S0 image brightness enhancement method and system based on polarization imaging mechanism

Abstract

The invention belongs to the technical field of photoelectric imaging and image processing, and discloses an S0 image brightness enhancement method and system based on a polarization imaging mechanism. The method comprises the following steps: performing polarization demosaicing and intensity denoising on a selected polarization image through a preprocessing module to obtain a polarization intensity image; and inputting the polarization intensity image to a parameter optimization weighted enhancement module for weight parameter optimization and weighted summation enhancement fusion to obtain a brightness enhanced S0 result image. According to the invention, under the condition of maintaining the integrity of polarization information in the S0 image, the problem that the S0 image is dark and weak due to a polarization imaging mechanism is effectively solved. By improving the brightness of the S0 image, the identification degree of the target object and the overall definition of the image are improved, more visual and accurate visual information is provided for a user, and the problem that the brightness of the S0 image is insufficient due to limitation of an imaging mechanism of an existing polarization camera is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of photoelectric imaging and image processing, and particularly relates to a method for performing S 0 Image brightness enhancement method and system. Background Art

[0002] As a special imaging method, polarization imaging technology has shown unique advantages in many fields such as environmental monitoring, medical diagnosis, and material analysis. It can provide information about the polarization state of light in the scene, which is essential for tasks such as identifying surface characteristics of objects and detecting hidden features. However, due to the influence of the polarization imaging mechanism, polarization images often have a low brightness problem, which seriously affects the information extraction effect of the image.

[0003] Traditional polarization image brightness enhancement methods include but are not limited to histogram equalization, contrast stretching, and gamma correction. Although these methods can improve the visibility and contrast of images in some cases, they are not ideal for dark images caused by the polarization imaging mechanism. On the one hand, these traditional methods are difficult to accurately assign appropriate weight values ​​to polarization images at different angles; on the other hand, traditional methods may introduce additional computational complexity and longer processing time, which is particularly insufficient in real-time applications.

[0004] In recent years, with the development of artificial intelligence technology, swarm intelligence optimization algorithms (such as ant colony optimization, particle swarm optimization, etc.) have received more and more attention in the field of image processing due to their global optimization capabilities and high adaptability. These algorithms imitate the group behavior patterns in nature, can effectively search the solution space, and automatically find the optimal or near-optimal solution. However, in the field of polarization imaging, how to use swarm intelligence algorithms to solve the problem of insufficient brightness of polarization images is still a topic that needs to be studied in depth.

[0005] In view of the limitations of the existing technology, it is urgent to develop new methods and technologies to solve the problem of insufficient image brightness caused by the polarization imaging mechanism. 0 The image brightness enhancement method and system aims to improve the brightness level of polarized images through an intelligent weight selection strategy, thereby improving image quality and information extraction efficiency. Through this method, the present invention hopes to overcome the challenges faced by traditional technologies and provide a more efficient and accurate polarized image brightness enhancement solution. Summary of the invention

[0006] In order to overcome the deficiencies in the prior art, the embodiments disclosed in the present invention provide a method for imaging the image of a polarization-based image. 0The invention relates to an image brightness enhancement method and system. Specifically, it relates to an S-type image enhancement method and system based on a polarization imaging mechanism combined with an improved horned lizard optimization algorithm. 0 Image brightness enhancement method. The present invention aims to solve the problem of S 0 The problem of image brightness enhancement technology. By integrating polarization demosaicing, intensity denoising, swarm intelligence optimization algorithm and neural network technology, the present invention can effectively process the S caused by the polarization imaging mechanism without losing polarization information. 0 The image is dim and weak, which is significantly improved. 0 The recognition and overall clarity of the target in the image are improved, thereby providing users with more intuitive and accurate image information. In addition, the method of the present invention has been tested and verified on the public dataset RCPI, and the results show its effectiveness.

[0007] The technical solution is as follows: a method based on a polarization imaging mechanism 0 The image brightness enhancement method includes:

[0008] S1: performing polarization demosaicing and intensity denoising on the selected polarization image through a preprocessing module (1) to obtain a polarization intensity image;

[0009] S2: Input the polarization intensity image into the parameter optimization weighted enhancement module (2) to perform weight parameter optimization and weighted sum enhancement fusion to obtain the brightness enhanced S 0 Result image.

[0010] In step S1, in the acquired polarization intensity image, the polarization intensity image includes intensity images of any polarization direction from 0° to 360°, and the intensity images of any polarization direction are collectively referred to as I 0 .

[0011] The polarization intensity image comprises:

[0012] S101: Polarization image I 0 The image is input to the polarization demosaicing module (101) for polarization image demosaicing to generate an intensity image I containing any polarization direction. 11 ;

[0013] S102: transform the intensity image I of any polarization direction 11 Input to the intensity denoising module (102) for initial denoising to obtain the polarization intensity image I 1 .

[0014] In step S2, the brightness enhancement S is obtained. 0 The resulting images include:

[0015] S201: Utilize the improved horned lizard optimization algorithm module (201) to optimize the weighted sum enhancement fusion module (202) and output the optimal parameter intensity weight coefficient w 1 ,w 2 ,w 3 ,w 4 ;

[0016] S202: According to the optimal parameter strength weight coefficient w 1 ,w 2 ,w 3 ,w 4 Set the polarization intensity image I 1 Input to the optimal weighted sum enhancement fusion module (202) to achieve S after brightness enhancement 0 Result Image I 2 .

[0017] Furthermore, in step S201, the improved horned lizard optimization algorithm module (201) searches iteratively and continuously adjusts the weight coefficients, and finally finds the optimal weight coefficient combination in the complex search space, thereby significantly improving S 0 The weighted summation enhancement fusion module (202) takes the polarization intensity image as input, integrates the intensity information in multiple polarization directions, performs weighted summation according to the optimized weight coefficient, and generates a brightness enhanced S 0 Result Image I 2 ;

[0018] The specific steps of the improved horned lizard optimization algorithm module (201) are as follows:

[0019] S2011: Algorithm parameter initialization: The present invention uses Logistic Tent chaotic mapping to initialize the initial parameters of the HLOA optimization algorithm, including: population size N, maximum number of iterations T max , specify the number K of search agents.

[0020] Among them, because the horned lizard optimization algorithm is the same as most intelligent algorithms, its initial velocity population is randomly generated in the search space. This random generation mode will cause the diversity of the algorithm population to decrease in the later iteration, and it is easy to fall into the local optimal solution, which will have an adverse effect on the global extreme value search. Therefore, the present invention introduces the Logistic Tent chaotic map to improve the initialization of the HLOA algorithm. Using the Logistic Tent chaotic map instead of random parameters allows the algorithm to obtain an initial solution with good diversity in the search space, effectively improving the convergence speed and solution accuracy of the HLOA algorithm. Its formula is shown in (1):

[0021]

[0022] Among them, r∈(0,4), the chaotic orbit state value range is (0, 1).

[0023] S2012: Construct an objective function to calculate the fitness value. The objective function is:

[0024]

[0025] Among them, J is the objective function, is the calculated S 0 The brightness value of the image, L V(i) is the brightness value of the visible light image corresponding to the pixel position.

[0026] S2013: By simulating the horned lizard's defense strategy, the search agent's position update is carried out, including three strategies:

[0027] Strategy 1: Stealth behavior strategy. When the horned lizard is attacked and in camouflage, the horned lizard adopts a stealth behavior strategy to update the position of the horned lizard search agent. The update formula is shown in (3):

[0028]

[0029] in, represents the proxy position of the i-th horned lizard in the search space of the t+1 generation; represents the optimal solution with the highest fitness at the tth iteration; α is a constant with a value of 2; T max Indicates the maximum number of iterations of the population; r 1 , r 2 , r 3 and r 4 is a random integer, ranging from [1 to population size N]; and Represents the rth i The position vector of the search agent in the tth generation; σ represents a binary number, c 1 and c 2 represents a random number obtained from the standard color palette, and c 1 ≠c 2 .

[0030] Strategy 2: Moving escape strategy. When the horned lizard is attacked and not in camouflage, if it chooses to escape, it will move quickly and randomly to escape the predator. The moving escape strategy is used to update the position of the horned lizard search agent. The update formula is shown in (4):

[0031]

[0032] in, represents the proxy position of the i-th horned lizard in the search space of the t+1 generation; represents the optimal solution with the highest fitness at the tth iteration; w represents a random number in the range of [-1, 1]; Represents random numbers generated from a Cauchy distribution with mean 0 and variance 1.

[0033] Strategy 3: Improved blood spraying strategy. When the horned lizard is attacked and is not in camouflage, if it does not choose to escape, it will adopt an aggressive strategy to resist the enemy by spraying blood from its eyes.

[0034] In view of the shortcoming that the initial velocity of blood spraying of horned lizards is fixed in the original blood spraying strategy, the present invention designs a blood spraying strategy by introducing optimized inertia weight. The improved blood spraying strategy can adjust the initial velocity of the spraying according to the number of iterations of the horned lizard optimization algorithm and the fitness of the individual. The improved blood spraying formula is shown in (5):

[0035]

[0036] in, V represents the proxy position of the i-th horned lizard in the search space of the t+1 generation; 0 represents the initial velocity; γ is the angle at which the horned lizard ejects blood, and its value is T max Indicates the maximum number of iterations of the population; Take 1E-6; represents the optimal solution with the highest fitness at the tth iteration; g represents the earth's gravity, which is 0.009807km / s 2 ; represents the agent position of the i-th horned lizard in the t-th generation search space.

[0037] V 0 The formula is shown in (6):

[0038]

[0039] Among them, V 0 represents the initial velocity; Indicates the weight of the current search agent; Indicates the speed of the current search agent; w star and w end Represent the initial weight and final weight respectively; T max represents the maximum number of iterations of the population; t represents the current number of iterations; κ and ξ represent the design control coefficients; f max and f min Respectively represent the global worst fitness value and the global optimal fitness value; f i Represents the current agent's fitness value.

[0040] S2014: By simulating the adjustment strategy of the horned lizard's skin becoming brighter or darker, the worst position of the search agent is updated, that is, if the horned lizard's skin becomes brighter, the formula (7) can be used for update and replacement; if the horned lizard's skin becomes darker, the formula (8) can be used for update and replacement.

[0041]

[0042] in, Indicates the solution with the worst fitness at the tth iteration; represents the optimal solution with the highest fitness at the tth iteration; and Represents the rth i The position vector of the search agent in the tth generation; Λ represents a random number in binary; L 1 and L 2 is a random number, and its value range is [0, 0.4046661]; D 1 and D 2 is a random number, and its value range is [0.5440510, 1].

[0043] S2015: By observing the change of the α-melanocyte rate value of the horned lizard on both sides of the color of the horned lizard skin in S504, if the α-melanocyte rate value of the horned lizard is lower than 0.3, it is necessary to replace the search agent with a low value and select the best search agent value, as shown in formula (9):

[0044]

[0045] in, represents the proxy position of the i-th horned lizard in the search space of generation t; represents the optimal solution with the highest fitness at the tth iteration; and Represents the rth i The position vector of the search agents in generation t; μ represents a binary number.

[0046] S2016: Determine whether it is at the maximum number of iterations. If so, output the optimal parameter strength weight coefficient; if not, return to S502 to continue the optimization operation.

[0047] Furthermore, the weighted sum enhancement fusion module (202) uses the optimized weight coefficient to perform weighted summation on the input polarization intensity image. Specifically, the corresponding weight coefficient is multiplied in each polarization direction, and then all weighted intensity values ​​are added to obtain the brightness enhanced S 0 Result Image I 2 The formula is as follows:

[0048]

[0049] Among them, I x° , I y° , I z° and I k° represents the intensity image of any polarization direction; w 1 , w 2 , w 3 and w 4 Represents the weight value of each intensity image. Different weight values ​​represent different importance.

[0050] Another object of the present invention is to provide a method for imaging the polarization of the 0 An image brightness enhancement system, the system implementing the polarization-based imaging mechanism 0 An image brightness enhancement method, the system comprising:

[0051] A preprocessing module, used for performing polarization demosaicing and intensity denoising on the selected polarization image to obtain a polarization intensity image;

[0052] Furthermore, the polarization demosaicing module is used to transform the polarization image I 0 Perform polarization image demosaicing to generate an intensity image I containing any polarization direction 11 ;

[0053] The intensity denoising module is used to transform the intensity image I 11 Perform initial denoising to obtain the polarization intensity image I 1 .

[0054] The parameter optimization weighted enhancement module uses an improved horned lizard optimization algorithm. The algorithm searches for the optimal combination of weight coefficients in a complex search space by iterative search and continuous adjustment of weight coefficients. The polarization intensity image is provided as input to this module, which integrates the intensity information in multiple polarization directions. Based on the optimized weight coefficients, the module performs a weighted summation operation to generate the brightness enhanced S 0 Resulting image. This process is called weighted sum fusion and is designed to significantly improve image brightness and quality without losing polarization information.

[0055] Furthermore, the improved horned lizard optimization module (201) is used to generate the optimal parameter intensity weight coefficient w of the optimized weighted sum enhancement fusion module. 1 ,w 2 ,w 3 ,w 4 ;

[0056] The weighted summation enhanced fusion module (202) is used to calculate the optimal parameter intensity weight coefficient w1 ,w 2 ,w 3 ,w 4 , the polarization intensity image I 1 Perform weighted summation to generate brightness enhanced S 0 Result Image I 2 .

[0057] Furthermore, the low-light color polarization image enhancement system is deployed on a computer device based on a polarization imaging mechanism. The device includes at least one processor, a memory, and a dedicated computer program stored in the memory and running on the processor. When the processor executes these programs, the various functions of the low-light color polarization image enhancement system can be realized.

[0058] Based on the polarization imaging mechanism, S 0 The image enhancement system is mounted on a computer device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor executes the computer program to realize the above-mentioned polarization imaging mechanism. 0 Function of the image brightness enhancement system.

[0059] In combination with all the above technical solutions, the beneficial effects of the present invention are as follows: the present invention provides an S-type imaging system based on a polarization imaging mechanism. 0 The image brightness enhancement method and system are as follows. The method pre-processes the selected polarization image, including polarization demosaicing and intensity denoising, so as to obtain a polarization intensity image; then uses the improved horned lizard optimization algorithm module to optimize the weighted summation enhancement fusion module, outputs the optimal parameter intensity weight coefficient, and then performs weighted summation on the polarization intensity image according to these coefficients, finally realizing the brightness enhanced S 0 The present invention can not only 0 Based on the polarization information in the image, the S 0 The problem of image dimness has also been greatly improved, and the recognition of target objects and overall clarity of the image have been greatly improved, providing observers with more intuitive and accurate image information. Compared with the existing polarization image brightness enhancement method, the present invention significantly improves the image brightness level and image quality through an intelligent weight selection strategy, thereby improving the efficiency of information extraction. It is particularly suitable for environmental monitoring, remote sensing imaging, medical imaging, machine vision and autonomous driving, and agricultural monitoring, and provides strong technical support for applications in related fields.

[0060] Compared with the prior art, the advantages of the present invention further include: first, the present invention can effectively remove various interference factors in the original polarization image, such as mosaic effect and random noise, by adopting advanced preprocessing technologies such as polarization demosaicing and intensity denoising, so as to obtain a purer and higher-quality polarization intensity image, which lays a solid foundation for subsequent image enhancement work. Secondly, the present invention originality proposes an improved horned lizard optimization algorithm and applies it to the parameter optimization weighted enhancement module. This algorithm can not only automatically find the best weight parameter configuration, but also flexibly adjust the optimization strategy according to the characteristics of different images to ensure that each enhancement can achieve the desired effect. Finally, the weighted summation enhancement fusion module of the present invention can make full use of the optimized weight parameters to perform meticulous enhancement processing on the polarization intensity image. It can not only significantly improve the overall brightness of the image, but also highlight the outline and texture details of the target object without affecting the original polarization information, making the image more ornamental and practical.

[0061] In summary, the ultimate goal of the present invention is to maximize the retention of polarization information in the original image while enhancing the brightness of the polarization image to prevent information loss. In addition, the present invention uses an improved optimization algorithm to flexibly adjust the weight coefficient to adapt to different types of polarization images. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The accompanying drawings herein are incorporated in and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the description, serve to explain the principles of the present disclosure;

[0063] Figure 1 is S under the polarization imaging mechanism provided by the embodiment of the present invention 0 Flowchart of image brightness enhancement method;

[0064] Figure 2 is S under the polarization imaging mechanism provided by the embodiment of the present invention 0 Schematic diagram of image brightness enhancement system;

[0065] Figure 3 It is a schematic diagram of the structural principle of the improved horned lizard optimization algorithm module provided by an embodiment of the present invention.

[0066] In the figure: 1. Preprocessing module; 101. Polarization demosaicing module; 102. Intensity denoising module; 2. Parameter optimization weighted enhancement module; 201. Improved horned lizard optimization algorithm module; 202. Weighted sum enhancement fusion module. DETAILED DESCRIPTION

[0067] To make the purpose, features and advantages of the present invention clearer and easier to understand, specific embodiments will be described in detail below in conjunction with the accompanying drawings. In the following description, many specific details are provided to help a comprehensive understanding of the present invention. However, the present invention can be implemented in a variety of ways different from those described herein, and those skilled in the art can make similar improvements without departing from the core principles of the present invention. Therefore, the scope of the present invention is not limited by the following specific embodiments.

[0068] The innovation of the present invention is: to solve the current S based on polarization imaging mechanism 0 The image brightness enhancement method has the problem of insufficient brightness. A method based on the polarization imaging mechanism S is proposed, which combines the preprocessing module 1 and the parameter optimization weighted enhancement module 2. 0 Image brightness enhancement method. The preprocessing module 1 uses the polarization demosaicing module 101 to demosaic the polarization image; the intensity denoising module 102 is used to perform initial denoising; the parameter optimization weighted enhancement module 2 uses the improved optimized horned lizard optimization algorithm 201 to optimize the weighted sum enhancement fusion module to obtain the optimal parameter intensity weight coefficient; the weighted sum enhancement fusion module 202 is used to achieve the brightness enhancement of S according to the optimal parameter weight coefficient. 0 Result image.

[0069] Embodiment 1, as Figure 1 As shown, an embodiment of the present invention provides a method for enhancing the brightness of an S0 image based on a polarization imaging mechanism, the method comprising de-mosaicing and denoising and a swarm intelligence optimization algorithm, the method comprising a pre-processing module 1 and a parameter optimization weighted enhancement module 2; the method comprises the following steps:

[0070] S1: performing polarization demosaicing and intensity denoising on the selected polarization image through a preprocessing module (1) to obtain a polarization intensity image;

[0071] S2: Input the polarization intensity image into the parameter optimization weighted enhancement module (2) to perform weight parameter optimization and weighted sum enhancement fusion to obtain the brightness enhanced S 0 Result image.

[0072] In step S1, in the acquired polarization intensity image, the polarization intensity image includes intensity images of any polarization direction from 0° to 360°, and the intensity images of any polarization direction are collectively referred to as I 0 .

[0073] As a preferred embodiment, the intensity image of any polarization direction includes polarization states in four directions: 0°, 45°, 90° and 135°, collectively referred to as I 0 .

[0074] The polarization intensity image comprises the following steps:

[0075] S101: Polarization image I 0 The image is input to the polarization demosaicing module (101) for polarization image demosaicing to generate an intensity image I containing any polarization direction. 11 ;

[0076] In an embodiment, as a preferred method, the polarization demosaicing module used in the present invention adopts the EARI algorithm with relatively good demosaicing quality at present;

[0077] S102: transform the intensity image I of any polarization direction 11 Input to the intensity denoising module (102) for initial denoising to obtain the polarization intensity image I 1 .

[0078] In this embodiment, as a preferred method, the present invention uses a median filtering method to perform polarization intensity image denoising;

[0079] The present invention adopts the technical means of demosaicing and intensity denoising to more effectively process S 0 Image brightness is enhanced, S 0 Image brightness problem. Polarization demosaicing is used to generate an intensity image I containing any polarization direction. 11 ,Through intensity denoising, the intensity image of any polarization direction is initially denoised to obtain the polarization intensity image.

[0080] In step S2 of the embodiment of the present invention, Figure 2 As shown, the parameter optimization weighted enhancement module 2 specifically includes an improved horned lizard optimization algorithm module 201 and a weighted sum enhancement fusion module 202;

[0081] The improved horned lizard optimization module (201) is used to generate the optimal parameter intensity weight coefficient w of the optimized weighted sum enhancement fusion module 1 ,w 2 ,w 3 ,w 4 ;

[0082] The weighted summation enhanced fusion module (202) is used to calculate the optimal parameter intensity weight coefficient w 1 ,w 2 ,w 3 ,w 4 , the polarization intensity image I 1 Perform weighted summation to generate brightness enhanced S 0 Result Image I 2 .

[0083] The polarization intensity image is input into the parameter optimization weighted enhancement module 2 for weighted enhancement fusion to obtain the brightness enhanced S0 The result image, the specific steps are as follows:

[0084] S201: Utilize the improved horned lizard optimization algorithm module 201 to optimize the weighted sum enhancement fusion module 202, and output the optimal parameter intensity weight coefficient w 1 ,w 2 ,w 3 ,w 4 ;

[0085] It can be understood that the improved horned lizard optimization algorithm module 201 used in the present invention finds the optimal weight coefficient combination in the complex search space through iterative search and continuous adjustment of the weight coefficient, thereby significantly improving S 0 The brightness and clarity of the image;

[0086] S202: According to the optimal parameter strength weight coefficient w 1 ,w 2 ,w 3 ,w 4 Set the polarization intensity image I 1 Input to the optimal weighted sum enhancement fusion module 202 to achieve S after brightness enhancement 0 Result Image I 2 .

[0087] It can be understood that the weighted summation enhancement fusion module 202 used in the present invention takes the polarization intensity image as input, and integrates the intensity information in multiple polarization directions, performs weighted summation according to the optimized weight coefficient, and generates the brightness enhanced S 0 Result Image I 2 ;

[0088] Example 2, illustratively, as an implementation mode, in the example of the present invention, step S2 includes:

[0089] like Figure 3 As shown in the figure, it is a schematic diagram of the module structure of the improved horned lizard optimization algorithm. The specific steps of the process are as follows:

[0090] The present invention uses Logistic Tent chaotic mapping to initialize the initial parameters of the HLOA optimization algorithm, including: population size N, maximum number of iterations T max, specify the number of search agents K; construct an objective function for calculating the fitness value; update the position of the search agent by simulating the defense strategy of the horned lizard; update the worst position of the search agent by simulating the brightening or darkening adjustment strategy of the horned lizard's skin; observe the light and dark changes of the horned lizard's α-melanocyte rate value on the horned lizard's skin. If the α-melanocyte rate value of the horned lizard is lower than 0.3, it is necessary to replace the search agent at a low value and select the best search agent value; determine whether it is at the maximum number of iterations. If so, output the optimal parameter strength weight coefficient; if not, return to continue the optimization operation. In this embodiment of the present invention, the population size N = 50, the maximum number of iterations T max =200, the number of specified search agents K = 10;

[0091] In step S2 of this embodiment, the parameter optimization weighted enhancement module 202 of the present invention also includes a weighted sum enhancement fusion module 202. The input polarization intensity image is weighted summed using the optimized weight coefficients. Specifically, the corresponding weight coefficient is multiplied in each polarization direction, and then all weighted intensity values ​​are added to obtain the brightness enhanced S 0 Result image.

[0092] It can be seen from the above embodiments that the present invention also has the following advantages:

[0093] Environmental monitoring: Polarization imaging can be used to monitor environmental parameters such as air quality, water pollution, and vegetation health. Polarization information related to these parameters can be extracted by optimizing weights, providing strong support for environmental protection and disaster warning.

[0094] Remote sensing images: In the field of remote sensing, polarization information can be used to improve the accuracy of tasks such as object classification, vegetation monitoring, and water body detection by optimizing weights to enhance the useful information in polarization images obtained from satellites or drones.

[0095] Medical imaging: In medical imaging, polarization imaging technology can be used to detect abnormal structures or functional changes in tissues and extract clearer pathological features from optimized polarization images to assist diagnosis and treatment.

[0096] Machine vision and autonomous driving: Polarization information helps enhance image contrast, reduce glare, and improve target detection capabilities. The optimized polarization images can be applied to autonomous driving vehicle camera systems to improve the accuracy of road and obstacle recognition.

[0097] Agricultural monitoring: Optimized polarization images can be used for crop pest and disease monitoring, crop growth status assessment, etc. They can highlight abnormal changes in crop leaves and provide technical support for precision agriculture.

[0098] This invention combines polarization demosaicing and intensity denoising technology with parameter optimization weighted enhancement technology for the first time to analyze the S 0 The image is enhanced with a relatively new algorithm to make up for the current S 0 The algorithm of the present invention effectively solves the problem of dark and weak images under the polarization imaging mechanism, and fills the technical gap in the current image brightness enhancement under the polarization imaging mechanism.

[0099] The present invention uses a rich data set covering a variety of scenarios and environments for training and evaluation, which helps ensure that the algorithm has good generalization capabilities, rather than just performing well under specific conditions. In the process of training data, we eliminated factors that may cause technical deviations, such as insufficient ambient light intensity (lux). For the above embodiments, the focus of each example is different; if some content is not described or recorded in detail in an embodiment, it can be found in the relevant description of other embodiments.

[0100] The information interaction, execution process and other contents between the above-mentioned devices are based on the same concept as the embodiment of the method of the present invention. Their specific functions and technical effects can be found in the method implementation part and will not be repeated here.

[0101] Professionals in the relevant field can understand that for the convenience of description and simplified explanation, only the specific division of each functional unit and module is used as an example for explanation. In practical applications, the above functions can be allocated to different functional units or modules as needed to achieve, that is, the internal structure of the device can be divided into different units or modules according to functional requirements to complete all or part of the functions. These functional units or modules can be integrated in a single processing unit, or they can exist as independent physical entities, and the functions of two or more units / modules can be integrated into one unit. This integrated unit can be implemented in hardware form or in the form of software functional units. In addition, the specific naming of each functional unit and module is only to distinguish them, and does not limit the protection scope of the present invention. The specific operation process of each unit and module in the system can refer to the corresponding process in the aforementioned method embodiment.

[0102] The embodiment of the present invention also introduces a computer device, which includes at least one processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor runs these computer programs, it will execute the steps described in the above-mentioned various method embodiments.

[0103] The embodiment of the present invention also provides an information data processing terminal, which can be run on various electronic devices and provide an input interface for users to execute the steps in the above method embodiments. The application of this information data processing terminal is not limited to mobile phones, computers, switches and other devices.

[0104] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiment method can be implemented in the present application by guiding the relevant hardware to complete through a computer program. The computer program can be stored in a computer-readable storage medium, and when these programs are executed by a processor, the steps of the above-mentioned various method embodiments can be implemented. The computer program mentioned here includes computer program code, which can be one of source code, object code, executable file or some intermediate form. Computer-readable media may include any entity or device that can carry computer program code to a target device (such as a camera or terminal), such as a recording medium, a computer memory, a ROM (read-only memory), a RAM (random access memory), an electrical signal, a software distribution medium, etc. Specific examples include USB flash drives, mobile hard drives, magnetic disks, optical disks, etc.

[0105] To further illustrate the relevant effects of the embodiments of the present invention, the following experiments are conducted: the present invention selects the 2024 public dataset RCPI, and adopts the polarization demosaicing and intensity denoising modules to better demosaic and denoise the polarization image; the improved horned lizard optimization algorithm module is used to generate the optimal parameter intensity weight coefficient w of the optimized weighted sum enhancement fusion module. 1 ,w 2 ,w 3 ,w 4 The weighted summation enhanced fusion module is used to obtain the optimal parameter intensity weight coefficient w 1 ,w 2 ,w 3 ,w 4 , and the polarization intensity image are weighted and summed to generate the brightness enhanced S 0 Result image.

[0106] Table 1 Comparison results of different optimization algorithms

[0107]

[0108] The above content describes the specific implementation of the present invention, but the protection scope of the present invention is not limited thereto. Personnel familiar with the technical field should understand that within the technical scope disclosed by the present invention, all modifications, equivalent substitutions and improvements within the spirit and basic principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for enhancing the brightness of an S0 image based on a polarization imaging mechanism, characterized in that: The method includes: S1: performing polarization demosaicing and intensity denoising on the selected polarization image through a preprocessing module (1) to obtain a polarization intensity image; S2: Input the polarization intensity image into the parameter optimization weighted enhancement module (2) to perform weight parameter optimization and weighted sum enhancement fusion to obtain the S0 result image after brightness enhancement. In step S1 , in the acquired polarization intensity image, the polarization intensity image includes intensity images in any polarization direction from 0° to 360°, and the intensity images in any polarization direction are collectively referred to as I0. The polarization intensity image comprises: S101: Input the polarization image I0 to the polarization demosaicing module (101) to perform polarization image demosaicing to generate an intensity image I containing any polarization direction. 11 ; S102: transform the intensity image I of any polarization direction 11 The image is input to the intensity denoising module (102) for initial denoising to obtain a polarization intensity image I1. In step S2, obtaining the brightness enhanced S0 result image includes: S201: Utilizing the improved horned lizard optimization algorithm module (201) to optimize the weighted sum enhancement fusion module (202), and outputting the optimal parameter intensity weight coefficients w1, w2, w3, w4; S202: According to the optimal parameter intensity weight coefficients w1, w2, w3, w4, the polarization intensity image I1 is input into the optimal weighted sum enhancement fusion module (202) to realize the S0 result image I2 after brightness enhancement. In step S201, the improved horned lizard optimization algorithm module (201) searches iteratively and continuously adjusts the weight coefficients, and finally finds the optimal weight coefficient combination in a complex search space, thereby significantly improving the brightness and clarity of the S0 image; the weighted summation enhancement fusion module (202) takes the polarization intensity image as input, and integrates the intensity information in multiple polarization directions, performs weighted summation according to the optimized weight coefficients, and generates the S0 result image I2 after brightness enhancement.

2. The method for enhancing the brightness of an S0 image based on a polarization imaging mechanism according to claim 1, characterized in that: In step S201, the improved horned lizard optimization algorithm module (201) specifically comprises the following steps: S2011: Algorithm parameter initialization: The present invention uses Logistic Tent chaotic mapping to initialize the initial parameters of the HLOA optimization algorithm, including: population size N, maximum number of iterations T max , specify the number K of search agents. Among them, because the horned lizard optimization algorithm is the same as most intelligent algorithms, its initial velocity population is randomly generated in the search space. This random generation mode will cause the diversity of the algorithm population to decrease in the later iteration, and it is easy to fall into the local optimal solution, which will have an adverse effect on the global extreme value search. Therefore, the present invention introduces the Logistic Tent chaotic map to improve the initialization of the HLOA algorithm. Using the Logistic Tent chaotic map instead of random parameters allows the algorithm to obtain an initial solution with good diversity in the search space, effectively improving the convergence speed and solution accuracy of the HLOA algorithm. Its formula is shown in (1): Among them, r∈(0,4), the chaotic orbit state value range is (0, 1). S2012: Construct an objective function to calculate the fitness value. The objective function is: Among them, J is the objective function, is the calculated brightness value of the S0 image, L V(i) is the brightness value of the visible light image corresponding to the pixel position. S2013: By simulating the horned lizard's defense strategy, the search agent's position update is carried out, including three strategies: Strategy 1: Stealth behavior strategy. When the horned lizard is attacked and in camouflage, the horned lizard adopts a stealth behavior strategy to update the position of the horned lizard search agent. The update formula is shown in (3): in, represents the proxy position of the i-th horned lizard in the search space of the t+1 generation; represents the optimal solution with the highest fitness at the tth iteration; α is a constant with a value of 2; T max Indicates the maximum number of iterations of the population; r1, r2, r3 and r4 are random integers, ranging from [1 to population size N]; and Represents the rth i The position vector of the search agent in the tth generation; σ represents a binary number, c1 and c2 represent random numbers taken from the standard palette, and c1≠c2. Strategy 2: Moving escape strategy. When the horned lizard is attacked and not in camouflage, if it chooses to escape, it will move quickly and randomly to escape the predator. The moving escape strategy is used to update the position of the horned lizard search agent. The update formula is shown in (4): in, represents the proxy position of the i-th horned lizard in the search space of the t+1 generation; represents the optimal solution with the highest fitness at the tth iteration; w represents a random number in the range of [-1, 1]; Represents random numbers generated from a Cauchy distribution with mean 0 and variance 1. Strategy 3: Improved blood spraying strategy. When the horned lizard is attacked and is not in camouflage, if it does not choose to escape, it will adopt an aggressive strategy to resist the enemy by spraying blood from its eyes. In view of the shortcoming that the initial velocity of blood spraying of horned lizards is fixed in the original blood spraying strategy, the present invention designs a blood spraying strategy by introducing optimized inertia weight. The improved blood spraying strategy can adjust the initial velocity of the spraying according to the number of iterations of the horned lizard optimization algorithm and the fitness of the individual. The improved blood spraying formula is shown in (5): in, represents the proxy position of the ith horned lizard in the t+1 generation search space; V0 represents the initial velocity; γ is the angle at which the horned lizard sprays blood, and its value is T max Indicates the maximum number of iterations of the population; Take 1E-6; represents the optimal solution with the highest fitness at the tth iteration; g represents the earth's gravity, which is 0.009807km / s 2 ; represents the agent position of the i-th horned lizard in the t-th generation search space. The V0 formula is shown in (6): Among them, V0 represents the initial velocity; Indicates the weight of the current search agent; Indicates the speed of the current search agent; w star and w end Represent the initial weight and final weight respectively; T max represents the maximum number of iterations of the population; t represents the current number of iterations; κ and ξ represent the design control coefficients; f max and f min Respectively represent the global worst fitness value and the global optimal fitness value; f i Represents the current agent's fitness value. S2014: By simulating the adjustment strategy of the horned lizard's skin becoming brighter or darker, the worst position of the search agent is updated, that is, if the horned lizard's skin becomes brighter, the formula (7) can be used for update and replacement; if the horned lizard's skin becomes darker, the formula (8) can be used for update and replacement. in, Indicates the solution with the worst fitness at the tth iteration; represents the optimal solution with the highest fitness at the tth iteration; and Represents the rth i is the position vector of the search agent in the tth generation; Λ represents a random number in binary; L1 and L2 are random numbers, both in the range of [0, 0.4046661]; D1 and D2 are random numbers, both in the range of [0.5440510, 1]. S2015: By observing the change of the α-melanocyte rate value of the horned lizard on both sides of the color of the horned lizard skin in S504, if the α-melanocyte rate value of the horned lizard is lower than 0.3, it is necessary to replace the search agent with a low value and select the best search agent value, as shown in formula (9): in, represents the proxy position of the i-th horned lizard in the search space of generation t; represents the optimal solution with the highest fitness at the tth iteration; and Represents the rth i The position vector of the search agents in the tth generation; μ represents a binary number. S2016: Determine whether it is at the maximum number of iterations. If so, output the optimal parameter strength weight coefficient; if not, return to S502 to continue the optimization operation.

3. The method for enhancing the brightness of an S0 image based on a polarization imaging mechanism according to claim 1, characterized in that: In step S202, the weighted sum enhancement fusion module (202) uses the optimized weight coefficient to perform weighted summation on the input polarization intensity image. Specifically, the corresponding weight coefficient is multiplied in each polarization direction, and then all weighted intensity values ​​are added to obtain the brightness enhanced S0 result image I2. The formula is as follows: I2=w1I x +w2I y +w3I z +w4I k (10) Among them, I x , I y , I z and I k represents the intensity image of any polarization direction; w1, w2, w3 and w4 represent the weight value of each intensity image, and different weight values ​​represent different importance.

4. A S0 image brightness enhancement system based on polarization imaging mechanism, characterized in that: The system implements the S0 image brightness enhancement method based on the polarization imaging mechanism as described in any one of claims 1 to 3, and the system includes: A preprocessing module (1) is used to perform polarization demosaicing and intensity denoising on the selected polarization image to obtain a polarization intensity image; The parameter optimization weighted enhancement module (2) is used to perform weight parameter optimization and weighted sum enhancement fusion on the polarization intensity image to obtain an S0 result image after brightness enhancement.

5. The S0 image brightness enhancement system based on polarization imaging mechanism according to claim 4, characterized in that: The preprocessing module (1) comprises: a polarization demosaicing module (101) for demosaicing polarization images and an intensity denoising module (102) for denoising intensity images. The polarization demosaicing module (101) is used to perform polarization image demosaicing processing on the polarization image I0 to generate an intensity image I containing any polarization direction. 11 ; The intensity denoising module (102) is used to convert the intensity image I 11 Perform initial denoising to obtain the polarization intensity image I1.

6. The S0 image brightness enhancement system based on polarization imaging mechanism according to claim 4, characterized in that: The parameter optimization weighted enhancement module (2) comprises: an improved horned lizard optimization algorithm module (201) that eventually finds the optimal weight coefficient combination in a complex search space through iterative search and continuous adjustment of the weight coefficient, and a weighted sum enhancement fusion module (202) that takes the polarization intensity image as input, integrates the intensity information in multiple polarization directions, performs weighted summation according to the optimized weight coefficient, and generates a brightness enhanced S0 result image I3; The improved horned lizard optimization module (201) is used to generate the optimal parameter intensity weight coefficients w1, w2, w3, w4 of the optimized weighted sum enhancement fusion module; The weighted sum enhancement fusion module (202) is used to perform weighted sum processing on the polarization intensity image I1 according to the above-mentioned optimal parameter intensity weight coefficients w1, w2, w3, w4, so as to generate a brightness enhanced S0 result image I2.

Images