Optical-digital collaborative optimization zoom imaging system design method

Abstract

The application discloses a zoom imaging system design method based on light-digital collaborative optimization, belongs to the technical field of computational imaging, adopts a four-piece zoom structure as an initial optical system, uses Zemax to simulate output blurred images and a spatial variable point spread function (PSF), and constructs an end-to-end joint optimization platform through VS Code and Matlab; the digital processing part is based on a block strategy, divides the images and the PSF into n*n blocks, combines boundary extension and gradual fusion technology, uses a Lucy-Richardson algorithm to perform non-blind deblurring processing, and optimizes brightness through dichotomy; finally, the mean square error of the restored image and the original image is fed back to Zemax as an evaluation function to drive the optimization of the optical system. While maintaining the system complexity, the application significantly improves the restoration effect in short-focus and other scenes, and the imaging quality is better than that of a traditional independent design mode.

Description

Technical Field

[0001] This invention belongs to the field of computer imaging technology, specifically relating to a design method for a zoom imaging system with optical-digital collaborative optimization. Background Technology

[0002] With the rapid development of optoelectronic imaging technology, the market demands increasingly higher performance from optical systems, including requirements for high resolution, wide field of view, and long detection distance, while also meeting design goals of lightweight, miniaturization, and low cost. Traditional zoom optical systems typically employ increased lens quantity, complex surface designs, and precision mechanical structures to correct aberrations and achieve focal length adjustment, but this leads to increased system size and cost, contradicting current application requirements for portability and integration.

[0003] like Figure 1 As shown, in traditional design methods, optical systems and image processing algorithms are often optimized independently: optical design pursues the minimization of aberrations at the hardware level, while digital algorithms are only used as a post-processing remedial measure. This fragmented optimization model makes it difficult to achieve global optimization and easily leads to resource waste—the optical system may become overly complex due to over-design, while the potential of digital algorithms is not fully utilized. Especially in zoom systems, the coordinated movement of the zoom group and the compensation group has a significant impact on image quality, and traditional methods struggle to balance aberration correction and system complexity across the entire focal length range.

[0004] Optical-digital joint optimization techniques offer a novel approach to solving this problem. By co-optimizing optical design and image processing algorithms, their respective strengths can be fully leveraged, achieving a better balance between system complexity and image quality. However, existing joint design methods still have shortcomings in zoom systems, such as the lack of adaptive handling for spatially variable aberrations or the absence of an efficient end-to-end optimization framework. Therefore, a more efficient optical-digital joint design method is urgently needed to improve the overall performance of zoom systems. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a design method for a zoom imaging system based on optical-digital collaborative optimization. Compared to traditional design methods, this method achieves better overall imaging performance for zoom optical systems of equal complexity. It also achieves equal or better overall imaging performance for systems with higher complexity, thus solving the problem of increased complexity caused by traditional designs that only obtain locally optimal solutions.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] A design method for a zoom imaging system with optical-digital co-optimization, the method comprising:

[0008] Step 1: The optical system outputs a blurred image of the initial optical structure and a spatially variable PSF dataset through Zemax;

[0009] Step 2: In Matlab, perform image block restoration processing on the blurred image and PSF dataset to obtain the restored image. The image block restoration processing includes boundary expansion compensation and linear gain method merging to eliminate artifacts at the stitching boundaries.

[0010] Step 3: Matlab compares the image evaluation metrics of the restored image with those of the initial optical structure. If the comparison result does not meet the design specifications, the comparison result is fed back to the optical system. Steps 1-3 are executed again with the restored image as the initial optical structure to form a closed-loop optimization until the imaging quality requirements are met.

[0011] In a second aspect, the present invention provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned optical-digital co-optimized zoom imaging system design method.

[0012] Thirdly, the present invention provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enable the processor to implement the aforementioned optical-digital co-optimized zoom imaging system design method.

[0013] The beneficial effects of this invention are as follows:

[0014] The optical-digital co-design zoom system and method based on the Lucy-Richardson algorithm proposed in this invention have significant technical advantages and application value. Compared with the traditional independent design mode, this method achieves an overall improvement in system performance by constructing a collaborative optimization framework for optical design and digital image processing. At the optical design level, this method overcomes the limitations of traditional methods that overly rely on hardware correction. By rationally allocating the optimization weights of the optical system and digital algorithm, it achieves better image quality while maintaining or reducing system complexity. Especially in scenarios that are difficult to handle with traditional designs, such as short focal lengths, the block processing strategy based on spatially variable PSF and the gradual fusion technology effectively improve the image restoration effect, significantly improving key indicators such as the mean square error, peak signal-to-noise ratio, and structural similarity of the restored image.

[0015] The end-to-end optimization platform built by this system boasts outstanding versatility and scalability. By integrating the Zemax and Matlab development environments through VS Code, it not only achieves seamless integration of optical simulation and image processing but also provides flexible configuration options for different initial structures and application scenarios. The Lucy-Richardson algorithm, after targeted optimization and combined with precise brightness adjustment and boundary processing mechanisms, significantly improves the algorithm's application performance in zoom systems. Experimental results show that the four-element zoom system optimized using this method achieves significantly better image quality than traditional design methods, especially in key areas such as edge fields of view and short focal lengths. This innovative design pattern provides a new technical path for optical-digital co-optimization in computational imaging, effectively improving the overall imaging performance of zoom systems while ensuring system compactness and economy. Attached Figure Description

[0016] Figure 1 A flowchart for image restoration after traditional optical design;

[0017] Figure 2 This is a flowchart of a zoom imaging system design method for optical-digital collaborative optimization according to the present invention;

[0018] Figure 3 A two-dimensional diagram of a traditional optical four-element zoom system.

[0019] Figure 4 A two-dimensional structural diagram of the four-element zoom system with optical-digital co-design proposed in this invention;

[0020] Figure 5 The image used in the simulation example of this invention is the one used to simulate the target to be imaged;

[0021] Figure 6 To restore the image from the telephoto lens in traditional optical design;

[0022] Figure 7 This is a reconstruction image of the central focal point in traditional optical design;

[0023] Figure 8 To restore images from short focal lengths in traditional optical designs;

[0024] Figure 9 The image restored by telephoto lens in the optical-digital joint design proposed in this invention;

[0025] Figure 10 This is the restored image of the central focal point in the optical-digital joint design proposed in this invention;

[0026] Figure 11 This is the restored image of the short focal length in the optical-digital joint design proposed in this invention;

[0027] Figure 12 The MSE curves are shown for comparing the present invention with those of traditional optical design methods.

[0028] Figure 13 The PSNR comparison curves are shown for this invention and traditional optical design methods.

[0029] Figure 14 The SSIM comparison curves are shown for this invention and traditional optical design methods. Detailed Implementation

[0030] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0031] like Figure 2 As shown, this invention proposes a design method for a zoom imaging system based on optical-digital collaborative optimization, comprising a closed-loop optimization architecture consisting of an optical design system and a digital image processing system. The optical design system utilizes Zemax; the digital image processing system is based on the Matlab image processing module. Both systems form an end-to-end bidirectional data channel through a VS Code API interface. The optical design system provides the blurred image and the system's point spread function for image restoration, while Matlab provides objective metrics corresponding to the restored image, such as mean square error (MSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM), to guide optical system optimization. The method specifically includes:

[0032] Step 1: Optical simulation; The optical system outputs a blurred image of the initial optical structure and a spatially variable PSF dataset through Zemax;

[0033] The system loads the initial optical structure in Zemax, performs optical simulation on the input calibration image, and outputs a blurred image of the current structure and the corresponding spatially variable point spread function (PSF) dataset.

[0034] Step 2, Image Processing; In Matlab, perform image block restoration processing on the blurred image and PSF dataset to obtain the restored image. The image block restoration processing includes boundary expansion compensation and linear gain method merging to eliminate artifacts at the stitching boundaries.

[0035] The Matlab digital image processing system executes a block-based restoration algorithm. First, the input blurred image undergoes polynomial distortion correction preprocessing. Then, considering the spatial variation characteristics of the optical system, the preprocessed blurred image and the PSF dataset are divided into n×n processing blocks, simplifying the complex spatially variable deconvolution problem into multiple spatially invariant deconvolution problems. Next, an improved Lucy-Richardson algorithm is used for deblurring the RGB channels. Finally, a binary search method is used to find the optimal image brightness to obtain the restored image. The image block segmentation employs an extended boundary processing technique: first, b pixels are extended outside the original block boundary to form an overlapping area, then a pixels are extended (i.e., ab pixels are extended again) to compensate for boundary pixel clipping. Finally, a linear gain method is used for merging. This method eliminates the significant deviations in the restoration results on both sides of the seam caused by the inconsistency in the use of PSF on both sides during deconvolution processing of adjacent blocks, which manifests as abrupt changes in image intensity values ​​and ultimately forms obvious artifact traces at the stitching boundary.

[0036] Step 3: Fusion Feedback; Matlab compares the image evaluation metrics of the restored image with those of the initial optical structure. If the comparison result does not meet the design specifications, the comparison result is fed back to the optical system. Steps 1-3 are executed again with the restored image as the initial optical structure to form a closed-loop optimization until the imaging quality requirements are met.

[0037] After image restoration in Matlab, objective evaluation metrics such as MSE, PSNR, and SSIM are obtained by comparing the restored image with the original image. These metrics are then fed back to Zemax via the API interface in the form of User Operands (UDOC). Zemax then optimizes the current optical system using damped least squares or orthogonal descent methods based on these metrics until the system meets the requirements.

[0038] like Figure 3 , Figure 4 As shown, a new system was obtained after using the optical-digital joint design zoom system design method of the present invention. While the imaging quality of the optical system is slightly inferior to the original system, the reconstruction structure is superior (as shown in Tables 1 and 2). Figure 3 A 3D diagram of a traditional optical four-element zoom system. Figure 4 Table 1 shows the three-dimensional configuration of the four-element zoom system with optical-digital co-design proposed in this invention. Figure 3 The performance parameters of the system shown are corresponding to those in Table 2. Figure 4The table shows the system's performance parameters. The optical system section refers to the objective evaluation metrics of the image output by the optical system, namely mean squared error (MSE), peak signal-to-noise ratio (PSNR), and structural similarity (SSIM). The restored image section refers to the objective evaluation metrics of the image obtained after Matlab image processing.

[0039] Table 1

[0040]

[0041] Table 2

[0042]

[0043] See Figures 5-12 , Figure 5 The image used in the simulation example of this invention is the one used to simulate the target to be imaged; Figure 6 , Figure 7 , Figure 8 These are the restored images from traditional optical designs at telephoto, medium telephoto, and short telephoto focal lengths, respectively; for comparison. Figure 9 , Figure 10 , Figure 11 The images shown are the restored images of telephoto, mid-range, and short-range focal lengths obtained based on the optical-digital collaborative optimization zoom imaging system design method of this invention. It can be seen that, in terms of the telephoto restored image, its mean square error, peak signal-to-noise ratio, and structural similarity are all better than the original system; in terms of the mid-range restored image, its mean square error, peak signal-to-noise ratio, and structural similarity are comparable to the original system; and in terms of the short-range restored image, its mean square error, peak signal-to-noise ratio, and structural similarity are all better than the original system, and the image quality is significantly improved.

[0044] Figure 12 , Figure 13 , Figure 14 The figures show the MSE, PSNR, and SSIM comparison curves for the design methods of this invention (new system) and traditional optics (old system). The optical-digital joint zoom system slightly balances the three structures, ensuring that each structure achieves good quality after image restoration. Specifically, it reduces MSE by 60%, increases PSNR by 20%, and improves SSIM by 2% in the short, medium, and long focal lengths; reduces MSE by 15%, increases PSNR by 5%, and improves SSIM by 0.8% in the medium focal length; and reduces MSE by 30%, increases PSNR by 10%, and improves SSIM by 1%. It can be seen that this invention achieves good computational accuracy in all aspects.

[0045] The present invention also provides an electronic device, comprising: one or more processors; and a memory for storing one or more programs; wherein, when the one or more programs are executed by the one or more processors, the one or more processors implement the aforementioned optical-digital co-optimized zoom imaging system design method.

[0046] The present invention also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enable the processor to implement the aforementioned optical-digital co-optimized zoom imaging system design method.

[0047] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A design method for a zoom imaging system with optical-digital collaborative optimization, characterized in that, The method includes: Step 1: The optical system outputs a blurred image of the initial optical structure and a spatially variable PSF dataset through Zemax; Step 2: In Matlab, perform image block restoration processing on the blurred image and PSF dataset to obtain the restored image. The image block restoration processing includes boundary expansion compensation and linear gain method merging to eliminate artifacts at the stitching boundaries. Step 3: Matlab compares the image evaluation metrics of the restored image with those of the initial optical structure. If the comparison result does not meet the design specifications, the comparison result is fed back to the optical system. Steps 1-3 are executed again with the restored image as the initial optical structure to form a closed-loop optimization until the imaging quality requirements are met. Step 2 includes: The input blurred image is preprocessed with polynomial distortion correction. The preprocessed blurred image and PSF dataset are divided into n×n processing blocks. The improved Lucy-Richardson algorithm is used to deblur the RGB three channels respectively. The optimal image brightness is found using the binary search method to obtain the restored image. The division into n×n processing blocks adopts the boundary extension processing technique. First, b pixels are extended outside the original block boundary to form an overlapping area, and then a pixels are extended, where a>b, to compensate for the clipping of boundary pixels. Finally, the linear gain method is used for merging.

2. The optical-digital collaborative optimization zoom imaging system design method according to claim 1, characterized in that, The Zemax and Matlab form an end-to-end bidirectional data channel through the VS Code API interface.

3. The optical-digital collaborative optimization zoom imaging system design method according to claim 1, characterized in that, In step 3, the optical system optimization adopts either the damped least squares method or the orthogonal descent method.

4. The optical-digital collaborative optimization zoom imaging system design method according to claim 1, characterized in that, The image evaluation metrics in step 3 include mean squared error, peak signal-to-noise ratio, and structural similarity.

5. The optical-digital collaborative optimization zoom imaging system design method according to claim 1, characterized in that, In step 3, the image evaluation metrics are fed back to Zemax in the form of user operands via the API interface.

6. The optical-digital collaborative optimization zoom imaging system design method according to claim 1, characterized in that, The method is applicable to short-focal-length, medium-focal-length, and long-focal-length scenarios.

7. An electronic device, characterized in that, include: One or more processors; Memory, used to store one or more programs; When one or more programs are executed by the one or more processors, the one or more processors implement the optical-digital collaborative optimization zoom imaging system design method according to any one of claims 1-6.

8. A computer-readable storage medium, characterized in that, It stores executable instructions that, when executed by a processor, enable the processor to implement the optical-digital co-optimized zoom imaging system design method as described in any one of claims 1-6.

Images