is an imaging sensor.

15. A method for measuring electromagnetic radiation, A step of modifying the components of a reconfigurable filter assembly having a series of components, wherein each component in the series filters for a different wavelength band, the wavelength bands of adjacent components in the series overlap, thereby defining a differential subband. For each of the aforementioned series of components, the step of measuring the electromagnetic radiation filtered by the reconfigurable filter assembly with a sensor, A step of deriving a measurement value for the differential subband using the aforementioned sensor measurement value. Methods that include...

16. For each of the aforementioned series of multiple components, In the wavelength band of the aforementioned component, a first differential subband is added and included compared to the wavelength band of an adjacent component, and a second differential subband is optionally excluded. The method according to claim 15, further comprising the step of deriving a measurement value for the first differential subband.

17. The method according to claim 16, comprising the step of deriving a measurement for the first differential subband using a process that includes determining the difference between a sensor measurement related to the wavelength band of the component and a sensor measurement related to the wavelength band of the adjacent component, and optionally adjusting to a measurement related to the second differential subband.

18. A method according to any one of claims 15 to 17, comprising the step of inputting sensor measurements associated with the series of components into a fitting process to determine measurements for the differential subband that fit the sensor measurements, wherein the fitting process incorporates details of the wavelength band associated with the series of components.

19. The method according to any one of claims 15 to 17, comprising the step of proceeding sequentially through the series of components.

20. The method according to any one of claims 15 to 17, comprising the steps of: measuring a first feature of electromagnetic radiation filtered by the filter assembly with the sensor for a series of different components; and using the measured values ​​of the first feature to derive a measured value for the differential subband related to the first feature.

21. A method according to any one of claims 15 to 17, comprising the step of modulating radiation incident on a target region into pulses, A method comprising the step of measuring electromagnetic radiation filtered by the reconfigurable filter assembly with the sensor, the step of measuring the time decay of electromagnetic radiation reflected, scattered or emitted from the target region and filtered by the filter assembly with the sensor.

22. The method according to any one of claims 15 to 17, comprising the step of deriving a measurement for the differential subband with respect to each of a plurality of pixels.

23. An executable program configured to perform the method described in any one of claims 15 to 17 when executed on the device described in any one of claims 1 to 4.

Images