A single-pixel ultraviolet-visible-infrared wide spectral range multispectral imaging system and imaging method

Abstract

The application discloses a kind of single-pixel ultraviolet-visible-infrared wide spectral range multispectral imaging system and imaging method, the system includes: light source module, modulation device, lens group, multispectral detection module, multichannel data acquisition module and data processing module;The present application is introduced multiple narrow-band detectors with different spectral responses in multispectral detection module, avoid increasing additional light splitting device and filter device, and the field of view, viewing angle, pixel number and other parameters of the image obtained using different narrow-band detectors are exactly the same, there is no image registration problem, convenient for data analysis and image fusion;In addition, the size of narrow-band detector is not limited basically, can reach centimeter level, the imaging capability is improved under weak light condition;The multispectral imaging system of the application has the characteristics of single-pixel imaging, large number of spectra, wide spectral range, fast imaging speed, good space-time resolution and small volume.

Description

Technical Field

[0001] This invention belongs to the field of multispectral imaging technology, and particularly relates to a single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system and imaging method. Background Technology

[0002] Traditional optical imaging systems use area-array silicon photodetectors (such as CCDs and CMOS sensors). These array detectors typically have over a million pixels, and fabricating high-resolution multispectral images is often complex and expensive. Furthermore, the size of individual image sensor pixels in array detectors is usually very small, resulting in poor imaging capabilities in low-light conditions, especially in multispectral imaging where only a portion of the energy within a narrow spectral range can be utilized. Single-point detectors can reach centimeter-scale sizes, significantly improving imaging capabilities in low-light conditions, but this improvement still has limitations. In extremely low-light or long-distance imaging scenarios, even large single-point detectors may face insufficient light collection efficiency. The light-collecting area of ​​a single detector is still limited relative to the entire imaging scene, restricting its performance in capturing weak signals. Traditional image sensors do not possess spectral resolution capabilities. To achieve spectral resolution, narrowband single-wavelength array detectors can be fabricated, or methods such as beam splitting and filtering of composite light can be employed. For narrowband single-wavelength array detectors, specific materials are required to achieve the desired wavelength selectivity. Silicon detectors, due to their large-scale production and mature technology, have seen a significant reduction in cost. However, achieving better spectral resolution often requires very expensive fabrication of narrowband single-wavelength detectors or array detectors based on other materials. Existing technologies typically use spectroscopic devices such as prisms, gratings, and interferometers, which increase system complexity and impact overall system size. Commonly used filters include rotary filters, linearly graded filters, and acousto-optic tunable filters; however, using these filters reduces the number of available spectra and requires longer switching wavelengths, thus slowing down imaging. Due to the bandgap differences among various semiconductor materials, each material has its specific optimal wavelength response range. For example, silicon's response range is approximately 300-1100 nm, while arsenic's is approximately 900-2000 nm. Therefore, to achieve multispectral imaging with a wide spectral range, multiple multispectral cameras must be used to photograph the same object from different angles. However, this method is not only bulky, but the differences in pixel count and shooting angles between different cameras can lead to registration problems in the synthesized image.

[0003] Tian Jie et al. (Hu Z, Fang C, Li B, et al. First-in-human liver-tumour surgery guided by multispectral fluorescence imaging in the visible and near-infrared I / II windows[J]. Nature Biomedical Engineering, 2020, 4(3): 1-13. DOI: 10.1038 / s41551-019-0494-0.) proposed a first-in-human liver tumor surgery guided by multispectral fluorescence imaging in the visible and near-infrared I / II windows. The article used three cameras: a visible light camera, a near-infrared I-zone camera, and a near-infrared II-zone camera. Since the number of pixels and shooting angles of different cameras may be different, there is an image registration problem, which brings difficulties to image fusion and image analysis. Summary of the Invention

[0004] To overcome the shortcomings of the prior art, the present invention aims to provide a single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system and method. By introducing multiple narrowband detectors with different spectral responses, the present invention avoids the need for additional spectroscopic devices and filters compared to traditional multispectral imagers. Furthermore, images obtained using different narrowband detectors have identical parameters such as field of view, viewing angle, and pixel count, eliminating image registration issues and facilitating data analysis and image fusion. In addition, the size of the narrowband detectors is largely unrestricted, reaching the centimeter level, which enhances imaging capabilities under low-light conditions. The multispectral imaging system of the present invention features single-pixel imaging, a large number of spectra, a wide spectral range, fast imaging speed, good spatiotemporal resolution, and small size.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system includes:

[0007] Light source module 1: used to emit uniform light, and the uniform light output terminal of light source module 1 is connected to the uniform light input terminal of modulation device 2;

[0008] Modulation device 2: Modulates the uniformly distributed outgoing light according to the loaded modulation matrix to convert it into structured light. The structured light emitted by the modulation device 2 is focused onto the imaging object 4 by the lens group 3.

[0009] Lens group 3: used to focus structured light onto imaging object 4, and the structured light reflected by imaging object 4 is focused onto multispectral detection module 5;

[0010] Multispectral detection module 5: Used to detect optical signals and convert them into photoelectric signals covering the ultraviolet-visible-infrared bands;

[0011] Multi-channel data acquisition module 7: used to acquire photoelectric signals. The photoelectric signal output terminal of the multi-channel data acquisition module 7 is connected to the photoelectric signal input terminal of the data processing module 8.

[0012] Data processing module 8: Used to control the modulation process of modulation device 2 and the sampling rate of multi-channel data acquisition module 7 when acquiring photoelectric signals, and to process the photoelectric signals acquired by multi-channel data acquisition module 7 to obtain a multispectral image with a wide ultraviolet-visible-infrared spectral range.

[0013] The uniform light in the light source module 1 is natural light or broadband light, and the wavelength coverage of broadband light is from ultraviolet to infrared.

[0014] The modulation device 2 is a digital micromirror array modulator (DMD).

[0015] The modulation matrix loaded on the modulation device 2 is a random speckle matrix, a Hadamard matrix, or a Fourier matrix.

[0016] The lens group 3 includes a condenser and a reflector. The structured light is focused by the condenser onto the reflector and then reflected onto the imaging object 4.

[0017] The multispectral detection module 5 has a barrel-shaped structure, including a barrel-shaped shell 10. Inside the barrel-shaped shell 10, there are multiple narrowband detectors 9 with different spectral responses. The multiple narrowband detectors 9 are arranged in concentric circles to form multiple detection rings, and each detection ring corresponds to a specific spectral range. The number of narrowband detectors 9 in the multispectral detection module 5 corresponds one-to-one with the number of channels in the multichannel data acquisition module 7.

[0018] The multispectral imaging system also includes an amplification module 6, which amplifies the photoelectric signal output by the multispectral detection module 5. The photoelectric signal output terminal of the amplification module 6 is connected to the photoelectric signal input terminal of the multichannel data acquisition device 7.

[0019] The data processing module 8 uses a single-pixel restoration algorithm to reconstruct the photoelectric signals acquired by the multi-channel data acquisition module 7 into a multispectral image; it then uses an image overlay algorithm or a signal overlay algorithm to overlay and synthesize the multispectral image; finally, it uses an image fusion algorithm to synthesize a multispectral image with a wide ultraviolet-visible-infrared spectral range.

[0020] A single-pixel multispectral imaging method covering a wide ultraviolet-visible-infrared spectral range includes:

[0021] Step 1: Obtain uniformly distributed emitted light through light source module 1;

[0022] Step 2: Determine the modulation matrix type, and modulate the uniformly distributed outgoing light by adjusting the determined modulation matrix pattern to convert it into structured light, which is then focused onto the multispectral detection module 5 to detect the photoelectric signal;

[0023] Step 3: The photoelectric signal detected in Step 2 is directly acquired by the multi-channel data acquisition module 7 and stored in the data processing module 8; or, the photoelectric signal detected in Step 2 is amplified by the amplification module 6, and then acquired by the multi-channel data acquisition module 7 and stored in the data processing module 8.

[0024] Step 4: The photoelectric signal stored in the data processing module 8 in step 3 is used to reconstruct the image using a single-pixel restoration algorithm to restore it into a multispectral image; then, the multispectral image is superimposed and synthesized using an image superposition algorithm or a signal superposition algorithm; finally, the image fusion algorithm is used to fuse the images and synthesize a multispectral image covering a wide spectral range of ultraviolet-visible-infrared.

[0025] In step 2, by adjusting the number of playback frames, playback rate, and number of sampling points of the modulation matrix pattern, the modulation device 2 is controlled to change the intensity distribution of the uniformly distributed emitted light, thereby obtaining structured light.

[0026] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0027] 1. Compared with existing high-density array detectors with complex processes, this invention employs a multispectral detection module with a barrel structure composed of multiple narrowband detectors. This allows each narrowband detector to capture light within a specific wavelength range, significantly improving the imaging capability of the imaging system under low-light conditions. Furthermore, the barrel structure increases the light-collecting area, further enhancing the sensitivity of the narrowband detectors to low light, thus significantly improving the performance of the entire imaging system when capturing weak signals. Using narrowband detectors avoids the need for additional spectrometers or filters, simplifying the system's complexity, reducing its size, and increasing imaging speed. Simultaneously, multiple narrowband detectors can cover multiple different spectral ranges, allowing for the simultaneous acquisition of information from multiple spectral bands, increasing the number and range of spectra.

[0028] 2. This invention generates optical signals by setting multiple narrowband detectors with different spectral responses in the multispectral detection module and converting the optical signals into photoelectric signals covering the ultraviolet-visible-infrared bands. At the same time, the data processing module uses a single-pixel restoration algorithm to reconstruct the photoelectric signals to obtain multispectral images. This avoids the need to add additional spectrometers and filters, and solves the technical problems of large system size, slow imaging speed and small number of spectra. Moreover, in single-pixel imaging, the parameters such as field of view, viewing angle and number of pixels of images obtained using different narrowband detectors are exactly the same, there is no image registration problem, which facilitates data analysis and image fusion.

[0029] 3. This invention enhances the intensity of photoelectric signals by employing an amplification module, ensuring that information is not lost due to signal attenuation during transmission, which helps the multi-channel data acquisition module acquire photoelectric signals. In addition, the amplification module significantly improves the signal-to-noise ratio of the system, ensuring the clarity and reliability of photoelectric signals, thereby optimizing the overall system performance.

[0030] In summary, compared with traditional multispectral imagers, this invention, by introducing multiple narrowband detectors with different spectral responses, avoids the need for additional spectroscopic and filtering devices. Moreover, images obtained using different narrowband detectors have identical parameters such as field of view, viewing angle, and pixel count, eliminating image registration issues and facilitating data analysis and image fusion. Furthermore, the size of the narrowband detectors is virtually unlimited, reaching the centimeter level, which enhances imaging capabilities under low-light conditions. The multispectral imaging system of this invention features single-pixel imaging, a large number of spectra, a wide spectral range, fast imaging speed, good spatiotemporal resolution, and small size. Attached Figure Description

[0031] Figure 1 This is a schematic diagram of the structure of the single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system provided by the present invention.

[0032] Figure 2 This is a schematic diagram of the structure of the barrel-shaped multispectral detection module provided by the present invention.

[0033] In the figure: 1-Light source module, 2-Modulation device, 3-Lens group, 4-Imaging object, 5-Multispectral detection module, 6-Magnification module, 7-Multi-channel data acquisition module, 8-Data processing module, 9-Narrowband detector, 10-Barrel-shaped housing.

[0034] Figure 3 The flowchart of the CNN multispectral image fusion model provided by the present invention.

[0035] Figure 4 The multispectral image provided by this invention before image fusion; wherein, Figure 4(a) has a wavelength of 433.8 nm. Figure 4 (b) has a wavelength of 515.2 nm. Figure 4 (c) has a wavelength of 539.4 nm. Figure 4 (d) has a wavelength of 638.3 nm. Detailed Implementation

[0036] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings.

[0037] like Figure 1 As shown, a single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system includes:

[0038] Light source module 1: used to emit uniform light, and the uniform light output terminal of light source module 1 is connected to the uniform light input terminal of modulation device 2;

[0039] The uniform light is natural light or broadband light; a xenon lamp is used to emit broadband light; the xenon lamp is a Zolihan GLORIA-X450A, and its wavelength coverage range is from ultraviolet to infrared.

[0040] Modulation device 2: Modulates the uniformly distributed outgoing light according to the loaded modulation matrix to convert it into structured light. The structured light emitted by the modulation device 2 is focused onto the imaging object 4 by the lens group 3.

[0041] The modulation device 2 is a digital micromirror array modulator (DMD), which is composed of tens of thousands of independently flippable micromirrors. Each micromirror can be flipped ±12° on the array plane. The model of the digital micromirror array modulator (DMD) is the V-6501VIS spatial light modulator from ViALUX, Germany. It is composed of 1920×1080 digital micromirrors of 7.56μm size, which can achieve light modulation in the range of 400-2200nm.

[0042] The modulation matrix loaded on the modulation device 2 is a random speckle matrix, a Hadamard matrix, or a Fourier matrix;

[0043] In this embodiment, the modulation matrix is ​​a Fourier matrix;

[0044] Lens group 3: used to focus structured light onto imaging object 4, including a condenser and a reflector. The structured light is focused by the condenser onto the reflector and then reflected onto imaging object 4. The structured light reflected by imaging object 4 is focused onto multispectral detection module 5.

[0045] The condenser lens is an LBTEK MAC4202; the reflector is a custom-made LBTEK model with a diameter of 25mm, a focal length of 100mm, and a wavelength coverage range of 350nm-2.1μm.

[0046] Multispectral detection module 5: used to detect optical signals and convert them into photoelectric signals covering the ultraviolet-visible-infrared bands. The photoelectric signal output terminal of the multispectral detection module 5 is connected to the photoelectric signal input terminal of the amplification module 6.

[0047] like Figure 2 As shown, the multispectral detection module 5 has a barrel-shaped structure, including a barrel-shaped housing 10. Multiple narrowband detectors 9 with different spectral responses are disposed inside the barrel-shaped housing 10. The multiple narrowband detectors 9 are arranged in concentric circles to form multiple detection rings, each corresponding to a specific spectral range. The barrel-shaped structure allows each narrowband detector 9 to capture light only within a specific wavelength range, significantly improving the imaging capability of the imaging system under low-light conditions. Furthermore, the barrel-shaped structure increases the light-collecting area, further enhancing the sensitivity of the narrowband detectors 9 to low light, thus significantly improving the performance of the entire imaging system when capturing weak signals.

[0048] Amplification module 6: Used to amplify the photoelectric signal output by the multispectral detection module 5. The photoelectric signal output terminal of the amplification module 6 is connected to the photoelectric signal input terminal of the multichannel data acquisition device 7.

[0049] The amplification module 6 uses an OPA1611 operational amplifier;

[0050] Multi-channel data acquisition module 7: used to acquire photoelectric signals. The photoelectric signal output terminal of the multi-channel data acquisition module 7 is connected to the photoelectric signal input terminal of the data processing module 8. The number of narrowband detectors 9 in the multispectral detection module 5 corresponds one-to-one with the number of channels in the multi-channel data acquisition module 7.

[0051] The multi-channel data acquisition module 7 used adopts a multi-channel data acquisition card, model NI 6361, with a maximum single-ended analog input channel count of 16; the use of a multi-channel acquisition card ensures that multiple narrowband detectors can work simultaneously, while also meeting the requirements for imaging speed.

[0052] Data processing module 8: used to control the modulation process of modulation device 2 and the sampling rate of multi-channel data acquisition module 7 when acquiring photoelectric signals, and to process the photoelectric signals acquired by multi-channel data acquisition module 7 to obtain a multispectral image with a wide ultraviolet-visible-infrared spectral range; the data processing module 8 is a computer;

[0053] The processing procedure is as follows:

[0054] The photoelectric signals acquired by the multi-channel data acquisition module 7 are reconstructed into a multispectral image using a single-pixel restoration algorithm. The multispectral image is then superimposed and synthesized using an image superposition algorithm or a signal superposition algorithm to eliminate the shadows caused by the angle limitation of the narrowband detector 9. Finally, an image fusion algorithm is used to synthesize a multispectral image with a wide ultraviolet-visible-infrared spectral range.

[0055] A single-pixel multispectral imaging method covering a wide ultraviolet-visible-infrared spectral range includes:

[0056] Step 1: Obtain uniformly distributed emitted light through light source module 1;

[0057] Step 2: Determine the modulation matrix type, and modulate the uniformly distributed outgoing light by adjusting the determined modulation matrix pattern to convert it into structured light, which is then focused onto the multispectral detection module 5 to detect the photoelectric signal;

[0058] By adjusting the number of playback frames, playback rate, and number of sampling points of the modulation matrix pattern, the intensity distribution of the uniformly distributed emitted light is changed by controlling the modulation device 2 to obtain structured light;

[0059] In this embodiment, the number of modulation matrix patterns played is 10,000, the playback rate is 100 frames / second, and the sampling rate of the multi-channel data acquisition device 7 is 100,000 Hz, that is, 1,000 points are collected for each modulation matrix pattern.

[0060] Step 3: The photoelectric signal detected in Step 2 is directly acquired by the multi-channel data acquisition module 7 and stored in the data processing module 8; or, the photoelectric signal detected in Step 2 is amplified by the amplification module 6, and then acquired by the multi-channel data acquisition module 7 and stored in the data processing module 8.

[0061] Step 4: The photoelectric signal stored in the data processing module 8 in step 3 is used to reconstruct the image using a single-pixel restoration algorithm to restore it into a multispectral image; then, an image superposition algorithm or a signal superposition algorithm is used to superimpose and synthesize the multispectral image to eliminate the shadows caused by the angle limitation of the narrowband detector; finally, an image fusion algorithm is used to fuse the images and synthesize a multispectral image with a wide ultraviolet-visible-infrared spectral range.

[0062] The image reconstruction process is as follows:

[0063] In the spatial domain, the basic building blocks of an image are pixels; in the Fourier domain, they are Fourier coefficients. To obtain a specific Fourier coefficient C(fx, fy) in the Fourier spectrum of a multispectral image, the modulation device 2 projects four Fourier matrix patterns with spatial frequencies (fx, fy) and initial phases of 0, π / 2, π, and 3π / 2, respectively, onto the imaging object 4. The multispectral detection module 5 then obtains the corresponding four response values ​​D1, D2, D3, and D4. Finally, the required Fourier coefficient is calculated using the following formula:

[0064] C(fx,fy)=[D1(fx,fy)-D3(fx,fy)]+j[D2(fx,fy)-D4(fx,fy)]

[0065] Where x, y represent the rectangular coordinates of the plane containing the imaging object, and f x and f y These are the spatial frequencies corresponding to the x and y directions, respectively. C(fx,fy) represents the Fourier coefficients, and D1, D2, D3, and D4 represent the four response values ​​obtained by the Fourier matrix patterns with spatial frequencies (fx,fy) and initial phases of 0, π / 2, π, and 3π / 2 through the multispectral detection module 5. j represents the imaginary number.

[0066] Each Fourier coefficient C(fx,fy) can be obtained by projecting four Fourier matrix patterns with the same spatial frequency onto the imaging object, but with initial phases of 0, π / 2, π, and 3π / 2 respectively. By projecting four-step phase-shifted Fourier matrix patterns with different spatial frequencies onto the target object, the Fourier coefficients corresponding to different spatial frequencies in the Fourier spectrum of the imaging object image can be obtained. Performing an inverse Fourier transform on the obtained Fourier spectrum of the object image allows for the reconstruction of the multispectral image I(x,y), as shown in the following expression:

[0067] I(x,y)=F- 1 {C(fx,fy)}

[0068] =F- 1 {[D1(fx,fy)-D3(fx,fy)]+j[D2(fx,fy)-D4(fx,fy)]}

[0069] Among them, F- 1 {*} represents the inverse Fourier transform;

[0070] Multispectral images are superimposed and synthesized using either an image superposition algorithm or a signal superposition algorithm to eliminate shadows caused by the 9-angle limitation of the narrowband detector; the image superposition algorithm process is as follows:

[0071] Shadow regions are identified by analyzing the standard deviation and mean of the image; shadow regions typically have low reflectivity and high standard deviation. The signals obtained from individual narrowband detectors are superimposed with specific weights to enhance the features of non-shadow regions and suppress shadow regions. By comparing the superimposed signal with the original signal, shadow regions are identified and corrected, thereby effectively eliminating shadows and improving the quality of multispectral images.

[0072] The basic principle of signal superposition algorithms can be described by the following expression:

[0073] y(t)=x1(t)+x2(t)+x3(t)+x4(t)+...x n (t)

[0074] Where y(t) represents the superimposed output signal, x1(t), x2(t), x3(t)…x n (t) represents the output signal of a single narrowband detector, x n (t) consists of the response values ​​obtained by the multispectral detection module 5;

[0075] The image fusion process is as follows:

[0076] Image fusion methods are used to fuse shadowless multispectral images. A CNN-based multispectral image fusion model is adopted, such as... Figure 3 As shown, end-to-end feature extraction, feature fusion, and image reconstruction are achieved by designing the network structure and loss function, thus avoiding the tedious process of manually designing fusion rules. The network layer contains multiple convolutional layers, which are responsible for extracting features from the input multispectral images. The loss function is constructed by measuring the similarity between the input n multispectral images and the multispectral fused image to guide the network in end-to-end training. Finally, the fused ultraviolet-visible-infrared wide-spectral multispectral images are output, which are enhanced in both spatial resolution and spectral information.

[0077] Figure 4 The multispectral image provided by this invention before image fusion, such as Figure 4 As shown, a narrowband detector combined with a single-pixel imaging algorithm can achieve multispectral imaging. Figure 4 Different colored gourds respond differently to different wavelengths of light. Multispectral imaging can see things that color images (RGB, i.e., 3-spectrum imaging) cannot. It can simultaneously acquire image and spectral information of the object under test, and has the advantage of image and spectrum integration. It can play an important role in agriculture, industrial testing, medical and health industry, basic scientific research, cutting-edge technology exploration and military applications.

Claims

1. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system, characterized in that, include: Light source module (1): used to emit uniform light, the uniform light output end of the light source module (1) is connected to the uniform light input end of the modulation device (2); Modulation device (2): Modulates the uniformly distributed outgoing light according to the loaded modulation matrix to convert it into structured light. The structured light emitted by the modulation device (2) is focused onto the imaging object (4) by the lens group (3). Lens group (3): used to focus structured light onto the imaging object (4), and the structured light reflected by the imaging object (4) is focused onto the multispectral detection module (5); Multispectral detection module (5): used to detect light signals and convert them into photoelectric signals covering the ultraviolet-visible-infrared bands; the multispectral detection module (5) is a barrel-shaped structure, including a barrel-shaped shell (10), and multiple narrowband detectors (9) with different spectral responses are arranged inside the barrel-shaped shell (10). The multiple narrowband detectors (9) are arranged in a concentric circle manner to form multiple detection rings, and each detection ring corresponds to a specific spectral range; Multi-channel data acquisition module (7): used to acquire photoelectric signals. The photoelectric signal output terminal of the multi-channel data acquisition module (7) is connected to the photoelectric signal input terminal of the data processing module (8). Data processing module (8): Used to control the modulation process of the modulation device (2) and the sampling rate of the multi-channel data acquisition module (7) when acquiring photoelectric signals, and to process the photoelectric signals acquired by the multi-channel data acquisition module (7) to obtain a multispectral image with a wide ultraviolet-visible-infrared spectral range.

2. The single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1, characterized in that: The uniform light in the light source module (1) is natural light or broadband light, and the wavelength coverage of broadband light is from ultraviolet to infrared.

3. The single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1, characterized in that: The modulator (2) is a digital micromirror array modulator (DMD).

4. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1 or 3, characterized in that: The modulation matrix loaded on the modulation device (2) is a random speckle matrix, a Hadamard matrix, or a Fourier matrix.

5. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1, characterized in that: The lens group (3) includes a condenser and a reflector. The structured light is focused by the condenser and then reflected onto the imaging object (4).

6. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1, characterized in that: The number of narrowband detectors (9) in the multispectral detection module (5) corresponds one-to-one with the number of channels in the multichannel data acquisition module (7).

7. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1, characterized in that: The multispectral imaging system also includes an amplification module (6) for amplifying the photoelectric signal output by the multispectral detection module (5). The photoelectric signal output terminal of the amplification module (6) is connected to the photoelectric signal input terminal of the multichannel data acquisition device (7).

8. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging system according to claim 1, characterized in that: The data processing module (8) uses a single-pixel restoration algorithm to reconstruct the photoelectric signals acquired by the multi-channel data acquisition module (7) into a multispectral image; it uses an image overlay algorithm or a signal overlay algorithm to overlay and synthesize the multispectral image; and then it uses an image fusion algorithm to synthesize a multispectral image with a wide ultraviolet-visible-infrared spectral range.

9. A single-pixel ultraviolet-visible-infrared broadband multispectral imaging method, characterized in that, include: Step 1: Obtain uniformly distributed emitted light through the light source module (1); Step 2: Determine the modulation matrix type, and modulate the uniformly distributed outgoing light by adjusting the determined modulation matrix pattern to convert it into structured light, which is then focused onto the multispectral detection module (5) to detect the photoelectric signal; Step 3: The photoelectric signal detected in Step 2 is directly acquired by the multi-channel data acquisition module (7) and stored in the data processing module (8); or, the photoelectric signal detected in Step 2 is amplified by the amplification module (6), and then acquired by the multi-channel data acquisition module (7) and stored in the data processing module (8). Step 4: The photoelectric signal stored in the data processing module (8) in step 3 is used to reconstruct the image using a single-pixel restoration algorithm to restore it into a multispectral image; then the multispectral image is superimposed and synthesized using an image superposition algorithm or a signal superposition algorithm; finally, the image fusion algorithm is used to fuse the image and synthesize a multispectral image covering a wide spectrum of ultraviolet-visible-infrared.

10. A single-pixel ultraviolet-visible-infrared wide-spectral-range multispectral imaging method according to claim 9, characterized in that: In step 2, by adjusting the number of playback frames, playback rate, and number of sampling points of the modulation matrix pattern, the modulation device (2) is controlled to change the light intensity distribution of the uniformly distributed emitted light, thereby obtaining structured light.

Images