A high light flux multi-purpose imaging device

Abstract

The application provides a high-light-flux multi-purpose imaging device, aiming at solving the technical bottleneck of low light flux, multi-modal sampling exposure mismatch and light field three-dimensional perception failure of the existing snapshot multi-purpose imaging device. The device comprises a constant-aperture image-side telecentric lens and a light field camera body, and the light field camera body comprises a microlens array and a photosensitive element; the microlens unit is square, and the side length is N times (N>2 and is a positive integer) of the pixel side length, the corresponding areas of the two are coincident and the focal ratios are consistent; a tightly superimposed multi-dimensional array light information modulation element (N*N arrangement) and an anti-diffraction frame element are arranged at the lens diaphragm, the former realizes multi-modal information spatial multiplexing, and the latter blocks light cross leakage. The application adopts a wideband notch filter to improve the light flux, is compatible with the light field three-dimensional reconstruction function, realizes the integration of various imaging functions such as light field, multispectral and polarization, and has low processing difficulty, and can be widely applied in scientific research, agriculture, anti-drone and other fields.

Description

Technical Field

[0001] This invention relates to the field of image data acquisition and processing technology, and in particular to a high-throughput multi-purpose imaging device, especially a device that can simultaneously achieve high spatial resolution light field imaging, multispectral imaging, polarization imaging and high dynamic range imaging. Background Technology

[0002] In complex applications such as target recognition and tracking, to improve the accuracy of target recognition and the success rate of tracking, it is often necessary to simultaneously acquire multi-dimensional physical information of the target, including its light field, spectrum, polarization, and high dynamic range. Specifically: light field imaging mainly acquires the direction information of the target's radiated light rays, which can be used for depth estimation and 3D reconstruction of the target scene; multispectral imaging mainly extracts the spectral reflectance characteristics of the target, which is the core means of target material identification and classification; polarization imaging mainly acquires the polarization state of light rays, which can effectively highlight the target outline and suppress background stray light; and high dynamic range imaging acquires as much brightness detail information as possible about the target, which is used to achieve clear imaging in extreme scenes with strong backlight or large differences in illumination.

[0003] Traditionally, light field cameras, multispectral cameras, polarization cameras, and high dynamic range cameras are structurally independent, single-function devices. In practical engineering applications, using multiple different devices separately stitched together or cascaded in a collaborative manner presents extremely fatal drawbacks: First, the system is bulky and large, with high weight and power consumption, making it difficult to meet the requirements of lightweight and compact modern equipment; second, the inherent spatial parallax and temporal synchronization differences between independent devices at different physical locations and viewpoints make it extremely difficult to achieve accurate registration of the acquired multimodal images at the pixel level, causing the backend data fusion algorithm to often produce severe artifacts or even completely fail due to registration errors.

[0004] To overcome the drawbacks of separate multi-device stitching, the industry has developed single-camera snapshot-type multi-purpose imaging systems that integrate multiple functions into a single device. However, existing integrated systems of this type mainly rely on on-chip coating processes at the sensor pixel level (such as Bayer spectral arrays or micro-polarizer arrays), which also have significant physical limitations: First, there is significant light energy loss. Traditional pixel-level narrowband filters sacrifice spectral purity by physically blocking some wavelengths of light. The more spectral channels there are, the lower the overall light throughput of the system (usually less than 5%), making it unsuitable for low-light or high-speed detection tasks.

[0005] Secondly, there is the mutual constraint between spatial-spectral and polarization information. Different bandpass filters produce significantly different imaging brightness for targets with different properties in a scene, leading to brightness imbalance of signals in each channel during a single exposure and causing severe mosaic effects in the reconstructed image. Such systems face a physical bottleneck where "high spectral / polarization resolution, high spatial resolution, and high luminous flux" cannot be simultaneously achieved and are mutually constrained.

[0006] Third, the failure of light field spatial perception capability and a significant degradation in spatial resolution. The core advantage of light field cameras lies in using the parallax between images of different sub-apertures for 3D reconstruction (depth estimation) and relying on computational imaging techniques such as multi-view rendering fusion to restore and improve the spatial resolution of the image. However, as with the predicament faced by publicly disclosed patent solutions based on conventional light field imaging structures (such as CN111856770B and CN111866316B), when each sub-aperture is given narrowband filtering characteristics with huge differences in transmittance and response spectrum, the brightness and texture features of the same target under different sub-apertures will undergo drastic changes. This physical-level change not only makes the parallax matching and refocusing algorithms of traditional light field cameras completely ineffective, causing the device to essentially lose its original 3D spatial perception capability; more fatally, the failure of feature matching prevents the system from using multi-view feature fusion algorithms to compensate for and improve the spatial resolution of the image, resulting in a significant reduction in the resolution of the final output image, completely exposing and exacerbating the inherent "low spatial resolution" defect of conventional light field imaging structures.

[0007] Therefore, neither traditional multi-device discrete cascading nor existing on-chip coating snapshot systems can perfectly achieve high-quality synchronous acquisition of multimodal information. In order to completely solve the registration problem of multimodal images and achieve high-throughput, high-resolution light field, multispectral, polarization and high dynamic multifunctional integrated sensing on a compact device, the inventors have conducted a lot of research and exploration. Summary of the Invention

[0008] The technical problem to be solved by the invention is: addressing the technical bottlenecks of existing snapshot-type multi-purpose imaging devices, such as low light flux, multimodal sampling exposure mismatch, and failure of three-dimensional light field perception and reduced spatial resolution due to abrupt changes in spectral characteristics. The invention provides a high-light-flux multi-purpose imaging device that can simultaneously achieve high-light-flux and high-spatial-resolution light field imaging, multispectral imaging, polarization imaging, and high dynamic range imaging.

[0009] The technical solution of this invention is: A multi-purpose imaging device includes a lens and a light field camera body, the light field camera body including a microlens array and a photosensitive element. Its features are: The lens is a constant aperture, image-side telecentric lens, and the aperture opening is square. The photosensitive element of the light field camera body includes several closely arranged pixels, each pixel is a square structure, and the physical size of each pixel is not less than 5 micrometers; The microlens array of the light field camera body includes several microlens units, each of which is square, with equal side length and equal focal length. The side length of each microlens unit is N times the side length of the pixel of the photosensitive element (N is a positive integer, N>2). Adjacent microlens units are arranged closely together without gaps. In the microlens array of the light field camera body, the square planar area where each microlens unit is located coincides with the square planar area where the N×N pixels of the photosensitive element are located on a plane perpendicular to the imaging optical axis. When focusing throughout the lens's entire travel, its focal ratio remains constant, and the physical size of the aperture remains constant; When focusing throughout the lens's entire travel, the lens's focal ratio and the focal ratio of each microlens unit are the same; A multi-dimensional optical encoder array is installed at the aperture of the lens. It consists of two closely stacked, identical elements: a multi-dimensional array optical information modulation element and an anti-diffraction frame element. The multi-dimensional array optical information modulation element is the core physical encoder for achieving fully coaxial, zero-parallax multimodal imaging with a single lens, by spatially multiplexing and feature mapping multi-dimensional information in the aperture domain. The anti-diffraction frame element works in conjunction with the multi-dimensional array optical information modulation element to separate modules with different optical information modulation functions, preventing cross-leakage of light.

[0010] Furthermore, the multi-dimensional array optical information modulation element is a square structure consisting of N×N optical information modulation elements arranged together.

[0011] Furthermore, the anti-diffraction frame element is a square structure with N×N through holes.

[0012] The beneficial effects of this invention are: (1) Perfect intensity matching and collaborative sampling of multi-modal modulation are achieved. The broadband transmission characteristics of the notch filter bring the reference transmittance differences between the spectral modulation region and the polarization modulation region (linear polarization or ellipsoidal polarization) and the high dynamic modulation region (attenuator) closer. This design ensures that the sampled values ​​of the photosensitive pixels corresponding to each optical information modulation region are in the same reasonable linear dynamic response range, completely eliminating the "underexposure" or "overexposure" phenomenon caused by the large difference in channel transmittance in traditional systems, and realizing the physical matching of multi-modal signals under a single global exposure.

[0013] (2) High light field energy utilization and weak light detection capability. This invention abandons the traditional narrowband filter array scheme and creatively uses a broadband notch filter for spectral modulation in the multi-dimensional array optical information modulation element. The notch filter only blocks a narrower specific wavelength band, while allowing the vast majority of the remaining broadband spectral energy to pass through. Compared with the existing narrowband filter scheme, the overall light flux is generally less than 5%, while the physical architecture of this invention achieves multispectral detection and increases the overall light flux of the system to more than 70%, fundamentally breaking through the application bottleneck of multispectral imaging in low-light environments or high-speed dynamic scenes.

[0014] (3) It is compatible with and enhances the three-dimensional spatial reconstruction and super-resolution functions of light field cameras. The notch filter selected in this invention only filters out extremely narrow frequency bands, and the images acquired in each spectral modulation sub-region maintain a very high consistency in macroscopic visual features and structural textures. This allows the device to still support the parallax matching algorithm of traditional light field cameras while performing multi-dimensional analysis. It can not only output high-precision three-dimensional depth maps, but also successfully provide a physical basis for multi-view feature fusion algorithms, thereby greatly improving the spatial resolution of the final output image and solving the problem of low resolution of conventional light field cameras.

[0015] (4) Hardware decoupling, low processing difficulty and significantly reduced cost. This invention provides a multi-purpose imaging device with low processing difficulty and high yield. Under existing technical conditions, whether it is a multispectral imaging device, a polarization imaging device, or a high dynamic range imaging device, they generally adopt on-chip coating process. This process involves coating an array of thin films on the surface of the camera's photosensitive element. For example, for multispectral imaging devices, it is coating an array of spectral filters; for polarization imaging devices, it is coating an array of polarization filters; and for high dynamic range imaging devices, it is coating an array of attenuation films. Existing on-chip coating processes are cumbersome, complex, have low yield, and are costly. For example, to manufacture a 4×4 array spectral filter, it is necessary to repeatedly coat it 16 times. Any coating failure during this process will lead to the scrapping of the entire product, and all previous efforts will be wasted. This invention avoids the complex on-chip pixel-level coating process and places a large-size multi-dimensional optical coding chip array at the aperture of the lens, which greatly reduces the processing difficulty, increases the yield, and has high industrial mass production value. IV. Description of the attached drawings

[0016] Figure 1 This is a schematic diagram of the principle of the high-light-throughput multi-purpose imaging device provided by the present invention; Figure 2 These are four specific implementations of the multi-dimensional array optical information modulation element provided by the present invention; Figure 3 These are two specific embodiments of the anti-diffraction frame element provided by the present invention; Figure 4 These are two specific implementations of the multi-dimensional optical coding chip array provided by the present invention.

[0017] Figure 5 This is a close-up image of the high-throughput multi-purpose imaging device and its lens provided by the present invention. V. Detailed Implementation Methods

[0018] Figure 1 This is a schematic diagram of the high-throughput multi-purpose imaging device provided by the present invention, including a lens ① and a light field camera body ②. The light field camera body ② includes a microlens array ③ and a photosensitive element ④. The plane containing the photosensitive element ④ is parallel to the plane containing the microlens array ③, and both are perpendicular to the optical axis of the lens ①. The pixels of the photosensitive element ④ have a square structure, and the physical size is not less than 5 micrometers. This size limitation is to minimize the adverse effects of optical diffraction and ensure sufficient single-pixel signal-to-noise ratio when calculating spectral, polarization, and light field information. The pixel size of existing imaging cameras is typically less than 5 micrometers. A pixel binning technique can be used to expand it to more than 5 micrometers. For example, a pixel size of 4.5 micrometers can be binned using 2×2 pixels, which is equivalent to expanding the pixel size to 9 micrometers. If 4×4 pixels are binned, it is equivalent to expanding the pixel size to 18 micrometers.

[0019] Figure 1 In the microlens array ③, each microlens unit is a square with the same side length and focal length. All microlens units have the same focal ratio as lens ①, and adjacent microlens units are arranged closely together without gaps. The square planar region containing each microlens unit coincides with the square planar region containing the N×N pixels of the imaging sensor on a plane perpendicular to the imaging optical axis. Each microlens unit is square, and its side length is N times the pixel size (N is a positive integer greater than 2). The distance between the plane containing all microlens units in microlens array ③ and the imaging sensor surface of the camera sensor ④ is equal to the focal length of the microlens.

[0020] Figure 1 In the image, a multi-dimensional optical coding chip array ⑤ is arranged at the aperture ⑧ of lens ①. The multi-dimensional optical coding chip array ⑤ consists of an anti-diffraction frame element ⑥ and a multi-dimensional array optical information modulation element ⑦. There are no requirements for the superposition order of the two.

[0021] The multi-dimensional array optical information modulation element ⑦ satisfies the following conditions: it consists of N×N identical square optical information modulation elements. There are three types of optical information modulation elements: spectral filters (for spectral information modulation), attenuators (for intensity information modulation), and polarizers (for polarization information modulation). By selecting different combinations of optical information modulation elements, the imaging device of this invention can have a specific imaging function, or multiple imaging functions simultaneously.

[0022] The three types of optical modulation elements can be combined by splicing or by applying multiple coatings to the same substrate material (usually quartz glass). The total number of optical modulation elements must be at least nine. The number of polarizers can be between 0 and 4, the number of attenuators between 0 and 3, and the remainder are spectral filters. The optimal combination is N = 4, meaning 16 optical modulation elements, including 10 spectral filters, 2 attenuators, and 4 polarizers. In this configuration, the imaging device has the maximum functionality, the minimum light crosstalk, and achieves optimal imaging performance for each application.

[0023] The relative number of the three types of optical information modulation elements—spectral filters, attenuators, and polarizers—can be increased or decreased according to actual needs. For example, when N = 4, attenuators or polarizers can be replaced with spectral filters, thereby increasing the number of spectral filters to 12 or even 16, thus increasing the spectral resolution of the multi-purpose imaging device. However, the corresponding high dynamic range imaging function or polarization imaging function will be sacrificed, retaining only the light field imaging function and multispectral imaging function.

[0024] When there are two polarizers, both are linear polarizers with polarization directions differing from each other by 90°. When there are three polarizers, all three are linear polarizers with polarization directions differing from each other by 60°. When there are four polarizers, the most common combinations are: Combination 1: All four polarizers are linear polarizers, with the polarization direction of one polarizer as the reference, and the polarization directions of the other three polarizers differing by 45°, 90°, and 135° respectively. Combination 2: Of the four polarizers, three are linear polarizers with polarization directions differing from each other by 60°, and one is a left-handed or right-handed circular polarizer. It should be noted that the above specific angle configurations are only typical embodiments for polarization dimensionality reduction sampling in the system, and the polarization state combinations of the multi-dimensional array optical information modulation elements in this invention are not limited to these. In practical applications, to minimize error propagation and noise variance during the calculation of the target's full Stokes polarization parameters, an optimal sampling combination based on a system noise model (such as Gaussian noise or a photon noise-dominated environment) is generally adopted. Any polarization modulation configuration scheme that can achieve effective decoupling and dimensionality reduction sampling of the optical field polarization dimension falls within the protection scope of this invention.

[0025] When the number of attenuators is greater than or equal to 2, the light transmittance of the multiple attenuators needs to differ from each other by at least one order of magnitude. For example, for three attenuators, the transmittance of attenuator 1 is 10%, the transmittance of attenuator 2 is 1%, and the transmittance of attenuator 3 is 1‰.

[0026] The number of spectral filters falls within the range of [N×N-7, N×N]. Notch filters are generally chosen, with their spectral range completely covering the sensitive spectral band of the camera's image sensor. For example, for ordinary CMOS or CCD image sensors used in the visible light range, the spectral range is 360nm-1100nm; for image sensors used in short-wave infrared imaging, the spectral range is generally 900nm-1700nm; and for broadband image sensors encompassing both visible and near-infrared light (such as the Sony IMX990 chip), the spectral range is 400nm-1700nm. For all spectral filters, the notch spectral band is broadbandly modulated within the sensitive spectral band of the camera's image sensor (such as visible light 360nm-1100nm or broadband 400nm-1700nm). This invention is not limited to a specific notch center wavelength distribution formula or a fixed linewidth configuration; its core design principle lies in optimizing the mathematical transfer matrix for multimodal broadband sampling. Specifically, the combined design of spectral filters must simultaneously satisfy the following optimization principles: First, maintain the overall high luminous flux of the system; second, minimize the condition number of the system's spectral modulation matrix to ensure the inverse spectral reconstruction algorithm possesses extremely high numerical stability and noise robustness; third, the higher the independence (i.e., the lower the correlation) between the spectral response curves of each channel, the better. In practical applications, no specific transmittance curve design scheme is limited; any spectral modulation array constructed based on the above-mentioned dimensionality reduction sampling optimization principles falls within the protection scope of the claims of this invention. To further illustrate the above optimization principles, this embodiment provides a method for a 400nm-1000nm wide-spectrum photosensitive chip (effective spectral range...). This example describes a specific transmittance curve design scheme for 10 notch spectral filters. To maximize luminous flux and ensure the independence of each channel, the spectral response curves of the 10 notch filters must meet the following specific parameter requirements: 1. Broadband high transmittance background: The reference transmittance of all filters in the non-notch band. All are greater than or equal to 90%; minimum transmittance in the center band of the notch wave. Less than or equal to 5%. 2. Center wavelengths are evenly spaced: the center wavelengths of the 10 notch filters... Arranged at equal intervals within the 400nm-1000nm range. For example, the center wavelengths of each filter can be set as follows: 430nm, 490nm, 550nm, 610nm, 670nm, 730nm, 790nm, 850nm, 910nm, 970nm. 3. Full width at half maximum (FWHM) matching: The blocking band FWHM of each notch filter is set to... This causes the edges of the notch blocking bands of adjacent filters to be connected end to end or to have slight overlap (overlap rate). This design example perfectly aligns with the core optimization principle of this invention, and its physical and mathematical advantages are reflected in: First, extremely high luminous flux. Because each filter only blocks 60nm of the spectrum, allowing the remaining 540nm of the broadband spectrum to pass through, ideally, the light energy utilization of a single filter can reach... Compared to traditional narrowband bandpass filters with only 5% transmittance, this example increases the luminous flux several times, completely solving the underexposure problem in low-light detection. Secondly, it features an extremely low spectral modulation matrix condition number. The discretized transmittance responses of the aforementioned 10 notch filters are combined into one... spectral modulation matrix Under this design, the matrix The structure closely approximates the well-known S-matrix in computational optics. Mathematically, the S-matrix possesses an extremely small condition number and optimal non-negative matrix inversion stability. This means that in real-world physical environments with sensor shot noise or dark current noise, noise is not amplified by the ill-conditioned matrix during inverse spectral reconstruction, resulting in a highly robust hyperspectral data cube with excellent noise resistance. Third, excellent channel independence. Although all filters are broadband (appearing highly overlapping), their unique "signal characteristics (notch dark bands)" are orthogonally distributed in the spectral domain, causing the spectral response curves of each channel to exhibit extremely low correlation in the derivative domain (rate of change), thus ensuring the accuracy of multimodal computational decoupling.

[0027] Figure 2 These are four specific implementations of the multi-dimensional array optical information modulation element provided by the present invention. Figure 2 (a) is the first implementation, which corresponds to the case of N=3, that is, it consists of 3×3 square optical information modulation elements, including 5 spectral filters, all of which are notch filters; 3 polarizers, all of which are linear polarizers, and the angle between their polarization directions is 60°; and 1 attenuator, which is located at the center of the multi-dimensional array optical information modulation elements. The attenuator is placed in the center to minimize the light diffraction and crosstalk effects on the surrounding elements. Figure 2 (b) is the second implementation, corresponding to N=3, which includes 6 notch spectral filters and 3 polarizers. This layout eliminates the attenuators, sacrificing high dynamic range imaging capabilities in exchange for higher multispectral resolution. In practical applications, it is sometimes not necessary to have all four imaging functions, so this invention provides the possibility of flexibly combining various functional combinations to suit different application scenarios.

[0028] Figure 2(c) is the third implementation, which corresponds to the case of N = 4, that is, it consists of 4×4 square optical information modulation elements, including 10 spectral filters, all of which are notch filters; 2 attenuators, the transmittance of attenuator 1 is 10%, and the transmittance of attenuator 2 is 1‰; and 4 polarizers, all of which are linear polarizers, with the polarization direction of one polarizer as the reference, and the polarization directions of the other three polarizers differing by 45°, 90° and 135° respectively.

[0029] Figure 2 (d) is the fourth implementation method, which corresponds to the case of N = 4. It consists of 4×4 square optical information modulation elements, including 12 spectral filters, all of which are notch filters; and 4 polarizers, three of which are linear polarizers with polarization directions 60° apart, and the remaining one is a left-handed or right-handed circular polarizer.

[0030] Figure 3 These are two specific embodiments of the anti-diffraction frame element provided by the present invention. The anti-diffraction frame element meets the following conditions: it is square, with the same size as the aperture stop ⑧ of the lens, and its surface is treated with anti-reflective coating, for example, for metal materials, it can be anodized blackened, coated with black paint, or other light-absorbing materials, such as ultra-black nano-coating, etc. It has N×N equally spaced holes of the same size inside. The holes can be square or circular. If they are square, their side length is smaller than that of the square optical information modulation element; if they are circular, their diameter is smaller than that of the square optical information modulation element. Figure 3 (a) corresponds to the implementation where the internal hole is square. Figure 3 (b) corresponds to the implementation with a circular aperture. It is important to note that the size of the internal aperture needs to balance two factors: the suppression effect of light diffraction and the light energy utilization rate. If the aperture is too large, although the light energy utilization rate is relatively high, the suppression effect of light diffraction is relatively weak; if the aperture is too small, the suppression effect of light diffraction is good, but the light energy utilization rate will decrease. The most suitable aperture size and position should be such that the center of the aperture coincides with the center of the square optical information modulation element, and the area of ​​the aperture is between 0.6 and 0.8 times the area of ​​the square optical information modulation element.

[0031] Figure 4 These are two specific implementations of the multi-dimensional optical encoding chip array ⑤ provided by the present invention. It consists of two closely stacked parts: a multi-dimensional array optical information modulation element ⑦ and an anti-diffraction frame element ⑥. The physical arrangement of this structure at the lens aperture ⑧ ensures the multimodal consistency and high spatial resolution characteristics of the device output data from the source.

[0032] Figure 5The present invention provides a physical example of the high-throughput multi-purpose imaging device and a close-up image of its lens ①. The multi-dimensional optical coding chip array ⑤ placed at the aperture stop ⑧ of the lens ① can be seen. The layout of the multi-dimensional array optical information modulation element ⑦ corresponding to this multi-dimensional optical coding chip array ⑤ is similar to... Figure 2 The second implementation in (b) is the same. Furthermore, the anti-diffraction frame element ⑥ is not visible in the figure because it is placed behind the multi-dimensional array optical information modulation element ⑦ and is obscured by it.

[0033] After developing a physical high-throughput multi-purpose imaging device using the solution described in this invention, the applicant conducted extensive market research and potential user surveys, reaching a clear conclusion: The product of this invention, with its core technological advantages such as simultaneous acquisition of multiple imaging modes (multispectral, polarization, etc.), zero parallax fusion registration, and flexible function switching, has high application potential in multiple mainstream application markets including scientific research, agriculture, anti-drone, industry, cultural relic protection, and autonomous driving. Furthermore, its comprehensive performance in various sub-scenarios surpasses current mature products, resulting in high market acceptance. In addition, the components and modules used in this invention are entirely domestically produced, ensuring leading performance while possessing core competitiveness based on independent control, further enhancing market acceptance.

[0034] Firstly, in the field of scientific research, this product can flexibly switch between multiple imaging channels, perfectly adapting to the diverse imaging needs of various scientific experiments, effectively reducing redundant investment in scientific equipment, and minimizing laboratory space occupation and equipment maintenance costs. During a trial period with a university, we provided a high-throughput, multi-purpose imaging device with 16 imaging channels. It can simultaneously and accurately capture spectral and polarization imaging details and dynamic changes during experiments, significantly improving experimental efficiency. The university's researchers gave the device high praise.

[0035] In the field of agricultural remote sensing, this product can be easily mounted on various drones to achieve precise monitoring and early warning of crop diseases. Utilizing visible light and near-infrared multimodal imaging technology and Normalized Difference Vegetation Index (NDVI) analysis capabilities, it produces clear images without any parallax, eliminating the need for cumbersome image registration steps. It is particularly suitable for mountainous and hilly areas (such as tea plantations), significantly saving operation time and labor costs. During the trial period, it has gained high recognition from users in various agricultural planting scenarios.

[0036] In the field of anti-drone and target recognition, this product possesses a core advantage that distinguishes it from competitors. Due to the use of a notch filter, it significantly improves light energy utilization. This key design allows the camera frame rate to exceed 900 fps, far surpassing the conventional 30 fps frame rate of the market. This high frame rate perfectly meets the core needs of anti-drone scenarios, ensuring real-time response to drone targets, quickly capturing the dynamic trajectory of drones, and timely target recognition and early warning. It effectively solves the core pain points of competing products, such as response delays, untimely target tracking, and delayed recognition caused by insufficient frame rates. Furthermore, this invention can assist in the stable recognition of drone targets in complex backgrounds, backlighting, glare, and other adverse imaging environments, providing efficient, stable, and reliable technical support for anti-drone prevention and control work in various scenarios. Currently, there is strong demand in this field, and it has received unanimous praise from users during trials at various research institutions.

[0037] In the field of industrial quality inspection, this product can accurately locate and measure various defects in products under complex industrial environments such as high contrast, strong light, and shadow, effectively improving the accuracy and efficiency of industrial quality inspection and reducing inspection costs. Furthermore, in the field of cultural relic preservation, this product can accurately record the chemical composition of cultural relics through multispectral imaging and achieve glare removal and texture revealing through polarization imaging, assisting in the protection and restoration of cultural relics. In the field of autonomous driving, it can maintain stable recognition under adverse weather conditions such as rain, snow, backlight, and heavy fog, ensuring the safe and stable operation of autonomous driving systems.

[0038] In summary, this invention demonstrates core advantages over existing products in all the above application scenarios, has broad market application prospects, and is expected to generate good returns after being launched on the market.

Claims

1. A multi-purpose imaging device, comprising a lens and a light field camera body, the light field camera body including a microlens array and a photosensitive element, characterized in that, The lens is a constant aperture, image-side telecentric lens, and the aperture opening is square. The photosensitive element of the light field camera body includes several closely arranged pixels, each pixel is a square structure, and the physical size of each pixel is not less than 5 micrometers; The microlens array of the light field camera body includes several microlens units, each of which is square, with equal side length and equal focal length. The side length of each microlens unit is N times the side length of the pixel of the photosensitive element (N is a positive integer, N>2). Adjacent microlens units are arranged closely together without gaps. In the microlens array of the light field camera body, the square planar area where each microlens unit is located coincides with the square planar area where the N×N pixels of the photosensitive element are located on a plane perpendicular to the imaging optical axis. When focusing throughout the lens's entire travel, its focal ratio remains constant, and the physical size of the aperture remains constant; When focusing throughout the lens's entire travel, the lens's focal ratio and the focal ratio of each microlens unit are the same; Two closely stacked and identical components are installed at the aperture of the lens: one is a multi-dimensional array optical information modulation component, and the other is an anti-diffraction frame component. The multi-dimensional array optical information modulation element is the core physical encoder for realizing single-lens fully coaxial, zero-parallax multimodal imaging, by spatially multiplexing and feature mapping of multi-dimensional information in the aperture domain; The function of the anti-diffraction frame element is to work with the multi-dimensional array optical information modulation element to separate modules with different optical information modulation functions and block the cross-leakage of light.

2. The multi-purpose imaging device according to claim 1, characterized in that, The multidimensional array optical information modulation element is a square structure consisting of N×N optical information modulation elements arranged together.

3. The multi-purpose imaging device according to claim 1, characterized in that, The anti-diffraction frame element is a square structure with N×N through holes.

4. The multi-purpose imaging device according to claim 2, characterized in that, The optical information modulation element includes at least one of a spectral filter, an attenuator, and a polarizer, and the total number of optical information modulation elements is at least nine.

5. The multi-purpose imaging device according to claim 4, characterized in that, The spectral filter is a notch filter, whose spectral range completely covers the spectral band of the photosensitive element.

6. The multi-purpose imaging device according to claim 4, characterized in that, In the optical information modulation element, the number of polarizers is 0-4, the number of attenuators is 0-3, and the rest are spectral filters.

7. The multi-purpose imaging device according to claim 6, characterized in that, N=4, the total number of optical information modulation elements is 16, including 10 spectral filters, 2 attenuators and 4 polarizers.

8. The multi-purpose imaging device according to claim 4, characterized in that, When the number of attenuators is ≥2, the light transmittance of each attenuator differs by at least one order of magnitude.

9. The multi-purpose imaging device according to claim 3, characterized in that, The through-hole of the anti-diffraction frame element is square or circular, the center of the through-hole coincides with the center of the corresponding optical information modulation element, and the area of ​​the through-hole is 0.6-0.8 times the area of ​​the corresponding optical information modulation element.

Images