An adaptive arc area repair and enhancement method

Abstract

The present invention discloses an adaptive arc light area repair and enhancement method, including step S1: image segmentation; step S2: arc light interference model calculation; step S3: image classification; step S4: image repair and enhancement; step S5: image restoration. The present invention proposes an adaptive solution to the interference caused by arc light on the image, such as reduced contrast, color deviation, instantaneous overexposure, low illumination, detail loss, etc. Through the above steps, the arc light area can be effectively repaired, the image details can be enhanced, and the image quality can be improved. The present invention can be applied to the repair and enhancement of images in complex industrial scenes, providing help for the subsequent development of industrial intelligence, and has broad application prospects.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence and image processing, and in particular to an adaptive arc area repair and enhancement method. Background Art

[0002] In complex industrial scenes, there are many common interferences: low illumination, dim light, dust pollution, uneven illumination, etc. Especially in the industrial field, electric arc is often used as a process means for welding, smelting, etc. Because of its high brightness and fast change characteristics, it is easy to cause the following to the image: ① Instantaneous overexposure, light saturation, arc area cover loss interference; ② Instantaneous light intensity reduction, low illumination, detail loss interference; ③ Brightness changes drastically, light and dark area contrast interference; ④ Color shift, distortion interference; ⑤ Contrast reduction, image quality reduction, etc. This not only affects the quality of the collected images, but also seriously affects the subsequent intelligent function development that relies on images, such as safety monitoring, process flow control, process object feature recognition, detection, positioning, etc., limiting the application of computer vision technology in industrial scenes.

[0003] At present, although certain research progress and results have been achieved in the field of image enhancement and restoration, there is no targeted research on the field of arc interference restoration. Existing related research mainly addresses the interference caused by environmental factors such as fog and sandstorms on images, assuming that the ambient light is globally constant. However, arcs have the characteristics of fast changes and strong brightness, which leads to images with drastic light changes, color deviation, and many highlight areas, which are different from the above research assumptions. Therefore, the above research cannot be applied to the research field of this article. Summary of the invention

[0004] In order to overcome the deficiencies in the prior art, the present invention provides an adaptive arc area repair and enhancement method to solve the defects of the existing image repair and enhancement methods.

[0005] In order to achieve the above-mentioned invention object, the technical solution adopted to solve the technical problem is as follows:

[0006] An adaptive arc area repair and enhancement method comprises the following steps:

[0007] Step S1: Image segmentation: The degree of arc interference in different areas of the image is different, and it is impossible to achieve repair and enhancement in a global way. Therefore, the image to be processed needs to be segmented into several split images of the same size;

[0008] Step S2: Calculation of arc light interference model: By analyzing the characteristics of arc light and its interference mechanism on the image, an arc light interference model consisting of distance factor, light intensity factor and color factor is proposed to measure the degree of arc light interference on each split image. The model value is used as the weight for subsequent image restoration and enhancement processing;

[0009] Step S3: Image classification: according to the arc interference model value calculated in step S2, the split image is divided into class I, class II, and class III images;

[0010] Step S4: Image restoration and enhancement: Based on the image classification in step S3, each split image is adaptively processed by unified light intensity, multi-frame replacement, color restoration, and contrast enhancement modules;

[0011] Step S5: Image restoration: restore each split image to its original position, and finally achieve adaptive arc area repair and enhancement effects in complex industrial scenes.

[0012] Furthermore, in step S2, the numerical calculation step of the arc interference model includes:

[0013] Step S21: distance factor calculation: This factor measures the degree to which each split image is disturbed by the arc light by calculating the distance from each split image to the arc light center area;

[0014] Step S22: Calculation of light intensity factor: This factor measures the degree of arc light interference on each split image by calculating the light intensity of each split image;

[0015] Step S23: Color factor calculation: This factor measures the degree of arc interference of each split image by calculating the closeness of the RGB three-channel colors of each split image. The calculation method is shown in formula (1):

[0016]

[0017] Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; C(i,j) is the color factor of the split image (i,j), is the standard deviation of the split image (i, j); The calculation method is shown in formula (2):

[0018]

[0019] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N R (x,y),N G (x,y),N B (x,y),N grey (x, y) are the pixel values ​​of the R channel, G channel, B channel, and gray image in the split image (i, j);

[0020] According to the U(i,j) value, a decimal in the interval [0,1] is assigned to each split image as the color factor value. The greater the color difference, the closer the value is to 1, and the smaller the color difference, the closer the value is to 0.

[0021] Step S24: Arc interference model, defined as shown in formula (3):

[0022]

[0023] Wherein, i,j are the center coordinates of the split image; (i,j) is the split image with the center coordinates i,j; DLC(i,j) is the arc interference model value of the split image (i,j); D(i,j) is the distance factor of the split image (i,j); L(i,j) is the light intensity factor of the split image (i,j); C(i,j) is the color factor of the split image (i,j); a, b, c are the proportional coefficients.

[0024] Furthermore, the calculation steps of step S21 include:

[0025] Step S211: using image preprocessing techniques: pixel point traversal, grayscale image pixel threshold setting, and morphological operations, identifying and representing the arc center area with a binary image;

[0026] Step S212: Calculate the distance between the center point of each split image and the arc center area;

[0027] Step S213: According to the value calculated in step S212, a decimal in the interval [0,1] is assigned to each split image as the distance factor D(i,j) value. The closer the distance, the closer the value is to 1, and the farther the distance, the closer the value is to 0.

[0028] Furthermore, the calculation steps of step S22 include:

[0029] Step S221: image format conversion, converting the RGB image into a grayscale image to reduce the amount of calculation;

[0030] Step S222: Calculate the average illumination intensity of each split image, as shown in formula (4):

[0031]

[0032] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; L(i, j) is the light intensity factor of the split image (i, j); N sum (i, j) is the sum of the light intensity of each pixel in the split image (i, j); u and v are the length and width of the split image;

[0033] Step S223: The numerical value calculated in step S222 is in the range of [0,1]. The stronger the light intensity, the closer the value is to 1, and the weaker the light intensity, the closer the value is to 0.

[0034] Further, in step S3, according to the arc interference model value calculated in step S2, the split image is divided into class I, class II, and class III images, including:

[0035] Step S31: Category I: split images without arc light or with little arc light influence, in which case 0≤DLC(i,j)<m, need to be processed by uniform light intensity and contrast enhancement adaptive module;

[0036] Step S32: Category II: The arc light causes a certain impact and the arc light area needs to be repaired. At this time, m≤DLC(i,j)<n, and it needs to be processed by the uniform light intensity, color restoration, and contrast enhancement adaptive modules;

[0037] Step S33: Category III: The arc light is serious, causing irreparable damage to the picture and details. At this time, n≤DLC(i,j)≤1, and it needs to be processed by unified light intensity, multi-frame replacement, color restoration, and contrast enhancement adaptive modules.

[0038] Further, in step S4, according to the split image classification performed in step S3, each split image is adaptively selected to be processed by: uniform light intensity, multi-frame replacement, color restoration, and contrast enhancement modules, including:

[0039] Step S41: unify the light intensity module, and its implementation method is shown in formula (5):

[0040] L Gamma (i,j)=[L(i,j)] γ *255 (5)

[0041] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; L Gamma (i, j) is the intensity factor of the split image (i, j) after unification; L(i, j) is the original intensity factor of the split image (i, j); γ is the unified intensity correction index, as shown in formula (6):

[0042] γ=log2[L(i,j) aim -L(i,j) median ] (6)

[0043] Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L(i,j) aim is the target brightness mean of the split image (i, j); L(i, j) median is the median brightness of the split image (i, j);

[0044]

[0045] Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L(i,j) mean is the mean brightness of the split image (i, j);

[0046] Step S42: multi-frame replacement module, the Class III split image is seriously disturbed by arc light, causing irreparable damage to the picture and details, and the arc light area also has the phenomenon of covering loss. According to the arc light characteristics and the principle of inter-frame similarity, for the Class III image, the non-Class III split images of the four frames before and after it are subjected to multi-channel weighted average fusion processing to replace the original split image;

[0047] Step S43: Color restoration module, which is implemented as shown in equations (8, 9, 10):

[0048] The R channel transformation rule is shown in formula (8):

[0049]

[0050] The G channel transformation rule is shown in formula (9):

[0051]

[0052] The B channel transformation rule is shown in formula (10):

[0053]

[0054] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N R (x, y) is the pixel value of the R channel of the split image (i, j); u is the reference parameter; are the maximum and minimum pixel values ​​of the G channel of the split image (i, j); U(i, j) is the variance of the split image (i, j); U R (i, j) is the variance of the R channel of the split image (i, j); μ R (i, j) is the mean of the R channel of the split image (i, j);

[0055] Step S44: Contrast enhancement module, which needs to convert the split image format from RGB to HIS space to reduce the amount of calculation. Contrast enhancement only needs to operate on the I channel. Its implementation method is shown in formula (11):

[0056]

[0057] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N I (x, y) is the pixel value of each pixel in the I channel of the split image (i, j); μ I (i,j) is the pixel mean of channel I of the split image (i,j); is the pixel standard deviation of channel I of the split image (i, j).

[0058] Due to the adoption of the above technical solution, the present invention has the following advantages and positive effects compared with the prior art:

[0059] The present invention proposes an adaptive solution to the interference of arc light on images, such as reduced contrast, color deviation, instantaneous overexposure, low illumination, and detail loss. By constructing an arc light interference model and combining the newly proposed unified light intensity, multi-frame replacement, color restoration, and contrast enhancement modules, the method of the present invention can effectively filter out arc light interference, repair arc light areas, enhance image details, and improve image quality. It can be applied to image restoration and enhancement in complex industrial scenes, provide assistance for the subsequent development of industrial intelligence, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the 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 work. In the drawings:

[0061] Figure 1 It is a flow chart of an adaptive arc area repair and enhancement method according to a specific embodiment of the present invention;

[0062] Figure 2 This is an arc center area recognition effect diagram of a specific embodiment of the present invention;

[0063] Figure 3 It is a before-and-after comparison diagram of the original image without arc light and the processed image of a specific embodiment of the present invention. DETAILED DESCRIPTION

[0064] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] like Figure 1 This embodiment discloses an adaptive arc area repair and enhancement method, comprising the following steps:

[0066] Step S1: Image segmentation: The degree of arc interference in different areas of the image is different, and it is impossible to achieve repair and enhancement in a global way. Therefore, the image to be processed needs to be segmented into several split images of the same size;

[0067] Step S2: Calculation of arc light interference model: By analyzing the characteristics of arc light and its interference mechanism on the image, an arc light interference model consisting of distance factor, light intensity factor and color factor is proposed to measure the degree of arc light interference on each split image. The model value is used as the weight for subsequent image restoration and enhancement processing;

[0068] Step S3: Image classification: according to the arc interference model value calculated in step S2, the split image is divided into class I, class II, and class III images;

[0069] Step S4: Image restoration and enhancement: Based on the image classification in step S3, each split image is adaptively processed by unified light intensity, multi-frame replacement, color restoration, and contrast enhancement modules;

[0070] Step S5: Image restoration: restore each split image to its original position, and finally achieve adaptive arc area repair and enhancement effects in complex industrial scenes.

[0071] Furthermore, in step S2, the numerical calculation step of the arc interference model includes:

[0072] Step S21: distance factor calculation: This factor measures the degree to which each split image is disturbed by the arc light by calculating the distance from each split image to the arc light center area;

[0073] Specifically, the calculation steps of step S21 include:

[0074] Step S211: Through image preprocessing techniques: pixel point traversal, grayscale image pixel threshold setting, morphology and other operations, the arc center area is identified and represented by a binary image. The effect is as follows: Figure 2 As shown;

[0075] Step S212: Calculate the distance between the center point of each split image and the arc center area;

[0076] Step S213: According to the value calculated in step S212, a decimal in the interval [0,1] is assigned to each split image as the distance factor D(i,j) value. The closer the distance, the closer the value is to 1, and the farther the distance, the closer the value is to 0.

[0077] Step S22: Calculation of light intensity factor: This factor measures the degree of arc light interference on each split image by calculating the light intensity of each split image;

[0078] Specifically, the calculation steps of step S22 include:

[0079] Step S221: image format conversion, converting the RGB image into a grayscale image to reduce the amount of calculation;

[0080] Step S222: Calculate the average illumination intensity of each split image, as shown in formula (4):

[0081]

[0082] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; L(i, j) is the light intensity factor of the split image (i, j); N sum (i, j) is the sum of the light intensity of each pixel in the split image (i, j); u and v are the length and width of the split image;

[0083] Step S223: The numerical value calculated in step S222 is in the range of [0,1]. The stronger the light intensity, the closer the value is to 1, and the weaker the light intensity, the closer the value is to 0.

[0084] Step S23: Color factor calculation: This factor measures the degree of arc interference of each split image by calculating the closeness of the RGB three-channel colors of each split image. The calculation method is shown in formula (1):

[0085]

[0086] Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; C(i,j) is the color factor of the split image (i,j), is the standard deviation of the split image (i, j); The calculation method is shown in formula (2):

[0087]

[0088] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N R (x,y),N G (x,y),N B (x,y),N grey (x, y) are the pixel values ​​of the R channel, G channel, B channel, and gray image in the split image (i, j);

[0089] According to the U(i,j) value, a decimal in the interval [0,1] is assigned to each split image as the color factor value. The greater the color difference, the closer the value is to 1, and the smaller the color difference, the closer the value is to 0.

[0090] Step S24: Arc interference model, defined as shown in formula (3):

[0091]

[0092] Wherein, i,j are the center coordinates of the split image; (i,j) is the split image with the center coordinates i,j; DLC(i,j) is the arc interference model value of the split image (i,j); D(i,j) is the distance factor of the split image (i,j); L(i,j) is the light intensity factor of the split image (i,j); C(i,j) is the color factor of the split image (i,j); a, b, c are the proportional coefficients.

[0093] Further, in step S3, according to the arc interference model value calculated in step S2, the split image is divided into class I, class II, and class III images, including:

[0094] Step S31: Category I: split images without arc light or with little arc light influence, in which case 0≤DLC(i,j)<m, need to be processed by uniform light intensity and contrast enhancement adaptive module;

[0095] Step S32: Category II: The arc light causes a certain impact and the arc light area needs to be repaired. At this time, m≤DLC(i,j)<n, and it needs to be processed by the uniform light intensity, color restoration, and contrast enhancement adaptive modules;

[0096] Step S33: Category III: The arc light is serious, causing irreparable damage to the picture and details. At this time, n≤DLC(i,j)≤1, and it needs to be processed by unified light intensity, multi-frame replacement, color restoration, and contrast enhancement adaptive modules.

[0097] Further, in step S4, according to the split image classification performed in step S3, each split image is adaptively selected to be processed by: uniform light intensity, multi-frame replacement, color restoration, and contrast enhancement modules, including:

[0098] Step S41: unify the light intensity module, and its implementation method is shown in formula (5):

[0099] L Gamma (i,j)=[L(i,j)] γ *255 (5)

[0100] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; L Gamma (i, j) is the intensity factor of the split image (i, j) after unification; L(i, j) is the original intensity factor of the split image (i, j); γ is the unified intensity correction index, as shown in formula (6):

[0101] γ=log2[L(i,j)aim -L(i,j) median ] (6)

[0102] Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L(i,j) aim is the target brightness mean of the split image (i, j); L(i, j) median is the median brightness of the split image (i, j);

[0103]

[0104] Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L(i,j) mean is the mean brightness of the split image (i, j);

[0105] Step S42: multi-frame replacement module: the Class III split image is seriously disturbed by arc light, causing irreparable damage to the picture and details, and the arc light area also has the phenomenon of cover loss. According to the arc light characteristics and the principle of inter-frame similarity, for the Class III image, the non-Class III split images of the four frames before and after it are subjected to multi-channel weighted average fusion processing to replace the original split image;

[0106] Step S43: Color restoration module, which is implemented as shown in equations (8, 9, 10):

[0107] The R channel transformation rule is shown in formula (8):

[0108]

[0109] The G channel transformation rule is shown in formula (9):

[0110]

[0111] The B channel transformation rule is shown in formula (10):

[0112]

[0113] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N R (x, y) is the pixel value of the R channel of the split image (i, j); u is the reference parameter; are the maximum and minimum pixel values ​​of the G channel of the split image (i, j); U(i, j) is the variance of the split image (i, j); U R (i, j) is the variance of the R channel of the split image (i, j); μ R (i, j) is the mean of the R channel of the split image (i, j);

[0114] Step S44: Contrast enhancement module, which needs to convert the split image format from RGB to HIS space to reduce the amount of calculation. Contrast enhancement only needs to operate on the I channel. Its implementation method is shown in formula (11):

[0115]

[0116] Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N I (x, y) is the pixel value of each pixel in the I channel of the split image (i, j); μ I (i,j) is the pixel mean of channel I of the split image (i,j); is the pixel standard deviation of channel I of the split image (i, j).

[0117] The processed image is outputted in step S5. Finally, the effect comparison diagram of the method of the present invention is shown as follows: Figure 3 shown.

[0118] The present invention is aimed at studying the field of image restoration and enhancement in complex industrial scenes. Due to the appearance of arc light in industrial scenes, the image is disturbed by reduced contrast, color deviation, instantaneous overexposure, low illumination, detail loss, etc. Various image restoration and enhancement methods have been studied and it is found that due to the special properties of arc light, existing research cannot be effectively applied to the restoration and enhancement of arc light areas in complex industries. Therefore, the present invention proposes an adaptive arc light area restoration and enhancement method. The algorithm is verified through real industrial scenes, and the method has good performance, which proves that the invention can effectively filter out arc light interference, enhance image details, and improve image quality. It can be applied to image restoration and enhancement in complex industrial scenes, and provide help for the subsequent development of industrial intelligence.

[0119] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An adaptive arc area repair and enhancement method, characterized in that: The following steps are involved: Step S1: Image segmentation: The degree of arc interference in different areas of the image is different, and it is impossible to achieve repair and enhancement in a global way. Therefore, the image to be processed needs to be segmented into several split images of the same size; Step S2: Calculation of arc light interference model: By analyzing the characteristics of arc light and its interference mechanism on the image, an arc light interference model consisting of distance factor, light intensity factor and color factor is proposed to measure the degree of arc light interference on each split image. The model value is used as the weight for subsequent image restoration and enhancement processing; In step S2, the numerical calculation steps of the arc interference model include: Step S21: distance factor calculation: This factor measures the degree to which each split image is disturbed by the arc light by calculating the distance from each split image to the arc light center area; Step S22: Calculation of light intensity factor: This factor measures the degree of arc light interference on each split image by calculating the light intensity of each split image; Step S23: Color factor calculation: This factor measures the degree of arc interference of each split image by calculating the closeness of the RGB three-channel colors of each split image. The calculation method is shown in formula (1): Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; C(i,j) is the color factor of the split image (i,j), is the standard deviation of the split image (i, j); The calculation method is shown in formula (2): Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; N R (x,y),N G (x,y),N B (x,y),N grey (x, y) are the pixel values ​​of the R channel, G channel, B channel, and gray image in the split image (i, j); According to the U(i,j) value, a decimal in the interval [0,1] is assigned to each split image as the color factor value. The greater the color difference, the closer the value is to 1, and the smaller the color difference, the closer the value is to 0. Step S24: Arc interference model, defined as shown in formula (3): Wherein, i,j are the center coordinates of the split image; (i,j) is the split image with the center coordinates i,j; DLC(i,j) is the arc interference model value of the split image (i,j); D(i,j) is the distance factor of the split image (i,j); L(i,j) is the light intensity factor of the split image (i,j); C(i,j) is the color factor of the split image (i,j); a, b, c are the proportional coefficients; Step S3: Image classification: according to the arc interference model value calculated in step S2, the split image is divided into class I, class II, and class III images; Step S4: Image restoration and enhancement: Based on the image classification in step S3, each split image is adaptively processed by unified light intensity, multi-frame replacement, color restoration, and contrast enhancement modules; Step S5: Image restoration: restore each split image to its original position, and finally achieve adaptive arc area repair and enhancement effects in complex industrial scenes.

2. The method for adaptive arc area repair and enhancement according to claim 1, characterized in that: The calculation steps of step S21 include: Step S211: using image preprocessing techniques: pixel point traversal, grayscale image pixel threshold setting, and morphological operations, identifying and representing the arc center area with a binary image; Step S212: Calculate the distance between the center point of each split image and the arc center area; Step S213: According to the value calculated in step S212, a decimal in the interval [0,1] is assigned to each split image as the distance factor D(i,j) value. The closer the distance, the closer the value is to 1, and the farther the distance, the closer the value is to 0.

3. The method for adaptive arc area repair and enhancement according to claim 1, characterized in that: The calculation steps of step S22 include: Step S221: image format conversion, converting the RGB image into a grayscale image to reduce the amount of calculation; Step S222: Calculate the average illumination intensity of each split image, as shown in formula (4): Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; L(i, j) is the light intensity factor of the split image (i, j); N sum (i, j) is the sum of the light intensity of each pixel in the split image (i, j); u and v are the length and width of the split image; Step S223: The numerical value calculated in step S222 is in the interval [0,1], where the stronger the light intensity, the closer the value is to 1, and the weaker the light intensity, the closer the value is to 0.

4. The method for adaptive arc area repair and enhancement according to claim 1, characterized in that: In step S3, the split image is divided into images of types I, II and III according to the arc interference model value calculated in step S2, including: Step S31: Category I: split images without arc light or with little arc light influence, in which case 0≤DLC(i,j)<m, need to be processed by uniform light intensity and contrast enhancement adaptive module; Step S32: Category II: The arc light causes a certain impact and the arc light area needs to be repaired. At this time, m≤DLC(i,j)<n, and it needs to be processed by the uniform light intensity, color restoration, and contrast enhancement adaptive modules; Step S33: Category III: The arc light is serious, causing irreparable damage to the picture and details. At this time, n≤DLC(i,j)≤1, and it needs to be processed by unified light intensity, multi-frame replacement, color restoration, and contrast enhancement adaptive modules.

5. The method for adaptive arc area repair and enhancement according to claim 1, characterized in that: In step S4, according to the split image classification performed in step S3, each split image is adaptively selected to be processed by: uniform light intensity, multi-frame replacement, color restoration, and contrast enhancement modules, including: Step S41: unify the light intensity module, which is implemented as shown in formula (5): L Gamma (i,j)=[L(i,j)] γ *255 (5) Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L Gamma (i, j) is the intensity factor of the split image (i, j) after unification; L(i, j) is the original intensity factor of the split image (i, j); γ is the unified intensity correction index, as shown in formula (6): γ=log2[L(i,j) aim -L(i,j) median ] (6) Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L(i,j) aim is the target brightness mean of the split image (i, j); L(i, j) median is the median brightness of the split image (i, j); Where i,j are the center coordinates of the split image; (i,j) is the split image whose center coordinates are i,j; L(i,j) mean is the mean brightness of the split image (i, j); Step S42: multi-frame replacement module, the Class III split image is seriously disturbed by arc light, causing irreparable damage to the picture and details, and the arc light area also has the phenomenon of covering loss. According to the arc light characteristics and the principle of inter-frame similarity, for the Class III image, the non-Class III split images of the four frames before and after it are subjected to multi-channel weighted average fusion processing to replace the original split image; Step S43: Color restoration module, which is implemented as shown in equations (8, 9, 10): The R channel transformation rule is shown in formula (8): The G channel transformation rule is shown in formula (9): The B channel transformation rule is shown in formula (10): Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N R (x, y) is the pixel value of the R channel of the split image (i, j); u is the reference parameter; are the maximum and minimum pixel values ​​of the G channel of the split image (i, j); U(i, j) is the variance of the split image (i, j); U R (i, j) is the variance of the R channel of the split image (i, j); μ R (i, j) is the mean of the R channel of the split image (i, j); Step S44: Contrast enhancement module, which needs to convert the split image format from RGB to HIS space to reduce the amount of calculation. Contrast enhancement only needs to operate on the I channel. Its implementation method is shown in formula (11): Where i, j are the center coordinates of the split image; (i, j) is the split image whose center coordinates are i, j; N I (x, y) is the pixel value of each pixel in the I channel of the split image (i, j); μ I (i,j) is the pixel mean of channel I of the split image (i,j); is the pixel standard deviation of channel I of the split image (i, j).

Images