Silicon steel plate surface defect image enhancement method and system based on improved mycet algorithm

Abstract

The present application belongs to the technical field of digital image processing and machine vision detection, and particularly is a silicon steel plate surface defect image enhancement method and system based on an improved slime mold algorithm. The method comprises the following steps: S1: acquiring a low-illumination image to be enhanced collected under a silicon steel plate surface defect detection scene; S2: establishing an incomplete Beta function gray transformation model; S3: constructing a fitness function for evaluating the quality of an enhanced image; S4: initializing a slime mold population; S5: iteratively updating the slime mold population based on a slime mold algorithm; S6: introducing a quadratic interpolation strategy in the process of updating the current optimal solution; S7: when a preset termination condition is met; S8: performing gray transformation on the low-illumination image to be enhanced according to the optimal parameter combination. The present application can enhance the distinguishability of a silicon steel plate defect image, can reduce the influence of manual experience parameter selection on the enhancement result, and can provide more stable image input for subsequent defect segmentation, recognition and quality determination.

Description

Technical Field

[0001] This invention relates to the fields of digital image processing and machine vision inspection technology, specifically to a method and system for enhancing surface defect images of silicon steel plates based on an improved slime mold algorithm. Background Technology

[0002] Silicon steel sheets are important functional materials in motors, transformers, and power generation equipment. During rolling, annealing, handling, and inspection, these sheets may exhibit surface abnormalities such as cracks, inclusions, patches, pitting, iron oxide scale, and scratches. Failure to detect these abnormalities in a timely manner can affect the quality of subsequent laminations, the stability of magnetic properties, and the reliability of the finished product.

[0003] Machine vision inspection can complete image acquisition and defect analysis without contacting the surface of the board, making it suitable for deployment on production lines or offline inspection platforms. During actual acquisition, the angle of the ambient light source, board surface reflection, camera exposure time, transmission speed, environmental occlusion, and sensor noise all affect the image quality, causing some defect images to appear as low brightness, low contrast, and weak edges.

[0004] In the defect detection of silicon steel plates, subsequent segmentation, recognition, and classification typically rely on the grayscale difference, texture difference, and contour continuity between the defect area and the substrate area. If the dark texture of the input image is not unfolded, weak features such as fine cracks, shallow scratches, and localized pitting are easily mixed with the background texture, leading to missed detections or misjudgments.

[0005] Existing image enhancement methods include histogram equalization, linear stretching, Gamma transformation, and Retinex models. While these methods can improve image brightness or contrast to some extent, they are prone to problems such as local over-enhancement, noise amplification, edge distortion, or parameter dependence on human experience in low-light defect images of silicon steel plates.

[0006] Incomplete beta functions can alter the grayscale mapping curve through shape parameters, making them suitable for non-linear adjustment of dark, bright, or mid-grayscale areas. However, under different defect types, exposure conditions, and surface textures, fixed parameters struggle to simultaneously meet the requirements of brightness enhancement and detail preservation.

[0007] The slime mold optimization algorithm can be used for continuous parameter optimization, but when directly used for image enhancement parameter search, it may still be affected by uneven distribution of initial candidate solutions, search stagnation in the later stages of iteration, and insufficient refinement ability near the optimal solution, resulting in unstable enhancement parameters.

[0008] Therefore, an enhancement scheme is needed for surface defect images of silicon steel plates, which can adaptively determine grayscale transformation parameters under low illumination, weak contrast or local shadow conditions, while taking into account defect contour preservation, texture detail enhancement and input stability for subsequent detection tasks. Summary of the Invention

[0009] The purpose of this section is to outline some aspects of the embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0010] To address the aforementioned technical problems, according to one aspect of the present invention, the present invention provides the following technical solution:

[0011] A method for enhancing surface defects in silicon steel plates based on an improved slime mold algorithm includes the following steps:

[0012] S1: Acquire the low-light image to be enhanced under the scenario of detecting surface defects in silicon steel plate, and normalize the low-light image to be enhanced to obtain a normalized image.

[0013] S2: Establish an incomplete Beta function grayscale transformation model, and use the shape parameters in the incomplete Beta function grayscale transformation model as variables to be optimized;

[0014] S3: Construct a fitness function for evaluating the quality of the enhanced image, the fitness function including information entropy, edge content, logarithm of edge intensity, and average gradient;

[0015] S4: Initialize the slime mold population using the Singer chaotic mapping strategy to obtain a preset number of initial individuals, each of which corresponds to a set of variables to be optimized;

[0016] S5: The slime mold population is iteratively updated based on the slime mold algorithm, and a hyperbolic oscillation factor is introduced during the iterative update process to dynamically adjust the global search capability and local exploitation capability of the slime mold population.

[0017] S6: In the process of updating the current optimal solution, a quadratic interpolation strategy is introduced to generate candidate approximate optimal solutions, and the candidate approximate optimal solutions compete with the current optimal solution for selection;

[0018] S7: When the preset termination condition is met, output the optimal parameter combination of the incomplete Beta function grayscale transformation model;

[0019] S8: Perform grayscale transformation on the low-light image to be enhanced according to the optimal parameter combination to obtain the enhanced silicon steel plate surface defect image.

[0020] As a preferred embodiment of the silicon steel plate surface defect image enhancement method based on the improved slime mold algorithm described in this invention, in step S1, the silicon steel plate surface defect detection scenario includes at least one of rolling production, offline sampling inspection, or laboratory testing; the low-light image to be enhanced includes at least one defect type image of fine cracks, inclusions, patches, pitting, iron oxide scale rolling, or scratches.

[0021] The specific method of normalization processing is as follows: when the low-light image to be enhanced is a grayscale image, the grayscale pixel values ​​are normalized; when the low-light image to be enhanced is a color image, the color image is converted to the HSV color space, the luminance V channel is extracted and the luminance V channel is normalized, and then converted back to the RGB color space after enhancement.

[0022] As a preferred embodiment of the silicon steel plate surface defect image enhancement method based on the improved slime mold algorithm described in this invention, in step S2, the incomplete Beta function grayscale transformation model is expressed as follows: ,in, u is the normalized original pixel value, F(u) is the transformed pixel value, a and b are the shape parameters, and the values ​​of a and b are both in the range of (0,10).

[0023] In a preferred embodiment of the silicon steel plate surface defect image enhancement method based on the improved slime mold algorithm described in this invention, the fitness function in step S3 is expressed as: Where H is the information entropy of the enhanced image, EC is the edge content of the enhanced image, LSI is the logarithmic sum of the edge intensities of the enhanced image, and MG is the average gradient of the enhanced image. It is a positive control coefficient;

[0024] The edge content EC is determined based on the ratio of the number of edge pixels to the total number of pixels in the image; the edge intensity logarithm and LSI are determined based on the horizontal and vertical gradients obtained from Sobel edge detection; and the average gradient MG is determined based on the grayscale difference between adjacent pixels in the horizontal and vertical directions.

[0025] As a preferred embodiment of the silicon steel plate surface defect image enhancement method based on the improved slime mold algorithm described in this invention, in step S4, the Singer chaotic mapping strategy is expressed as follows: ,in, ∈(0.9,1.08), and map the chaotic sequence generated by the Singer chaotic mapping strategy to the search boundary of the variable to be optimized, so as to determine the position of each initial individual in the slime mold population.

[0026] As a preferred embodiment of the image enhancement method for surface defects of silicon steel plates based on the improved slime mold algorithm described in this invention, in step S5, the hyperbolic oscillation factor is a piecewise function: in the first half of the iteration, the search weight is nonlinearly changed based on the hyperbolic tangent function to expand the exploration range of the parameter space; in the second half of the iteration, local oscillation perturbation is formed based on the sine and cosine functions to enhance the development capability near the optimal solution and reduce the probability of getting trapped in local optima.

[0027] As a preferred embodiment of the image enhancement method for surface defects of silicon steel plates based on the improved slime mold algorithm described in this invention, step S6 includes: selecting two distinct individual positions X and Y and the current global optimal position Z from the slime mold population; constructing a quadratic interpolation model based on the fitness values ​​of X, Y, and Z to obtain a candidate approximate optimal solution R; if the fitness of the candidate approximate optimal solution R is better than that of the current optimal solution, then updating the current optimal solution with the candidate approximate optimal solution R.

[0028] As a preferred embodiment of the image enhancement method for surface defects of silicon steel plates based on the improved slime mold algorithm described in this invention, the size N of the slime mold population is 30, the maximum number of iterations tmax is 300, and the search dimension D of the variable to be optimized is 2.

[0029] As a preferred embodiment of the silicon steel plate surface defect image enhancement method based on the improved slime mold algorithm described in this invention, in step S7, the preset termination condition includes at least one of the following: reaching the maximum number of iterations, the change in the optimal fitness after a consecutive preset number of iterations being less than a preset threshold, or the current optimal fitness reaching a preset target value.

[0030] An image enhancement system for surface defects in silicon steel plates based on an improved slime mold algorithm includes:

[0031] The image acquisition module is used to acquire low-light images to be enhanced in the scenario of detecting defects on the surface of silicon steel plates and perform normalization processing; the transformation model construction module is used to establish an incomplete Beta function grayscale transformation model and determine the shape parameters to be optimized.

[0032] The fitness evaluation module is used to calculate individual fitness based on information entropy, edge content, logarithm of edge strength, and average gradient.

[0033] The population initialization module is used to generate chaotic sequences and complete the initialization of slime mold populations using the Singer chaotic mapping strategy.

[0034] An iterative optimization module is used to introduce a hyperbolic oscillation factor to adjust the search mode and generate candidate solutions through a quadratic interpolation strategy, iteratively searching for the optimal parameter combination. The iterative optimization module includes a global search unit, a local development unit, and a candidate solution competition unit. The global search unit is used to expand the search range of the slime mold population in the early stage of iteration. The local development unit is used to enhance the local search capability near the optimal solution in the later stage of iteration. The candidate solution competition unit is used to compare the fitness of the candidate approximate optimal solution generated by the quadratic interpolation strategy with the current optimal solution.

[0035] The image enhancement module is used to substitute the optimal parameter combination into the incomplete Beta function, perform grayscale transformation on the low-light image to be enhanced, and output the enhanced image.

[0036] Compared with the prior art, the beneficial effects of the present invention are: the present invention can enhance the recognizability of defect images of silicon steel plates; can highlight the edges and textures of defect areas such as fine cracks, scratches, patches, and pitting; can reduce the influence of manual experience in parameter selection on the enhancement results; and can provide more stable image input for subsequent defect segmentation, recognition, and quality judgment. Attached Figure Description

[0037] To more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0038] Figure 1 This is a flowchart of the image enhancement method for surface defects of silicon steel plates based on the improved slime mold algorithm of the present invention;

[0039] Figure 2 This is a flowchart of the parameter optimization process for the improved slime mold algorithm in the image enhancement method for surface defects of silicon steel plates based on the improved slime mold algorithm of the present invention.

[0040] Figure 3 This is a system diagram of the silicon steel plate surface defect image enhancement system based on the improved slime mold algorithm of the present invention;

[0041] Figure 4 This is a flowchart illustrating the engineering application of the present invention for detecting surface defects in silicon steel plates.

[0042] Figure 5 This is a schematic diagram comparing the low-light images of silicon steel plate surface defects before and after enhancement according to the present invention. The left column shows the low-light defect images before enhancement, and the right column shows the enhancement results obtained by using the method of the present invention under the same input conditions. Detailed Implementation

[0043] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0044] Secondly, the present invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of the present invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not according to the usual scale. Furthermore, the schematic diagrams are merely examples and should not limit the scope of protection of the present invention. In addition, actual fabrication should include three-dimensional spatial dimensions of length, width, and depth.

[0045] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0046] Please see Figure 1 A method for enhancing surface defect images of silicon steel plates based on an improved slime mold algorithm includes the following steps:

[0047] S1: Acquire the low-light image to be enhanced under the scenario of detecting surface defects in silicon steel plate, and normalize the low-light image to be enhanced to obtain a normalized image.

[0048] The surface defect detection scenarios for silicon steel plates include at least one of rolling production, offline sampling inspection, or laboratory testing; the low-light images to be enhanced include at least one defect type image of fine cracks, inclusions, patches, pitting, iron oxide scale rolling, or scratches.

[0049] The specific method of normalization processing is as follows: when the low-light image to be enhanced is a grayscale image, the grayscale pixel values ​​are normalized; when the low-light image to be enhanced is a color image, the color image is converted to the HSV color space, the luminance V channel is extracted and normalized, and then converted back to the RGB color space after enhancement. This processing method separates luminance enhancement from color information, reducing the interference of color drift on the judgment of defect texture.

[0050] S2: Establish an incomplete Beta function grayscale transformation model, and use the shape parameters in the incomplete Beta function grayscale transformation model as variables to be optimized;

[0051] When establishing the grayscale transformation model, the normalized pixel value u is input into an incomplete Beta function, and the mapping curve is controlled by shape parameters a and b. The incomplete Beta function can be expressed as follows: ,in, u is the normalized original pixel value, F(u) is the transformed pixel value, a and b are the shape parameters, and the search range of a and b can be set to (0,10) to cover different enhancement requirements such as dark area stretching, bright area compression and intermediate grayscale adjustment.

[0052] S3: Construct a fitness function for evaluating the quality of the enhanced image, the fitness function including information entropy, edge content, logarithm of edge intensity, and average gradient;

[0053] The fitness function is expressed as: Where H is the information entropy of the enhanced image, EC is the edge content of the enhanced image, LSI is the logarithmic sum of the edge intensities of the enhanced image, and MG is the average gradient of the enhanced image. It is a positive control coefficient; this function allows parameter search to simultaneously consider image sharpness and defect outline visibility.

[0054] The edge content EC is determined based on the ratio of the number of edge pixels to the total number of pixels in the image; the edge intensity logarithm and LSI are determined based on the horizontal and vertical gradients obtained from Sobel edge detection; and the average gradient MG is determined based on the grayscale difference between adjacent pixels in the horizontal and vertical directions.

[0055] S4: Initialize the slime mold population using the Singer chaotic mapping strategy to obtain a preset number of initial individuals, each of which corresponds to a set of variables to be optimized;

[0056] Singer's chaotic mapping strategy is represented as follows: ,in, ∈(0.9,1.08), and map the chaotic sequence generated by the Singer chaotic mapping strategy to the search boundary of the variable to be optimized, so as to determine the position of each initial individual in the slime mold population.

[0057] During the initialization phase, a Singer chaotic mapping is used to generate candidate sequences, which are then mapped to the parameter boundaries of a and b. This results in an initial population that does not rely on simple random sampling, reducing the probability of initial solutions concentrating in local intervals.

[0058] S5: Iteratively update the slime mold population based on the slime mold algorithm (see [link]). Figure 2 (and introduces a hyperbolic oscillation factor during the iterative update process to dynamically adjust the global search capability and local exploitation capability of the slime mold population);

[0059] The hyperbolic oscillation factor is a piecewise function: in the first half of the iteration, the search weights are made to change nonlinearly based on the hyperbolic tangent function to expand the exploration range of the parameter space; in the second half of the iteration, local oscillation perturbations are formed based on the sine and cosine functions to enhance the development capability near the optimal solution and reduce the probability of getting trapped in local optima.

[0060] During the iterative optimization phase, the enhanced image and fitness value corresponding to each candidate parameter combination are calculated, and the current optimal parameter combination is used as the reference position for subsequent updates. During the iteration process, a hyperbolic oscillation factor is introduced to adjust the shrinkage and perturbation intensity in the slime mold position update in stages.

[0061] In the early stage of iteration, the hyperbolic oscillation factor keeps the candidate solution with a large search span in order to discover the enhancement effect corresponding to different gray-scale transformation curves; in the later stage of iteration, the oscillation term causes the candidate solution to be perturbed in a small range around the better parameter region in order to improve the local parameter correction capability.

[0062] S6: In the process of updating the current optimal solution, a quadratic interpolation strategy is introduced to generate candidate approximate optimal solutions, and the candidate approximate optimal solutions compete with the current optimal solution for selection;

[0063] Select two distinct individual positions X and Y from the slime mold population, as well as the current global best position Z. Construct a quadratic interpolation model based on the fitness values ​​of X, Y, and Z to obtain a candidate approximate optimal solution R. If the fitness of the candidate approximate optimal solution R is better than that of the current optimal solution, then update the current optimal solution with the candidate approximate optimal solution R.

[0064] When further searching the currently better region, two candidate positions X and Y, as well as the current global best position Z, are selected from the population. A quadratic interpolation candidate solution R is constructed based on the fitness of these three positions. If the enhanced image corresponding to R has a higher fitness, then R replaces the current best position; otherwise, the original best position remains unchanged.

[0065] S7: When the preset termination condition is met, output the optimal parameter combination of the incomplete Beta function grayscale transformation model;

[0066] The preset termination conditions include at least one of the following: reaching the maximum number of iterations, the change in the optimal fitness after a preset number of consecutive iterations being less than a preset threshold, or the current optimal fitness reaching a preset target value.

[0067] The iteration stops and the optimal parameter combination is output when the maximum number of iterations is reached, the change in optimal fitness after multiple consecutive iterations is less than a threshold, or the optimal fitness reaches the target value. This parameter combination is then substituted into the incomplete Beta function to perform a grayscale transformation on the image to be enhanced, resulting in the enhanced image of the silicon steel plate surface defects.

[0068] S8: Perform grayscale transformation on the low-light image to be enhanced according to the optimal parameter combination to obtain the enhanced silicon steel plate surface defect image.

[0069] In one specific embodiment, the slime mold population size N is set to 30, the maximum number of iterations tmax is set to 300, and the dimension D of the variable to be optimized is set to 2. This setting is suitable for image enhancement tasks that optimize two shape parameters of an incomplete Beta function.

[0070] Please see Figure 4 In engineering applications, enhanced images can be used in multi-threshold segmentation, edge extraction, depth recognition models, or defect classification modules. For defects such as fine cracks and scratches, enhanced images help to increase the grayscale distance between slender edges and the substrate; for defects such as patches, pitting, and rolled iron oxide scale, enhanced images help to present the texture changes in the defect area; and for defects such as inclusions, enhanced images help to strengthen local contours and suppress background interference.

[0071] The present invention can adaptively determine enhancement parameters based on the grayscale distribution, edge response and local gradient of the input image, thus making it easier to adapt to different defect types, different lighting conditions and different board surface texture states.

[0072] To visually illustrate the enhancement effect of the method of this invention on surface defect images of low-light silicon steel plates, typical defect samples such as fine cracks, inclusions, pitting, and scratches were selected for processing. The processing results are as follows: Figure 5 As shown.

[0073] Depend on Figure 5 As can be seen, after processing with the method of the present invention, the gray level, edge contour and local texture of the defect area on the surface of silicon steel plate are clearer, and the defect information that is not easy to identify under low illumination conditions is highlighted, which is beneficial to subsequent defect segmentation, identification and quality judgment processing.

[0074] Please see Figure 3 The present invention also provides an image enhancement system for surface defects of silicon steel plates based on an improved slime mold algorithm:

[0075] It includes: an image acquisition module, used to acquire low-light images to be enhanced in the scenario of detecting defects on the surface of silicon steel plates and perform normalization processing; and a transformation model construction module, used to establish an incomplete Beta function grayscale transformation model and determine the shape parameters to be optimized.

[0076] The fitness evaluation module is used to calculate individual fitness based on information entropy, edge content, logarithm of edge strength, and average gradient.

[0077] The population initialization module is used to generate chaotic sequences and complete the initialization of slime mold populations using the Singer chaotic mapping strategy.

[0078] An iterative optimization module is used to introduce a hyperbolic oscillation factor to adjust the search mode and generate candidate solutions through a quadratic interpolation strategy, iteratively searching for the optimal parameter combination. The iterative optimization module includes a global search unit, a local development unit, and a candidate solution competition unit. The global search unit is used to expand the search range of the slime mold population in the early stage of iteration. The local development unit is used to enhance the local search capability near the optimal solution in the later stage of iteration. The candidate solution competition unit is used to compare the fitness of the candidate approximate optimal solution generated by the quadratic interpolation strategy with the current optimal solution.

[0079] The image enhancement module is used to substitute the optimal parameter combination into the incomplete Beta function, perform grayscale transformation on the low-light image to be enhanced, and output the enhanced image.

[0080] In another embodiment, the present invention also provides an electronic device comprising an image acquisition module, a processor, and a memory. The image acquisition module may be an industrial camera, a CMOS image sensor, or other machine vision acquisition unit; the processor is used to execute the image enhancement method; and the memory is used to store computer programs, parameter configurations, and detected images.

[0081] The electronic device can be deployed on a silicon steel plate surface quality inspection platform. After receiving a low-light defect image, the processor calls the program in the memory to complete image normalization, parameter optimization, grayscale transformation, and enhanced image output. The enhanced image can be further transmitted to the defect segmentation module, defect recognition module, or quality judgment module.

[0082] Although the present invention has been described above with reference to embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of the invention. In particular, as long as there is no structural conflict, the features in the disclosed embodiments can be combined with each other in any manner. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, the present invention is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm, characterized in that, Includes the following steps: S1: Acquire the low-light image to be enhanced under the scenario of detecting surface defects in silicon steel plate, and normalize the low-light image to be enhanced to obtain a normalized image. S2: Establish an incomplete Beta function grayscale transformation model, and use the shape parameters in the incomplete Beta function grayscale transformation model as variables to be optimized; S3: Construct a fitness function for evaluating the quality of the enhanced image, the fitness function including information entropy, edge content, logarithm of edge intensity, and average gradient; S4: Initialize the slime mold population using the Singer chaotic mapping strategy to obtain a preset number of initial individuals, each of which corresponds to a set of variables to be optimized; S5: The slime mold population is iteratively updated based on the slime mold algorithm, and a hyperbolic oscillation factor is introduced during the iterative update process to dynamically adjust the global search capability and local exploitation capability of the slime mold population. S6: In the process of updating the current optimal solution, a quadratic interpolation strategy is introduced to generate candidate approximate optimal solutions, and the candidate approximate optimal solutions compete with the current optimal solution for selection; S7: When the preset termination condition is met, output the optimal parameter combination of the incomplete Beta function grayscale transformation model; S8: Perform grayscale transformation on the low-light image to be enhanced according to the optimal parameter combination to obtain the enhanced silicon steel plate surface defect image.

2. The method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm according to claim 1, characterized in that, In S1, the surface defect detection scenario of silicon steel plate includes at least one of rolling production, offline sampling inspection or laboratory testing; the low-light image to be enhanced includes at least one defect type image of fine cracks, inclusions, patches, pitting, iron oxide scale rolling or scratches. The specific method of normalization processing is as follows: when the low-light image to be enhanced is a grayscale image, the grayscale pixel values ​​are normalized; when the low-light image to be enhanced is a color image, the color image is converted to the HSV color space, the luminance V channel is extracted and the luminance V channel is normalized, and then converted back to the RGB color space after enhancement.

3. The method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm according to claim 1, characterized in that, In S2, the incomplete Beta function grayscale transformation model is expressed as follows: ,in, u is the normalized original pixel value, F(u) is the transformed pixel value, a and b are the shape parameters, and the values ​​of a and b are both in the range of (0,10).

4. The method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm according to claim 1, characterized in that, In S3, the fitness function is expressed as: Where H is the information entropy of the enhanced image, EC is the edge content of the enhanced image, LSI is the logarithmic sum of the edge intensities of the enhanced image, and MG is the average gradient of the enhanced image. It is a positive control coefficient; The edge content EC is determined based on the ratio of the number of edge pixels to the total number of pixels in the image; the edge intensity logarithm and LSI are determined based on the horizontal and vertical gradients obtained from Sobel edge detection; and the average gradient MG is determined based on the grayscale difference between adjacent pixels in the horizontal and vertical directions.

5. The method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm according to claim 1, characterized in that, In S4, the Singer chaotic mapping strategy is expressed as follows: ,in, ∈(0.9,1.08), and map the chaotic sequence generated by the Singer chaotic mapping strategy to the search boundary of the variable to be optimized, so as to determine the position of each initial individual in the slime mold population.

6. The method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm according to claim 1, characterized in that, In S5, the hyperbolic oscillation factor is a piecewise function: in the first half of the iteration, the search weight is changed nonlinearly based on the hyperbolic tangent function to expand the exploration range of the parameter space; in the second half of the iteration, local oscillation disturbances are formed based on the sine and cosine functions to enhance the development capability near the optimal solution and reduce the probability of getting trapped in local optima.

7. The method for enhancing surface defect images of silicon steel plates based on the improved slime mold algorithm according to claim 1, characterized in that, S6 includes: selecting two distinct individual positions X and Y from the slime mold population and the current global optimal position Z; constructing a quadratic interpolation model based on the fitness values ​​of X, Y, and Z to obtain a candidate approximate optimal solution R; if the fitness of the candidate approximate optimal solution R is better than that of the current optimal solution, then updating the current optimal solution with the candidate approximate optimal solution R.

8. The method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm according to claim 1, characterized in that, The size N of the slime mold population is 30, the maximum number of iterations tmax is 300, and the search dimension D of the variable to be optimized is 2.

9. The method for enhancing surface defect images of silicon steel plates based on the improved slime mold algorithm according to claim 1, characterized in that, In S7, the preset termination conditions include at least one of the following: reaching the maximum number of iterations, the change in the optimal fitness after a preset number of consecutive iterations being less than a preset threshold, or the current optimal fitness reaching a preset target value.

10. A system for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm, used to implement the method for enhancing images of surface defects in silicon steel plates based on an improved slime mold algorithm as described in any one of claims 1-9, characterized in that, include: The image acquisition module is used to acquire low-light images to be enhanced in the scenario of detecting defects on the surface of silicon steel plates and to perform normalization processing. The transformation model construction module is used to establish an incomplete Beta function grayscale transformation model and determine the shape parameters to be optimized. The fitness evaluation module is used to calculate individual fitness based on information entropy, edge content, logarithm of edge strength, and average gradient. The population initialization module is used to generate chaotic sequences and complete the initialization of slime mold populations using the Singer chaotic mapping strategy. An iterative optimization module is used to introduce a hyperbolic oscillation factor to adjust the search mode and generate candidate solutions through a quadratic interpolation strategy, iteratively searching for the optimal parameter combination. The iterative optimization module includes a global search unit, a local development unit, and a candidate solution competition unit. The global search unit is used to expand the search range of the slime mold population in the early stage of iteration. The local development unit is used to enhance the local search capability near the optimal solution in the later stage of iteration. The candidate solution competition unit is used to compare the fitness of the candidate approximate optimal solution generated by the quadratic interpolation strategy with the current optimal solution. The image enhancement module is used to substitute the optimal parameter combination into the incomplete Beta function, perform grayscale transformation on the low-light image to be enhanced, and output the enhanced image.

Images