Imaging methods and apparatus for improving the resolution of multi-slit stripe laser imaging systems

Abstract

This invention relates to an imaging method and apparatus for improving the resolution of a multi-slit stripe laser imaging system, and pertains to the field of lidar imaging technology. It addresses the problem of low imaging resolution in existing lidar imaging technologies using multi-slit stripe arrays. The invention acquires target images based on laser detection and a multi-slit stripe camera. The field of view of the laser detection imaging is divided into M0×N0 blocks, each block containing m0×n0 pixels, for a total of (M0·m0)×(N0·n0) pixels in the entire detection field of view. The DMD is controlled to transform N times, obtaining N stripe image measurement results. The image obtained from one measurement is n = M0·N0 stripe images with m = m0·n0 horizontal pixels. Intensity image extraction is performed to obtain a low-resolution target intensity image of size m×n. Based on the intensity values ​​of each pixel measured N times, a reconstruction algorithm is used to synchronously reconstruct each m0×n0 block, thereby reconstructing the target intensity image resolution to a K×L resolution size, thus achieving imaging.

Description

Technical Field

[0001] This invention relates to the field of lidar imaging technology, and more specifically to a method and imaging device for improving the resolution of a multi-slit stripe laser imaging system. Background Technology

[0002] Stripe tube laser imaging radar is a type of radar that uses a stripe camera with high temporal resolution as a detector. It can simultaneously achieve distance imaging and intensity imaging of a target and has wide applications in aerospace, topographic mapping, underwater exploration and other fields.

[0003] Striped tube cameras utilize a large-area array technology, possessing ultrafast spatiotemporal characteristics that enable high spatial and distance resolution. They can be categorized into single-slit and multi-slit systems. Currently, single-slit striped tube imaging achieves high resolution, but still requires a precision scanning device to obtain a complete image of the target, and is prone to motion artifacts for moving targets. Multi-slit striped tube array imaging, on the other hand, eliminates the need for scanning equipment, allowing for direct single-frame array imaging, which offers advantages for imaging moving targets. However, the limited photocathode area of ​​the striped tube results in lower imaging resolution.

[0004] Computational imaging technology fully utilizes and compiles the obtained information through encoding principles, sensing mechanisms, and corresponding decoding methods, ultimately acquiring higher-resolution images through computational reconstruction. Compared with traditional physical optical imaging systems, it has significant advantages in improving imaging resolution, expanding detection distance, increasing imaging field of view, and reducing the size and power consumption of optical systems.

[0005] In summary, existing lidar imaging technologies cannot achieve high-resolution imaging of targets using a surface array imaging system. However, effectively combining computational imaging with multi-slit stripe laser imaging could achieve even higher resolution imaging, representing a significant innovation in stripe laser surface array imaging technology. Summary of the Invention

[0006] This invention aims to address the problem of low imaging resolution in existing lidar imaging technologies using multi-slit stripe arrays.

[0007] An imaging method for improving the resolution of a multi-slit stripe laser imaging system includes the following steps:

[0008] The target image is obtained based on laser detection and a multi-slit stripe camera; the field of view of laser detection imaging is divided into M0×N0 blocks for processing, each block is m0×n0 pixels; the total number of pixels in the detection field of view is (M0·m0)×(N0·n0);

[0009] A DMD with a resolution of K×L is used, where K=k1·m0·M0 and L=k2·n0·N0. The DMD is divided into (M0 m0)×(N0 n0) blocks. The DMD micromirror array is encoded by computer control, and the DMD is transformed N times to obtain N stripe image measurement results. In the multi-slit laser imaging field of view, the number of DMD pixels corresponding to one imaging unit is k1×k2.

[0010] A multi-slit stripe camera corresponds to an image with m horizontal pixels and n slits; m = m0·n0, n = M0·N0; the image obtained by a multi-slit stripe laser imaging radar in one measurement is n stripe images with m horizontal pixels. By extracting the intensity image from the stripe image, a low-resolution target intensity image of size m×n is obtained.

[0011] Based on the intensity values ​​of each pixel measured N times, each m0×n0 small block is reconstructed synchronously using a reconstruction algorithm, thereby reconstructing the resolution of the target intensity image to a K×L resolution size, thus achieving imaging.

[0012] Furthermore, in the process of obtaining an image with m horizontal pixels and n slits from a multi-slit stripe camera, one imaging unit is transformed into a row to correspond to one slit, that is: the m0×n0 region is transformed into m×1; there are a total of n slits.

[0013] Furthermore, based on the intensity values ​​measured N times for each pixel, the process of synchronously reconstructing each m0×n0 small block using a reconstruction algorithm includes the following steps:

[0014] Given that the DMD is encoded N times, and the stripe camera acquires N×n stripe images with N×m×n intensity values, the generative model for the system's acquired data is expressed as:

[0015] Ax = b

[0016] Where A∈R N×mn This is a light coding matrix, corresponding to N codes, each code containing K×L elements; x∈R K×L Let b be the column vector form of the target to be reconstructed; b∈R N This represents the column vector form of data acquired by the multi-slit stripe tube imaging system.

[0017] Reconstruct the target image x from the optical coding matrix A and the acquired data b; the reconstructed target image is represented as a quadratic minimization problem:

[0018]

[0019] The linear iterative gradient descent method is adopted to transform the reconstruction process into a process of minimizing the intensity values ​​of real data and measured data. The optimal solution is solved iteratively using the gradient and gradient descent step size, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

[0020] Furthermore, during the iterative process of finding the optimal solution, it is determined whether the average intensity of the pixel point measured N times is 0. If it is 0, the iterative solution is not performed and the value is directly assigned to 0; if it is not 0, the gradient descent method is used to solve the problem, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

[0021] An imaging device for improving the resolution of a multi-slit stripe laser imaging system includes: a laser, a transmitting optical system, a DMD digital micromirror array, a receiving optical system, a multi-slit stripe camera, an imaging CCD, and a computer;

[0022] The laser emits pulsed laser light, which travels through the emitting optical system to the DMD micromirror array. The DMD micromirror array spatially modulates the laser light, which then illuminates the target and is reflected. After passing through the receiving optical system, the laser light is imaged by a multi-slit fringe camera at the imaging CCD. Finally, the imaging CCD transmits the resulting target image to the computer, which reconstructs the image to obtain a high-resolution target intensity image.

[0023] The process of reconstructing a high-resolution target intensity image using a computer includes the following steps:

[0024] The process of synchronously reconstructing each m0×n0 block using a reconstruction algorithm based on the intensity values ​​measured N times for each pixel includes the following steps:

[0025] Given that the DMD is encoded N times, and the stripe camera acquires N×n stripe images with N×m×n intensity values, the generative model for the system's acquired data is expressed as:

[0026] Ax = b

[0027] Where A∈R N×mn This is a light coding matrix, corresponding to N codes, each code containing K×L elements; x∈R K×L Let b be the column vector form of the target to be reconstructed; b∈R N This represents the column vector form of data acquired by the multi-slit stripe tube imaging system.

[0028] Reconstruct the target image x from the optical coding matrix A and the acquired data b; the reconstructed target image is represented as a quadratic minimization problem:

[0029]

[0030] The linear iterative gradient descent method is adopted to transform the reconstruction process into a process of minimizing the intensity values ​​of real data and measured data. The optimal solution is solved iteratively using the gradient and gradient descent step size, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

[0031] Furthermore, during the iterative process of finding the optimal solution, it is determined whether the average intensity of the pixel point measured N times is 0. If it is 0, the iterative solution is not performed and the value is directly assigned to 0; if it is not 0, the gradient descent method is used to solve the problem, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

[0032] Furthermore, the device also includes a digital delay pulse generator; the laser emits pulsed laser light and pulsed electrical signal simultaneously, which is delayed by the digital delay pulse generator and then sent to the multi-slit stripe camera, thereby activating the multi-slit stripe camera to turn on the ramp voltage and put it into working state, i.e., to acquire stripe images. The delay time is the time it takes for the laser light emitted from the laser to reach the multi-slit stripe camera.

[0033] Furthermore, the laser is a pumped laser.

[0034] The beneficial effects of this invention are:

[0035] Based on existing multi-slit stripe tube imaging systems, this invention introduces computational imaging methods to spatially modulate lasers, utilizes non-scanning multiple encoding measurement imaging, and reconstructs high-resolution target intensity images through reconstruction algorithms to solve the problem of low resolution in current multi-slit stripe area array imaging. This enables high-resolution imaging of targets under area array imaging systems, promoting the development of high-resolution rapid imaging. Attached Figure Description

[0036] Figure 1 A schematic diagram of a device for improving the imaging resolution of a multi-slit stripe tube.

[0037] Figure 2 A schematic diagram of a method to improve the imaging resolution of multi-slit stripe tubes.

[0038] Figure 3 This is a schematic diagram of the random coded pattern of N×N blocks in a DMD micromirror array.

[0039] Figure 4 This is the original stripe intensity image acquired by the multi-slit stripe tube imaging system.

[0040] Figure 5 This is the high-resolution intensity image of the target obtained after processing by the reconstruction algorithm. Detailed Implementation

[0041] The apparatus and method of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0042] like Figure 1 As shown, an imaging device for improving the resolution of a multi-slit stripe laser imaging system includes:

[0043] 1. Laser, 2. Emitting optical system, 3. DMD digital micromirror array, 4. Receiving optical system, 5. Digital delayed pulse generator, 6. Multi-slit stripe camera, 7. Imaging CCD, 8. and computer.

[0044] Laser 1 is a pump laser. The laser emits pulsed laser light, which reaches DMD micromirror array 3 after passing through the emitting optical system 2. The DMD micromirror array spatially modulates the laser light. The modulated laser light illuminates the target 4 and is reflected. After passing through the receiving optical system 5, it is imaged at the imaging CCD 8 by the multi-slit fringe camera 7. Finally, the imaging CCD 8 transmits the obtained target image to computer 9 via USB. The computer performs image denoising, fringe feature extraction, and decoding and reconstruction on the obtained original fringe image to obtain a high-resolution target intensity image.

[0045] The laser emits pulsed laser light and pulsed electrical signal simultaneously. After being delayed by the digital delay pulse generator 6, the signal is sent to the multi-slit stripe camera 7, thereby activating the multi-slit stripe camera to turn on the ramp voltage and put it into working state so that it can acquire stripe images. The delay time is the time it takes for the laser light emitted from the laser to reach the multi-slit stripe camera.

[0046] like Figure 2 As shown, an imaging method for improving the resolution of a multi-slit stripe laser imaging system includes the following steps:

[0047] The highest resolution of the target image should be consistent with the DMD resolution. The field of view of the laser detection imaging is divided into M0×N0 blocks for processing, each block being m0×n0 pixels. Then the total number of pixels in the detection field of view should be: (M0·m0)×(N0·n0).

[0048] A DMD with a resolution of K×L is used, where K = k1·m0·M0 and L = k2·n0·N0. The DMD is divided into (M0·m0)×(N0·n0) blocks, and the micromirror array of the DMD is encoded by computer control. The encoding array is a pseudo-random 0 / 1 array, where 0 corresponds to the off state of the micromirrors and 1 corresponds to the on state of the micromirrors. By controlling the DMD to transform N times, N stripe image measurement results can be obtained.

[0049] In the multi-slit laser imaging field of view, the number of DMD pixels corresponding to one imaging unit is k1×k2;

[0050] In a multi-slit stripe laser imaging system (in a multi-slit stripe camera), an imaging unit is transformed into a row through an optical fiber converter to correspond to one slit, that is: the m0×n0 region is transformed into m×1; there are a total of n slits.

[0051] Where m is the number of horizontal pixels in the stripe image, and n is the number of slits; m = m0·n0, n = M0·N0;

[0052] Ultimately, the image obtained from a single measurement by the multi-slit stripe laser imaging radar consists of n stripe images with m horizontal pixels. By extracting the intensity values ​​from the stripe images, a low-resolution target intensity image of size m×n is obtained. Based on the intensity values ​​of each pixel measured N times, each m0×n0 small block is simultaneously reconstructed using a reconstruction algorithm. The optimal solution for the target intensity image can be obtained, and the resolution of the target intensity image is reconstructed to a resolution of size K×L, thereby improving the resolution of the multi-slit stripe laser imaging.

[0053] The process of synchronously reconstructing each m0×n0 block using a reconstruction algorithm based on the intensity values ​​measured N times for each pixel includes the following steps:

[0054] Given that the DMD is encoded N times, and the stripe camera acquires N×n stripe images with N×m×n intensity values, the generation model of the data acquired by this system can be expressed as:

[0055] Ax = b

[0056] Where A∈R N×mn This is a light coding matrix, corresponding to N codes, each code containing K×L elements; x∈R K×L Let b be the column vector form of the target to be reconstructed; b∈R N This represents the column vector form of the data acquired by the multi-slit stripe tube imaging system.

[0057] Reconstruct the target image x from the optical encoding matrix A and the acquired data b. The reconstructed target image can be represented as a quadratic minimization problem:

[0058]

[0059] A linear iterative gradient descent method is employed to transform the reconstruction process into minimizing the intensity values ​​of the real data and the measured data. The optimal solution is obtained iteratively using the gradient and the gradient descent step size. The algorithm checks if the average intensity of N measured values ​​for a pixel is 0. If it is 0, no iterative solution is performed, and the value is directly assigned to 0. If it is not 0, the gradient descent method is used to solve the problem, ultimately reconstructing a target intensity image with a resolution of m0×n0.

[0060] The overall target scene reconstruction involves reconstructing the image synchronously based on the measurement results of each pixel, and finally reconstructing the resolution of the target intensity image to a K×L resolution size, thereby improving the resolution of multi-slit stripe laser imaging.

[0061] Based on existing multi-slit stripe tube imaging systems, this invention introduces a computational imaging method to spatially modulate the laser, utilizes non-scanning multiple encoding measurement imaging, and reconstructs a high-resolution target intensity image through a reconstruction algorithm. This addresses the problem of low resolution in current multi-slit stripe array imaging and promotes the development of high-resolution rapid imaging.

[0062] Example

[0063] Following the above scheme, experiments were conducted on a fixed scene target. For ease of calculation, it was assumed that the DMD used a 400×400 resolution and was processed using 20×20 block encoding. For example... Figure 3 The image shows the pseudo-random coding pattern of each 20×20 DMD. The pseudo-random coding is generated randomly under computer control, and 200 intensity images are obtained by measuring 200 times of coding transformation. Figure 3 (a), (b), (c), and (d) are 20×20 DMD encoded arrays for the 1st, 2nd, 3rd, and 200th iterations, respectively. Figure 4 The image shown is an intensity image of a target in a scene captured by a multi-slit stripe camera, with an imaging resolution of 20×20. Figure 4 (a), (b), (c), and (d) are the target intensity images obtained by the multi-slit fringe camera during the 1st, 2nd, 3rd, and 200th scans, respectively. Figure 5 The image shown is an intensity image of the target scene reconstructed using the gradient descent algorithm, with an imaging resolution of 400×400.

[0064] The above examples of the present invention are merely illustrative of the computational model and process of the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is impossible to exhaustively list all possible implementations here. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. An imaging method for improving the resolution of a multi-slit stripe laser imaging system, characterized in that, Includes the following steps: Target images are obtained using laser detection and a multi-slit stripe camera; The field of view of laser detection imaging is divided into M0×N0 blocks for processing, each block being m0×n0 pixels; the total number of pixels in the entire detection field of view is (M0·m0)×(N0·n0); A DMD with a resolution of K×L is used, where K=k1·m0·M0 and L=k2·n0·N0. The DMD is divided into (M0 m0)×(N0n0) blocks, and the DMD micromirror array is encoded by computer control. The DMD is transformed N times to obtain N stripe image measurement results. In the multi-slit laser imaging field of view, the number of DMD pixels corresponding to one imaging unit is k1×k2. A multi-slit stripe camera corresponds to an image with m horizontal pixels and n slits; m = m0·n0, n = M0·N0; the image obtained by a multi-slit stripe laser imaging radar in one measurement is n stripe images with m horizontal pixels. By extracting the intensity image from the stripe image, a low-resolution target intensity image of size m×n is obtained. Based on the intensity values ​​of each pixel measured N times, each m0×n0 small block is reconstructed synchronously using a reconstruction algorithm, thereby reconstructing the resolution of the target intensity image to a K×L resolution size, thus achieving imaging.

2. The imaging method for improving the resolution of a multi-slit stripe laser imaging system according to claim 1, characterized in that, In the process of obtaining an image with m horizontal pixels and n slits from a multi-slit stripe camera, one imaging unit is transformed into a row to correspond to one slit, that is: the m0×n0 region is transformed into m×1; there are a total of n slits.

3. An imaging method for improving the resolution of a multi-slit stripe laser imaging system according to claim 1 or 2, characterized in that, The process of synchronously reconstructing each m0×n0 block using a reconstruction algorithm based on the intensity values ​​measured N times for each pixel includes the following steps: Given that the DMD is encoded N times, and the stripe camera acquires N×n stripe images with N×m×n intensity values, the generative model for the system's acquired data is expressed as: Ax = b Where A∈R N×mn This is a light coding matrix, corresponding to N codes, each code containing K×L elements; x∈R K×L Let b be the column vector form of the target to be reconstructed; b∈R N This represents the column vector form of data acquired by the multi-slit stripe tube imaging system. Reconstruct the target image x from the optical coding matrix A and the acquired data b; the reconstructed target image is represented as a quadratic minimization problem: The linear iterative gradient descent method is adopted to transform the reconstruction process into a process of minimizing the intensity values ​​of real data and measured data. The optimal solution is solved iteratively using the gradient and gradient descent step size, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

4. The imaging method for improving the resolution of a multi-slit stripe laser imaging system according to claim 3, characterized in that, During the iterative process of finding the optimal solution, it is determined whether the average intensity of the pixel point after N measurements is 0. If it is 0, the iteration is not performed and the value is directly assigned to 0; if it is not 0, the gradient descent method is used to solve the problem, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

5. An imaging device for improving the resolution of a multi-slit stripe laser imaging system, characterized in that, include: Laser, transmitting optical system, DMD digital micromirror array, receiving optical system, multi-slit stripe camera, imaging CCD and computer; The laser emits pulsed laser light, which then travels through the emitting optical system to the DMD micromirror array. The DMD micromirror array spatially modulates the laser, which then illuminates and reflects off the target. After passing through the receiving optical system, the laser is imaged at the imaging CCD by a multi-slit fringe camera. Finally, the imaging CCD transmits the resulting target image to a computer, which reconstructs the image to obtain a high-resolution target intensity image. The process of reconstructing a high-resolution target intensity image using a computer includes the following steps: The process of synchronously reconstructing each m0×n0 block using a reconstruction algorithm based on the intensity values ​​measured N times for each pixel includes the following steps: Given that the DMD is encoded N times, and the stripe camera acquires N×n stripe images with N×m×n intensity values, the generative model for the system's acquired data is expressed as: Ax = b Where A∈R N×mn This is a light coding matrix, corresponding to N codes, each code containing K×L elements; x∈R K×L Let b be the column vector form of the target to be reconstructed; b∈R N This represents the column vector form of data acquired by the multi-slit stripe tube imaging system. Reconstruct the target image x from the optical coding matrix A and the acquired data b; the reconstructed target image is represented as a quadratic minimization problem: The linear iterative gradient descent method is adopted to transform the reconstruction process into a process of minimizing the intensity values ​​of real data and measured data. The optimal solution is solved iteratively using the gradient and gradient descent step size, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

6. The imaging device for improving the resolution of a multi-slit stripe laser imaging system according to claim 5, characterized in that, During the iterative process of finding the optimal solution, it is determined whether the average intensity of the pixel point after N measurements is 0. If it is 0, the iteration is not performed and the value is directly assigned to 0; if it is not 0, the gradient descent method is used to solve the problem, and finally the target intensity image with a resolution of m0×n0 is reconstructed.

7. An imaging device for improving the resolution of a multi-slit stripe laser imaging system according to claim 6, characterized in that, The device also includes a digital delay pulse generator; The laser emits pulsed laser light and pulsed electrical signal simultaneously. After being delayed by a digital delay pulse generator, the signal is sent to the multi-slit stripe camera, thereby activating the multi-slit stripe camera to turn on the ramp voltage and put it into working state to acquire stripe images. The delay time is the time it takes for the laser light emitted from the laser to reach the multi-slit stripe camera.

8. An imaging device for improving the resolution of a multi-slit stripe laser imaging system according to claim 6 or 7, characterized in that, The laser is a pumped laser.

Images