An automatic image defogging processing method based on artificial intelligence

Abstract

The present invention belongs to the technical field of automatic image defogging processing, and relates to an automatic image defogging processing method based on artificial intelligence. By obtaining the image datasets of each area belonging to each designated transmission line tower in each shot, and further performing techniques such as polarization analysis, defogging processing, and image fusion on them, the defogged images of each area belonging to each designated transmission line tower are obtained, effectively reducing the impact of complex environments on target detection. Furthermore, the state of the transmission line corridor is clearly visible during various heavy fog periods, improving the state control ability of important transmission channels. By analyzing the evaluation of the defogging effect of the defogged images of each area belonging to each designated transmission line tower, and outputting and feeding it back, a more accurate and objective evaluation result is obtained, providing more accurate and reliable information for subsequent image analysis, processing, and decision-making, which is beneficial to observing the situation of the tower and the line more clearly.

Description

Technical Field

[0001] The present invention belongs to the technical field of automatic image dehazing, and relates to an automatic image dehazing method based on artificial intelligence. Background Art

[0002] In the natural environment, due to the existence of particulate matter, water vapor, smoke, etc. in the atmosphere, the target objects in the image are often blocked by haze, resulting in a decline in image quality and a blurred visual effect. Especially in bad weather conditions, such as haze days, rainy days, etc., the visibility and clarity of the image will be seriously affected, bringing difficulties to image processing and recognition. Transmission lines are the infrastructure for realizing power transmission and are the key to ensuring safe and stable power supply. However, transmission lines are often erected outdoors, with a harsh natural environment and a wide distribution range. As a result, the line channel images obtained by the transmission line visual inspection device are often blocked by fog, the image quality deteriorates, the state of the transmission channel and the ice coating state of the conductor cannot be seen completely, affecting the accurate analysis of the line operation state, it is difficult to detect potential hazards, and it is easy to misjudge the operation status, etc. Therefore, an automatic image dehazing system for transmission lines has very important significance and functions.

[0003] In the prior art, there are also some related solutions for image dehazing. For example, a patent application for an invention of an automatic image fog penetration processing method, system, terminal and storage medium with the Chinese patent publication number CN117392009A includes: using an image classification model to classify the fog concentration level of the image to be processed. If the fog concentration of the image to be processed is low, a dehazing algorithm is used to perform fog penetration processing on the image to be processed. If the fog concentration of the image to be processed is high, an optical lens is added for optical fog penetration, and color restoration is performed on the image obtained by optical fog penetration. The present invention uses an image classification model to classify the fog concentration level of the image to be processed, automatically enables a matching fog penetration processing strategy for the image to be processed based on the classification, realizes automatic fog penetration of the image, and automatically performs color compensation on the image after performing optical fog penetration, and finally obtains a clear color image.

[0004] Another invention patent application for a polarization-based image defogging method with Chinese Patent Publication No. CN118314045A includes: obtaining a foggy image and multiple polarization images corresponding to the foggy image, where the polarization angles of the multiple polarization images are different; constructing a network model, including a PEN module and a TDN module connected in sequence, and an SRN module is connected to the PEN module and the TDN module respectively; inputting the multiple polarization images into the network model, the PEN module obtains the polarization degree of transmitted light, the polarization degree of atmospheric light, the atmospheric light at infinity, the feature image, and the polarization degree of the feature image based on the multiple polarization images; the TDN module obtains the transmission map based on the polarization degree of transmitted light, the polarization degree of atmospheric light, the feature image, and the polarization degree of the feature image; the SRN module obtains the defogged image based on the atmospheric light at infinity, the feature image, and the transmission map. By using the polarization characteristics to achieve image defogging, the defogging performance for different degrees of haze scenes and real-world haze scenes is improved.

[0005] Although the above solution proposes some solutions for image defogging processing, there are still certain limitations: the existing solution classifies according to the fog concentration level and automatically enables the matching fog penetration processing strategy, or uses the polarization characteristics to achieve image defogging. However, the existing solution ignores the quantitative evaluation of the quality of the defogged image, so it is impossible to evaluate the results more accurately and objectively, and it is also impossible to process in time when the defogging effect is not ideal, which is not conducive to improving the usability of the image, and further not conducive to observing the situation of the tower and the line more clearly, and it is impossible to detect potential safety hazards in time and ensure the safe operation of the transmission line. Summary of the Invention

[0006] In view of this, to solve the problems proposed in the above background technology, a method for automatically processing image defogging based on artificial intelligence is proposed.

[0007] The object of the present invention can be achieved by the following technical solutions: The present invention provides a method for automatically processing image defogging based on artificial intelligence, including: S1. Image dataset acquisition: Divide each detected transmission line tower according to the region and record it as each designated transmission line tower belonging to each region. Then, according to a preset fixed monitoring interval period, each designated transmission line tower belonging to each region is photographed at a set short interval duration to obtain an image dataset of each designated transmission line tower belonging to each region in the current time period, and it is recorded as the image dataset of each designated transmission line tower belonging to each region in each photographing, where the image dataset includes each picture corresponding to each polarization angle.

[0008] S2. Image dataset analysis: Analyze the polarization parameters of each pixel point of the images of each designated transmission line tower belonging to each region in each photographing.

[0009] S3. Image dataset parsing: Analyze the fog influence degree index of each pixel point of the images of the poles and towers of each specified transmission line belonging to each region, and then obtain the sub-regions of each fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region.

[0010] S4. Obtaining defogged images: Analyze the atmospheric light intensity and transmission rate of the images of the poles and towers of each specified transmission line belonging to each region, and accordingly obtain the defogged images corresponding to each polarization angle of the poles and towers of each specified transmission line belonging to each region, and fuse them to obtain the defogged images of the poles and towers of each specified transmission line belonging to each region.

[0011] S5. Analyzing the defogging effect of images: Obtain the effect data parameters of the defogged images of the poles and towers of each specified transmission line belonging to each region, and analyze the defogging effect evaluation of the defogged images of the poles and towers of each specified transmission line belonging to each region.

[0012] S6. Management of defogged images: If the defogging effect evaluation of the defogged images of the poles and towers of each specified transmission line belonging to each region is excellent, output and feedback them; otherwise, execute S1.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. By obtaining the image datasets of the poles and towers of each specified transmission line belonging to each region in each shot, and further performing polarization analysis, defogging processing, image fusion and other technologies on them, the defogged images of the poles and towers of each specified transmission line belonging to each region are obtained, which helps to obtain higher image clarity and contrast, and at the same time retains more detailed information of the target, effectively reducing the influence of complex environments on target detection. Furthermore, the state of the transmission line corridor can be clearly visible during various heavy fog periods, potential hazards can be identified, the state control ability of important transmission channels is improved, safe operation is guaranteed, and the cost and difficulty of manual intervention are reduced.

[0014] 2. By obtaining the effect data parameters of the defogged images of the poles and towers of each specified transmission line belonging to each region, analyzing the defogging effect evaluation of the defogged images of the poles and towers of each specified transmission line belonging to each region, and outputting and feedbacking them, a more accurate and objective evaluation result can be obtained through quantitative evaluation of the quality of the defogged images, which helps to improve the usability of the images, provides more accurate and reliable information for subsequent image analysis, processing and decision-making, and further facilitates clearer observation of the poles and lines, timely discovery of potential safety hazards, and guarantee of the safe operation of the transmission lines. Description of the Drawings

[0015] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0016] Figure 1 It is a schematic diagram of the implementation steps of the method of the present invention. Specific implementation manners

[0017] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0018] Please refer to Figure 1 As shown, the present invention provides an automatic image dehazing processing method based on artificial intelligence, and the specific steps are as follows: S1. Image dataset acquisition: Divide each detected transmission line tower according to regions and record it as each designated transmission line tower belonging to each region. Then, according to a preset fixed monitoring interval period, each designated transmission line tower belonging to each region is photographed at a set short time interval to obtain an image dataset of each designated transmission line tower belonging to each region in the current time period, and it is recorded as the image dataset of each designated transmission line tower belonging to each region in each photograph. The image dataset includes images corresponding to each polarization angle.

[0019] S2. Image dataset analysis: Analyze the polarization parameters of each pixel point of the images of each designated transmission line tower belonging to each region in each photograph.

[0020] As a preferred feasible embodiment, the polarization parameters of each pixel point of the images of each designated transmission line tower belonging to each region in each photograph include degree of polarization and polarization angle.

[0021] As a preferred feasible embodiment, the specific acquisition method of the degree of polarization of each pixel point of the images of each designated transmission line tower belonging to each region in each photograph is: Extract the images corresponding to each polarization angle of each designated transmission line tower belonging to each region in each photograph, process them using an image processing library to obtain the light intensity values of each pixel point corresponding to the images corresponding to each polarization angle of each designated transmission line tower belonging to each region in each photograph, and then obtain the average light intensity values of each pixel point corresponding to each polarization angle of each designated transmission line tower belonging to each region in each photograph.

[0022] It should be further noted that the specific method for obtaining the average light intensity values of each pixel corresponding to the polarization angles of each designated transmission line tower in each area of each shooting is as follows: According to the calculation formula the average light intensity value G of each pixel corresponding to the polarization angles of each designated transmission line tower in each area of each shooting is obtained fabhd , where G′ fabhyd is the average light intensity value of the d-th pixel corresponding to the h-th polarization angle of the b-th designated transmission line tower in the a-th area of the f-th shooting, f = 1, 2,..., F, f is the number of each shooting, F is the number of shootings, a = 1, 2,..., i, a is the number of each area, i is the number of areas, b = 1, 2,..., B, b is the number of each designated transmission line tower, B is the number of designated transmission line towers, h = 1, 2,..., H, h is the number of each polarization angle, H is the number of polarization angles, y = 1, 2,..., Y, y is the number of each picture, Y is the number of pictures, d = 1, 2,..., D, d is the number of each pixel, D is the number of pixels.

[0023] Sort the average light intensity values of each pixel corresponding to the polarization angles of each designated transmission line tower in each area of each shooting from large to small according to different polarization angles, and obtain the maximum average light intensity value and the minimum average light intensity value of each pixel of each designated transmission line tower in each area in different polarization angles of each shooting, which are respectively denoted as where f = 1, 2,..., F, f is the number of each shooting, F is the number of shootings, a = 1, 2,..., i, a is the number of each area, i is the number of areas, b = 1, 2,..., B, b is the number of each designated transmission line tower, B is the number of designated transmission line towers, d = 1, 2,..., D, d is the number of each pixel, D is the number of pixels. According to the calculation formula the polarization degree P of each pixel of the image of each designated transmission line tower in each area of each shooting is obtained fabd .

[0024] As a preferred feasible embodiment, the specific method for obtaining the polarization angle of each pixel of the image of each designated transmission line tower in each area of each shooting is as follows: Select the average light intensity values of each pixel of each designated transmission line tower in each area of each shooting corresponding to the polarization angles of 0°, 45°, 90° and 135° of each shooting According to the calculation formula the polarization angle θ of each pixel of the image of each designated transmission line tower in each area of each shooting is obtained fabd .

[0025] S3. Image dataset parsing: Analyze the fog influence degree index of each pixel point of the images of the poles and towers of each specified transmission line belonging to each region, and then obtain the sub-regions of fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region.

[0026] As a preferred feasible embodiment, the specific analysis method for the fog influence degree index of each pixel point of the images of the poles and towers of each specified transmission line belonging to each region is: extract the polarization degree and polarization angle of each pixel point of the images of the poles and towers of each specified transmission line belonging to each region taken each time, and analyze the fog influence degree index of each pixel point of the images of the poles and towers of each specified transmission line belonging to each region. Where P( f+1 ) abd and θ( f+1 ) abd are respectively the polarization degree and polarization angle of the d-th pixel point of the image of the b-th specified transmission line pole and tower belonging to the a-th region taken at the (f + 1)-th time, and △P 0 and △θ 0 are respectively the set permitted polarization degree difference and permitted polarization angle difference.

[0027] It should be further noted that the set permitted polarization degree difference and permitted polarization angle difference can be 0.01 and 0.01 degrees respectively, which means that there can be a 1% deviation in both the polarization degree and polarization angle of the pixel point.

[0028] As a preferred feasible embodiment, the specific acquisition method for the sub-regions of fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region is: record the region enclosed by adjacent pixel points of the images of the poles and towers of each specified transmission line belonging to each region with the same fog influence degree index as the sub-region of fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region, and then obtain the sub-regions of fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region.

[0029] It should be further noted that after obtaining the sub-regions of fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region, different degrees of defogging treatment intensity are applied to the sub-regions of fog influence degree corresponding to the images of the poles and towers of each specified transmission line belonging to each region based on their fog concentration influence degree.

[0030] S4. Defogged image acquisition: Analyze the atmospheric light intensity and transmission rate of the images of the poles and towers of each specified transmission line belonging to each region, and accordingly obtain the defogged images corresponding to each polarization angle of the poles and towers of each specified transmission line belonging to each region, and fuse them to obtain the defogged images of the poles and towers of each specified transmission line belonging to each region.

[0031] As a preferred feasible embodiment, the specific analysis method for the atmospheric light intensity and transmittance of each specified transmission line tower image in each region is as follows: extract the average light intensity values of each pixel corresponding to each polarization angle of each specified transmission line tower in each region for each shot, and screen out the pixels corresponding to each polarization angle of each specified transmission line tower in each region with the largest average light intensity value for each shot. Denote them as the high-brightness pixels corresponding to each polarization angle of each specified transmission line tower in each region for each shot. Further, denote the average light intensity value of the high-brightness pixels corresponding to each polarization angle of each specified transmission line tower in each region for each shot as the atmospheric light intensity A of the image of each specified transmission line tower in each region ab 。

[0032] Extract the polarization degree of each pixel of the image of each specified transmission line tower in each region for each shot, and substitute it into a specific calculation formula to calculate the transmittance t of the image of each specified transmission line tower in each region ab 。

[0033] It should be further noted that the specific calculation formula is P = k * t, where P is the polarization degree of the pixel, t is the transmittance, and k is a constant

[0034] As a preferred feasible embodiment, the specific method for obtaining the defogged image corresponding to each polarization angle of each specified transmission line tower in each region is as follows: according to the calculation formula obtain the defogged light intensity value J of each pixel corresponding to each polarization angle of each specified transmission line tower in each region abhd ,where G fabhd is the average light intensity value of the d-th pixel corresponding to the h-th polarization angle of the b-th specified transmission line tower in the a-th region for the f-th shot, h = 1, 2,..., H, h is the number of each polarization angle, and H is the number of polarization angles

[0035] Use an image processing library to reconstruct the defogged light intensity values of each pixel corresponding to each polarization angle of each specified transmission line tower in each region to obtain the defogged image corresponding to each polarization angle of each specified transmission line tower in each region

[0036] It should be further noted that the specific operation of reconstructing the defogging light intensity values corresponding to the polarization angles of each specified transmission line tower in each region for each pixel point by using the image processing library includes: using the image processing library to create a new image data structure according to the polarization angles of each specified transmission line tower in each region corresponding to the size of the original image, and constructing correspondingly in the new image data structure according to the positions of each pixel point in the original image corresponding to the polarization angles of each specified transmission line tower in each region, and performing color adjustment, so as to obtain the defogging images corresponding to the polarization angles of each specified transmission line tower in each region.

[0037] It should be further noted that the specific operation of fusing the defogging images corresponding to the polarization angles of each specified transmission line tower in each region to obtain the defogging images of each specified transmission line tower in each region includes: using the feature point matching algorithm to find the corresponding feature points in the defogging images corresponding to the polarization angles of each specified transmission line tower in each region of each specified transmission line in each region, calculating the transformation matrix between the images according to the positions of the feature points, including parameters such as translation, rotation and scaling, and applying the transformation matrix to align all the images under a common coordinate system.

[0038] The pixel values of the defogging images corresponding to the polarization angles of each specified transmission line tower in each region of each specified transmission line in each region are weighted and summed according to a certain weight to obtain the fused pixel values, and then the defogging images of each specified transmission line tower in each region are obtained.

[0039] By obtaining the image data sets of each specified transmission line tower in each region for each shooting, and further performing polarization analysis, defogging processing, image fusion and other technologies on them, the defogging images of each specified transmission line tower in each region are obtained in the present invention, which helps to obtain higher image clarity and contrast, and at the same time retains more detailed information of the target, effectively reduces the influence of complex environment on target detection, and then realizes that the state of the transmission line channel is clearly visible during various heavy fog periods, identifies potential hazards, improves the state control ability of important transmission channels, ensures safe operation, and reduces the cost and difficulty of manual intervention.

[0040] S5. Analysis of defogging effect of images: Obtain the effect data parameters of the defogging images of each specified transmission line tower in each region, and analyze the defogging effect evaluation of the defogging images of each specified transmission line tower in each region.

[0041] As a preferred feasible embodiment, the effect data parameters of the defogging images of each specified transmission line tower in each region include peak signal-to-noise ratio, structural similarity index and information entropy.

[0042] It should be further noted that the specific method for obtaining the peak signal-to-noise ratio of the defogged images of the specified transmission line towers in each region is as follows: The defogged images of the specified transmission line towers in each region are grayscaled to obtain the defogged grayscale images of the specified transmission line towers in each region and the grayscale values Gray of their corresponding pixels abq where q = 1, 2,..., Q, q is the number of each pixel in the defogged grayscale image of the specified transmission line tower, and Q is the number of pixels in the defogged grayscale image of the specified transmission line tower. Similarly, the grayscale values Gray' of the corresponding pixels in the standard grayscale images of the specified transmission line towers in each region can be obtained abq .

[0043] According to the analysis formula the mean squared error MSE of the defogged images of the specified transmission line towers in each region is obtained ab .

[0044] According to the analysis formula the peak signal-to-noise ratio PSNR of the defogged images of the specified transmission line towers in each region is obtained ab where Gray max is the maximum possible pixel value of the image. For example, for an eight-bit grayscale image, Gray max = 255

[0045] The specific method for obtaining the structural similarity index of the defogged images of the specified transmission line towers in each region is as follows: According to the grayscale values of the corresponding pixels in the defogged grayscale images of the specified transmission line towers in each region and the grayscale values of the corresponding pixels in the standard grayscale images of the specified transmission line towers in each region, calculate the average brightness of the defogged grayscale images of the specified transmission line towers in each region and the average brightness of the standard grayscale images of the specified transmission line towers in each region, where and and calculate the standard deviation of the defogged grayscale images of the specified transmission line towers in each region and the standard deviation of the standard grayscale images of the specified transmission line towers in each region, where and

[0046]

[0047] According to the analysis formula the brightness contrast index l of the defogged images of the specified transmission line towers in each region is obtained ab where μ 0 is a set constant

[0048] According to the analysis formula the contrast comparison index c of the defogged images of the transmission line towers belonging to each region is obtained ab , where μ 0 ′ is a set constant.

[0049] According to the analysis formula the contrast comparison index s of the defogged images of the transmission line towers belonging to each region is obtained ab , where μ0″ is a set constant.

[0050] Analyze the structural similarity index SSIM of the defogged images of the transmission line towers belonging to each region ab = l ab *c ab *s ab .

[0051] The specific way to obtain the information entropy of the defogged images of the transmission line towers belonging to each region is as follows: respectively extract the gray values of each pixel point corresponding to the defogged grayscale images of the transmission line towers belonging to each region, and further perform statistics on them to obtain the number m of each gray value of the pixel points in the defogged grayscale images of the transmission line towers belonging to each region abz , where z = 1, 2,..., W, z is the number of each gray value, and W is the number of gray values. According to the analysis formula the information entropy Entropy of the defogged images of the transmission line towers belonging to each region is obtained ab .

[0052] As a preferred feasible embodiment, the specific analysis method for evaluating the defogging effect of the defogged images of the transmission line towers belonging to each region is: extract the peak signal-to-noise ratio, structural similarity index and information entropy of the defogged images of the transmission line towers belonging to each region, and record them as PSNR ab , SSIM ab , Entropy ab , and analyze the defogging effect evaluation coefficient of the defogged images of the transmission line towers belonging to each region where PSNR 0 , SSIM 0 , Entropy 0 are respectively the peak signal-to-noise ratio threshold, structural similarity index threshold and information entropy threshold of the defogged images of the transmission line towers extracted from the information database.

[0053] Compare the defogging effect evaluation coefficients of the defogged images of the specified transmission line poles and towers in each area with the set defogging effect evaluation coefficient threshold respectively. If the defogging effect evaluation coefficient of the defogged image of a specified transmission line pole and tower in a certain area is greater than the defogging effect evaluation coefficient threshold, record the defogging effect of the defogged image of the specified transmission line pole and tower in this area as excellent.

[0054] S6. Defogged image management: If the defogging effect evaluation of the defogged images of the specified transmission line poles and towers in each area is excellent, output and feedback them; otherwise, execute S1.

[0055] As a preferred feasible embodiment, the specific operation of outputting and feedbacking the defogged images of the specified transmission line poles and towers in each area with excellent defogging effect evaluation is as follows: upload the defogged images of the specified transmission line poles and towers in each area to the external network platform through the public network 4G channel; otherwise, re-execute S1 until the defogging effect evaluation is excellent.

[0056] The present invention obtains the effect data parameters of the defogged images of the specified transmission line poles and towers in each area, analyzes the defogging effect evaluation of the defogged images of the specified transmission line poles and towers in each area, and outputs and feedbacks them. By quantitatively evaluating the quality of the defogged images, more accurate and objective evaluation results can be obtained, which helps to improve the usability of the images, provides more accurate and reliable information for subsequent image analysis, processing and decision-making, and further helps to observe the poles and towers and lines more clearly, timely discover potential safety hazards, and ensure the safe operation of the transmission line.

[0057] The above content is only an example and illustration of the concept of the present invention. Those skilled in the art of the present technology make various modifications or supplements or use similar methods to replace the specific embodiments described, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all belong to the protection scope of the present invention.

Claims

1. An automatic image defogging method based on artificial intelligence, characterized in that: include: S1. Acquisition of image data set: Divide each detected transmission line tower into designated transmission line towers belonging to each region according to the region, and then photograph each designated transmission line tower belonging to each region at a set short-term interval according to a preset fixed monitoring interval period, and obtain an image data set of each designated transmission line tower belonging to each region photographed each time in the current time period, and record it as an image data set of each designated transmission line tower belonging to each region photographed each time, wherein the image data set includes each picture corresponding to each polarization angle; S2, image data set analysis: analyzing the polarization parameters of each pixel of each designated transmission line tower image of each area captured each time; S3, image data set analysis: analyzing the fog impact index of each pixel point of each designated transmission line tower image belonging to each area, and then obtaining each fog impact sub-area corresponding to each designated transmission line tower image belonging to each area; S4, obtaining defogging images: analyzing the atmospheric light intensity and transmission rate of the images of the designated transmission line towers in each region, and obtaining defogging images corresponding to the polarization angles of the designated transmission line towers in each region, and fusing them to obtain defogging images of the designated transmission line towers in each region; S5, image defogging effect analysis: obtaining effect data parameters of the defogging images of each designated transmission line tower in each area, and analyzing the defogging effect evaluation of the defogging images of each designated transmission line tower in each area; S6, defogging image management: if the defogging effect of the defogging image of each designated transmission line tower in each area is evaluated as excellent, it will be output and fed back, otherwise, S1 will be executed; The defogging images corresponding to the polarization angles of the designated transmission line towers in each area are fused to obtain the defogging images of the designated transmission line towers in each area. The specific operations include: using a feature point matching algorithm to find the corresponding feature points in the defogging images corresponding to the polarization angles of the designated transmission line towers in each area, calculating the transformation matrix between the images according to the positions of the feature points, including translation, rotation and scaling, and applying the transformation matrix to align all the images to a common coordinate system; The pixel values ​​of the defogging images corresponding to the polarization angles of the designated transmission line towers in each area are weightedly summed according to certain weights to obtain the fused pixel values, and then the defogging images of the designated transmission line towers in each area are obtained.

2. The method for automatic image defogging based on artificial intelligence according to claim 1, characterized in that: The polarization parameters of each pixel point of each designated transmission line tower image of each area captured each time include polarization degree and polarization angle; The effect data parameters of the defogging images of each designated transmission line tower in each area include peak signal-to-noise ratio, structural similarity index and information entropy.

3. The method for automatic image defogging based on artificial intelligence according to claim 2 is characterized in that: The specific method for obtaining the polarization degree of each pixel point of each designated transmission line tower image of each area captured each time is as follows: Extract the pictures corresponding to the polarization angles of the designated transmission line towers in each area of ​​each shooting, process them using the image processing library, obtain the light intensity value of each pixel point corresponding to the pictures corresponding to the polarization angles of the designated transmission line towers in each area of ​​each shooting, and then obtain the average light intensity value of each pixel point corresponding to the polarization angles of the designated transmission line towers in each area of ​​each shooting; The average light intensity values ​​of each pixel corresponding to each polarization angle of each designated transmission line tower in each area of ​​each shooting are sorted from large to small according to the different polarization angles, and the maximum average light intensity value and the minimum average light intensity value of each pixel of each designated transmission line tower in each area of ​​each shooting at different polarization angles are obtained, which are recorded as ,in , is the number of each shot, is the number of shots, , is the number of each area, is the number of regions, , is the number of each designated transmission line tower, is the number of transmission line towers specified, , is the number of each pixel, is the number of pixels, according to the calculation formula Get the polarization degree of each pixel of each designated transmission line tower image in each area taken each time .

4. The method for automatic image defogging based on artificial intelligence according to claim 3 is characterized in that: The specific method for obtaining the polarization angle of each pixel point of each designated transmission line tower image of each area captured each time is as follows: The average light intensity values ​​of each pixel point of each designated transmission line tower in each area of ​​each shooting at each polarization angle corresponding to each pixel point are screened to obtain the average light intensity values ​​of each pixel point of each designated transmission line tower in each area of ​​each shooting at 0 degree, 45 degree, 90 degree and 135 degree polarization angles. , according to the calculation formula Get the polarization angle of each pixel of each designated transmission line tower image in each area captured each time .

5. The method for automatic image defogging based on artificial intelligence according to claim 4 is characterized in that: The specific analysis method of the fog influence index of each pixel point of each designated transmission line tower image in each area is as follows: Extract the polarization degree and polarization angle of each pixel point of each designated transmission line tower image in each area taken each time, and analyze the fog influence index of each pixel point of each designated transmission line tower image in each area ,in Respectively The first shot Region belongs to The first of the specified transmission line tower images The degree of polarization and polarization angle of each pixel, They are the set allowable polarization degree difference and allowable polarization angle difference respectively.

6. The method for automatic image defogging based on artificial intelligence according to claim 5, characterized in that: The specific method for obtaining each sub-region of fog influence degree corresponding to each designated transmission line tower image in each region is as follows: The area surrounded by adjacent pixel points of each designated transmission line tower image belonging to each area with consistent fog impact degree index is recorded as the fog impact degree sub-area corresponding to each designated transmission line tower image belonging to each area, and then the fog impact degree sub-area corresponding to each designated transmission line tower image belonging to each area is obtained.

7. The method for automatic image defogging based on artificial intelligence according to claim 4, characterized in that: The specific analysis method of the atmospheric light intensity and transmission rate of each designated transmission line tower image in each area is as follows: The average light intensity value of each pixel corresponding to each polarization angle of each designated transmission line tower in each area of ​​each shooting is extracted, and the pixel corresponding to each polarization angle of each designated transmission line tower in each area of ​​each shooting with the largest average light intensity value is screened out, and recorded as the high-brightness pixel corresponding to each polarization angle of each designated transmission line tower in each area of ​​each shooting, and further the average light intensity value of the high-brightness pixel corresponding to each polarization angle of each designated transmission line tower in each area of ​​each shooting is recorded as the atmospheric light intensity of the image of each designated transmission line tower in each area. ; Extract the polarization degree of each pixel of each designated transmission line tower image of each area taken each time, and put it into a specific calculation formula to calculate the transmission rate of each designated transmission line tower image of each area. .

8. The method for automatic image defogging based on artificial intelligence according to claim 7, characterized in that: The specific method for obtaining the defogging image corresponding to each polarization angle of each designated transmission line tower in each area is as follows: According to the calculation formula Get the defogging light intensity value of each pixel corresponding to each polarization angle of each designated transmission line tower in each area ,in For the The first shot Region belongs to No. of designated transmission line towers The polarization angle corresponds to The average light intensity value of each pixel, , is the number of each polarization angle, is the number of polarization angles; The image processing library is used to reconstruct the defogging light intensity value of each pixel corresponding to each polarization angle of each designated transmission line tower in each area, so as to obtain the defogging image corresponding to each polarization angle of each designated transmission line tower in each area.

9. The method for automatic image defogging based on artificial intelligence according to claim 3, characterized in that: The specific analysis method for evaluating the defogging effect of the defogging images of the designated transmission line towers in each area is as follows: The peak signal-to-noise ratio, structural similarity index and information entropy of the dehazed images of each designated transmission line tower in each area are extracted and recorded as , analyze the defogging effect evaluation coefficient of the defogging images of each designated transmission line tower in each area ,in are the peak signal-to-noise ratio threshold, structural similarity index threshold and information entropy threshold of the dehazed image of the transmission line tower extracted from the information database; The defogging effect evaluation coefficients of the defogging images of each designated transmission line tower in each area are compared with the set defogging effect evaluation coefficient thresholds. If the defogging effect evaluation coefficient of the defogging image of a designated transmission line tower in a certain area is greater than the defogging effect evaluation coefficient threshold, the defogging effect evaluation of the defogging image of the designated transmission line tower in the area is recorded as excellent.

10. The method for automatic image defogging based on artificial intelligence according to claim 9, characterized in that: The specific operations for outputting and feeding back the defogging images of the designated transmission line towers in each area where the defogging effect is evaluated as excellent are as follows: The defogging images of the designated transmission line towers in each area are uploaded to the external network platform through the public network 4G channel. Otherwise, S1 is re-executed until the defogging effect is evaluated as excellent.

Images