A monitoring device and a monitoring method having line icing monitoring function

Abstract

The present application relates to the field of image matching technology, and specifically to a monitoring device and a monitoring method with a line ice monitoring function. The method combines the position distribution of edge pixels around the pixel points in the line ice image with the pixel value distribution around the pixel points at the same position in the line ice infrared image to obtain the melting characteristic value; according to the degree of distribution disorder of the melting characteristic value around the pixel point, the initial cost weight of the pixel point is obtained; the SAD matching algorithm and the Census algorithm are combined to match the pixel points in the left grayscale image of the line ice and the right grayscale image of the line ice, as well as the initial cost weight, to obtain the line ice depth map to measure the ice thickness. The present application obtains the initial cost weight through the contour line fluctuations and temperature differences of ice and water, and combines the SAD algorithm with the Census algorithm for analysis, so that the ice depth map is more consistent with the actual ice situation on the line, and the accuracy of the line ice thickness measurement is improved.

Description

Technical Field

[0001] The present application relates to the field of image matching technology, and in particular to a monitoring device and a monitoring method with a line icing monitoring function. Background Art

[0002] Traditional operation and inspection models often require a large amount of human resources to carry out operation and maintenance, inspection and power supply protection. However, with the rapid development of the power grid, the contradiction between the shortcomings of the traditional operation and inspection model, such as long cycle, low efficiency, great impact on climate and environment, and limited scope, and the rapidly growing scale of the power grid is becoming increasingly prominent, and an operation and maintenance monitoring method that can be applied to the rapidly growing power grid is urgently needed. Intelligent terminal monitoring devices designed based on the monitoring sensor technology and data analysis technology of the Internet of Things have emerged.

[0003] Icing of transmission lines, temperature measurement of transmission lines, tilting of towers, dancing, and insulation damage are common phenomena that affect the safety of transmission lines, and are also the focus of transmission line operation and maintenance monitoring. Among them, the increase of icing on transmission lines will increase the mechanical load of transmission lines, insulator strings, and towers, which may cause accidents such as flashover tripping, communication interruption, and tower collapse, seriously threatening the safe and stable operation of the power grid. Therefore, it is extremely important to timely understand the thickness of ice on high-voltage transmission lines, make appropriate judgments on the severity of icing on the power grid, and obtain accurate monitoring results so that staff can take preventive measures.

[0004] When the intelligent terminal transmits the collected line ice images to the background for image analysis, a large number of images in the database are usually used for matching analysis. However, a single matching algorithm usually only matches through a single image information, such as texture, pixel value, etc. For example, when using the Sum of Absolute Differences (SAD) algorithm to match line ice images, the difference is only measured by the sum of the absolute differences of the pixel values ​​in the two images, without considering the feature information of the image. It is easily affected by factors such as image noise and texture, resulting in a high possibility of mismatching, making the accuracy of the obtained disparity map low, resulting in a decrease in the accuracy of the ice thickness measurement of the transmission line. Summary of the invention

[0005] In order to solve the technical problem that a single image matching algorithm is difficult to match the texture information of the image well, resulting in inaccurate measurement of ice thickness on the transmission line, the purpose of this application is to provide a monitoring device and a monitoring method with line ice monitoring function. The technical solutions adopted are as follows:

[0006] The present application proposes a monitoring device and a monitoring method with a line icing monitoring function, the method comprising:

[0007] The original image acquired by the monitoring device is processed to obtain a line ice-covered image and a line ice-covered infrared image respectively;

[0008] According to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image, and the difference between the pixel values ​​of the pixel at the same position in the line ice-covered infrared image and the pixel in the preset window, the melting characteristic value of each pixel in the line ice-covered image is obtained;

[0009] According to the degree of disorder of the melting characteristic value in the preset window of each pixel in the line ice-covered image, an initial cost weight of each pixel in the line ice-covered image is obtained;

[0010] The left grayscale image of the line ice cover and the right grayscale image of the line ice cover are matched by using the SAD matching algorithm and the Census matching algorithm respectively, and the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of the line ice cover are obtained, and the line ice cover depth map is obtained by combining the initial cost weights;

[0011] Measure the ice thickness of transmission lines based on the line ice depth map.

[0012] Furthermore, the method for obtaining the melting characteristic value includes:

[0013] According to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image, the line fluctuation degree of each pixel in the line ice-covered image is obtained;

[0014] According to the difference between the pixel values ​​of each pixel in the infrared image of the ice-covered line and the pixel values ​​in the preset window, the temperature difference characteristic value of each pixel in the infrared image of the ice-covered line is obtained;

[0015] The melting characteristic value of each pixel in the line ice-covered image is obtained by combining the line fluctuation degree of each pixel in the line ice-covered image and the temperature difference characteristic value of the pixel at the same position in the line ice-covered infrared image; the temperature difference characteristic value and the line fluctuation degree are both positively correlated with the melting characteristic value.

[0016] Furthermore, the method for obtaining the line fluctuation degree includes:

[0017] According to the number of edge pixels in a preset window of each pixel in the line ice-covered image, edge feature pixels in the line ice-covered image are screened out;

[0018] For any edge feature pixel point in the line ice-covered image, curve fitting is performed on all edge pixel points in a preset window of the edge feature pixel point to obtain an edge curve segment of the edge feature pixel point, and the absolute value of the difference between the horizontal coordinates of the pixel coordinates of the two end points of the edge curve segment is used as the edge horizontal difference of the edge feature pixel point; the absolute value of the difference between the vertical coordinates of the pixel coordinates of the two end points of the edge curve segment is used as the edge vertical difference of the edge feature pixel point;

[0019] The line fluctuation degree of the edge feature pixel points in the line ice-covered image is obtained by combining the edge transverse difference value, the edge longitudinal difference value and the length of the preset window of the edge feature pixel points;

[0020] The line fluctuation of pixels other than edge feature pixels in the line ice-covered image is set to 0.

[0021] Furthermore, the method for acquiring edge feature pixel points includes:

[0022] Perform edge detection on the line ice-covered image to obtain edge pixel points in the line ice-covered image;

[0023] The pixel points whose ratio of the number of edge pixels in a preset window of each pixel point in the line ice-covered image to the total number of pixels in the preset window is greater than a preset edge threshold are taken as edge feature pixels in the line ice-covered image.

[0024] Furthermore, the calculation formula for the line fluctuation degree of the edge feature pixel points in the line ice-covered image is as follows:

[0025] ;

[0026] In the formula, is the line fluctuation degree of the i-th edge feature pixel point in the line ice-covered image; is the edge lateral difference of the i-th edge feature pixel point in the line ice-covered image; is the edge longitudinal difference value of the ith edge feature pixel point in the line ice image; L is the length of the preset window; arcsin() is the inverse sine function; cos() is the cosine function.

[0027] Furthermore, the method for obtaining the temperature difference characteristic value includes:

[0028] Obtain the ambient temperature at the time when the infrared image of the line ice coverage is obtained;

[0029] Select any pixel point in the infrared image of ice covering the line as the analysis pixel point, and use the difference between the pixel value of the analysis pixel point and the pixel value of each pixel point in its preset window as the temperature difference value of each pixel point in the preset window of the analysis pixel point;

[0030] When the pixel value of the analysis pixel is equal to the preset first value and the ambient temperature is greater than the preset second value, the pixel whose temperature difference value is less than the preset temperature threshold in the preset window of the analysis pixel is marked;

[0031] The ratio of the number of marked pixels in the preset window of the analysis pixel to the total number of pixels in the preset window is used as the temperature difference characteristic value of the analysis pixel.

[0032] Furthermore, the method for obtaining the initial cost weight includes:

[0033] The information entropy of the melting characteristic value of each pixel in the preset window of the line ice-covered image is obtained as the temperature difference disorder degree of each pixel in the line ice-covered image;

[0034] According to the melting characteristic value and the temperature difference chaos of each pixel in the line ice-covered image, the initial cost weight of each pixel in the line ice-covered image is obtained; the melting characteristic value and the temperature difference chaos are both positively correlated with the initial cost weight.

[0035] Furthermore, the method for obtaining the ice depth map includes:

[0036] Any pixel point in the left grayscale image of the line ice is used as the left pixel point, and the left pixel point and each pixel point in the right grayscale image of the line ice form a pair to be matched; the maximum value of the initial cost weights of the two pixels in the pair to be matched is used as the matching cost weight when matching the two pixels of the pair to be matched;

[0037] According to the matching cost weights of the pixels in the left grayscale image of the line ice and the right grayscale image of the line ice, the SAD matching cost and the Census matching cost, the fused disparity cost of each pixel in the left grayscale image of the line ice under the disparity value is obtained; each pixel in the left grayscale image of the line ice has at least two fused disparity costs;

[0038] The line ice-covered image is filtered to obtain a filtered ice-covered image, wherein the filtered ice-covered image includes: a filtered ice-covered left image and a filtered ice-covered right image; the filtered ice-covered left image is matched with the filtered ice-covered right image to obtain a fusion disparity cost of each pixel point in the filtered ice-covered left image under a disparity value;

[0039] The disparity value corresponding to the minimum value of the fusion disparity cost of the pixel point at the same position in the left grayscale image of the line ice cover and the left grayscale image of the filtered ice cover is used as the final disparity value of the pixel point in the left grayscale image of the line ice cover;

[0040] The final disparity value of each pixel in the left grayscale image of the line ice coverage constitutes an initial disparity map; the initial disparity map is optimized to obtain an optimized disparity map; and the optimized disparity map is converted into a line ice coverage depth map using a conversion formula between the disparity map and the depth map.

[0041] Furthermore, the method of obtaining the fusion disparity cost of each pixel in the left grayscale image of the line ice coverage under the disparity value according to the matching cost weights of the pixels in the left grayscale image of the line ice coverage and the right grayscale image of the line ice coverage, the SAD matching cost and the Census matching cost, includes:

[0042] The calculation formula of the fusion disparity cost is as follows:

[0043] ;

[0044] ;

[0045] ;

[0046] In the formula, is the fusion disparity cost of the j-th pixel in the left grayscale image of the line ice; d is the disparity value of the j-th pixel in the left grayscale image of the line ice; is the matching cost weight of the jth pixel in the left grayscale image of the line ice cover; is the SAD matching cost of the jth pixel in the left grayscale image of the line ice cover; is the Census matching cost of the jth pixel in the left grayscale image of the line ice cover; is a positive constant; is a normal number; exp() is an exponential function with the natural constant e as the base.

[0047] In a second aspect, an embodiment of the present application provides a monitoring device with a line icing monitoring function, the device comprising:

[0048] The data acquisition module is used to process the original image acquired by the monitoring device to obtain the line ice-covered image and the line ice-covered infrared image respectively;

[0049] An image analysis module is used to obtain a melting characteristic value of each pixel in the line ice-covered image according to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image and the difference between the pixel values ​​of the pixel at the same position in the line ice-covered infrared image and the pixel in the preset window;

[0050] A weight calculation module, used for obtaining an initial cost weight of each pixel in the line ice-covered image according to the degree of disorder of the melting characteristic value in a preset window of each pixel in the line ice-covered image;

[0051] An image matching module is used to match the left grayscale image of line ice coverage with the right grayscale image of line ice coverage using the SAD matching algorithm and the Census matching algorithm respectively, obtain the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of line ice coverage, and obtain the line ice coverage depth map in combination with the initial cost weight;

[0052] The ice thickness monitoring module is used to measure the ice thickness of the transmission line based on the line ice depth map.

[0053] This application has the following beneficial effects:

[0054] In an embodiment of the present application, a monitoring device is used to obtain an original image of an ice-covered area on a transmission line, and the degree of melting of ice in the ice-covered area is analyzed from two aspects: the difference in line fluctuation degree and temperature difference between liquid water and solid ice in the ice-covered area. The position distribution of edge pixels in a preset window of pixels in the ice-covered line image presents the degree of line fluctuation around the pixels. The difference in pixel values ​​between pixels and surrounding pixels in the infrared image of the line ice reflects the temperature difference distribution around the pixels. Comprehensive analysis makes the melting characteristic value present the degree of ice melting with higher accuracy. The distribution of the melting characteristic value presents the contour texture of the ice-covered area after melting, so that the initial cost weight is more in line with the ice melting scene. The SAD algorithm is easily affected by factors such as image noise and texture, and the Census algorithm can better compensate for these shortcomings. The SAD algorithm and the Census algorithm are combined to match the left grayscale image of the line ice with the right grayscale image of the line ice, thereby improving the accuracy of image matching. Combined with the initial cost weight, the obtained line ice depth map is more suitable for the line ice scene and has a higher accuracy, thereby improving the accuracy of the ice thickness measurement of the transmission line. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0056] Figure 1 A method flow chart of a monitoring method with line icing monitoring function provided by an embodiment of the present application;

[0057] Figure 2 A schematic diagram of edge curve segment analysis of edge feature pixel points provided by one embodiment of the present application;

[0058] Figure 3A specific implementation flow chart of the fusion feature value provided by an embodiment of the present application;

[0059] Figure 4 A schematic diagram of the composition structure of a monitoring device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0060] A monitoring device and a monitoring method embodiment with line icing monitoring function:

[0061] In order to further explain the technical means and effects adopted by the present application to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, describes in detail a monitoring device and a monitoring method with line icing monitoring function proposed in the present application, its specific implementation method, structure, features and effects. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0062] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0063] The following is a detailed description of a specific scheme of a monitoring method with line icing monitoring function provided by the present application in conjunction with the accompanying drawings.

[0064] See also Figure 1 , which shows a method flow chart of a monitoring method with line icing monitoring function provided by an embodiment of the present application, the method comprising:

[0065] Step S001, processing the original image acquired by the monitoring device to obtain a line ice-covered image and a line ice-covered infrared image respectively.

[0066] Specifically, the monitoring device is fixed on the high-voltage line tower to make the monitoring device work normally, so as to use the built-in binocular camera of the monitoring device to collect the original image of the ice-covered area on the transmission line, and obtain the line ice-covered RGB left image and the line ice-covered RGB right image. Secondly, two thermal imaging cameras are fixed near the two cameras of the binocular camera, and the image obtained by the thermal imaging camera fixed near the left camera of the binocular camera is used as the initial ice-covered left infrared image; the image obtained by the thermal imaging camera fixed near the right camera of the binocular camera is used as the initial ice-covered right infrared image. It should be noted that the line ice-covered RGB left image corresponds to the pixels in the initial ice-covered left infrared image one by one, and the line ice-covered RGB right image corresponds to the pixels in the initial ice-covered right infrared image one by one.

[0067] Since the lens in the binocular camera deviates during the manufacturing process or the camera is assembled, the original image obtained will inevitably be distorted. In order to avoid the radial distortion or tangential distortion in the acquired left and right images affecting the results of subsequent analysis, the camera parameters are obtained by image correction. The Zhang calibration method is used to calibrate the line ice RGB left image and the line ice RGB right image, respectively, to obtain the line-aligned calibrated ice left image and calibrated ice right image.

[0068] Among them, image correction and Zhang's calibration method are well-known technologies to those skilled in the art and will not be described in detail here.

[0069] Furthermore, the top and left and right sides of the transmission line are covered with smooth ice. Due to factors such as temperature, the ice on the transmission line melts into water and freezes into ice below the transmission line, resulting in different shapes of ice ridges under the transmission line, with gaps in the middle. Due to the thickness of ice covering the transmission line, this application only analyzes the ice-covered area on the transmission line, and uses a convolutional neural network to obtain the line ice-covered area in the calibrated ice-covered left image, the calibrated ice-covered right image, the initial ice-covered left infrared image, and the initial ice-covered right infrared image, respectively, to obtain the ice-covered area left image, the ice-covered area right image, the line ice-covered left infrared image, and the line ice-covered right infrared image in turn. The specific process is as follows:

[0070] s1: respectively obtain A ice-covered left image, a calibrated ice-covered right image, an initial ice-covered left infrared image, and an initial ice-covered right infrared image. In this embodiment, the size of A takes an empirical value of 500;

[0071] s2: Use the LabelImg image annotation tool to label the pixel labels of the ice-covered area and the non-ice-covered area in the ice-covered left image, the calibrated ice-covered right image, the initial ice-covered left infrared image, and the initial ice-covered right infrared image as 1 and 0 respectively;

[0072] s3: Encode the label and the ice-covered left image, the calibrated ice-covered right image, the initial ice-covered left infrared image, and the initial ice-covered right infrared image respectively to obtain a training set for training the neural network model;

[0073] s4: The neural network structure is a convolutional neural network, the loss function is a cross entropy loss, and the optimizer is an adam optimizer. The training set in s3 is used to train the neural network to obtain a neural network model that can identify ice-covered areas. Neural network training is a well-known technology, and the specific process will not be repeated here.

[0074] s5: Input the ice-covered left image, calibrated ice-covered right image, initial ice-covered left infrared image and initial ice-covered right infrared image obtained on the transmission line into the trained neural network model, and obtain the ice-covered area left image, ice-covered area right image, line ice-covered left infrared image and line ice-covered right infrared image in turn.

[0075] Furthermore, weighted average grayscale is used to grayscale the left image of the ice-covered area and the right image of the ice-covered area, respectively, to obtain the left grayscale image of the line ice-covered and the right grayscale image of the line ice-covered. Among them, weighted average grayscale is a common technology in the field of image processing and will not be described in detail here.

[0076] The left grayscale image of ice-covered railway line and the right grayscale image of ice-covered railway line are recorded as ice-covered railway line images, and the left infrared image of ice-covered railway line and the right infrared image of ice-covered railway line are recorded as ice-covered railway line infrared images. It should be noted that the ice-covered railway line images correspond to the pixels in the ice-covered railway line infrared images one by one.

[0077] Step S002: Obtain the melting characteristic value of each pixel in the line ice-covered image according to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image, and the difference between the pixel value of the pixel at the same position in the line ice-covered infrared image and the pixel value in its preset window.

[0078] Specifically, for the ice-covered area on the transmission line, the contour shapes of the ice and the water formed by the melting of the ice are different. The position distribution of the edge pixels in the preset window of the pixel points in the line ice-covered image presents the edge contour information around the pixel points; due to the high thermal conductivity of water, water droplets will show different temperature characteristics on the ice surface. The difference in pixel values ​​between the pixel points in the line ice-covered infrared image and the surrounding pixel points reflects the melting of the ice. The edge information and pixel values ​​are used for comprehensive analysis to make the analyzed melting characteristic values ​​more accurate.

[0079] Preferably, the specific method for obtaining the melting characteristic value of each pixel in the line ice-covered image is as follows: first, according to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image, the line fluctuation degree of each pixel in the line ice-covered image is obtained; secondly, according to the difference between the pixel value of each pixel in the line ice-covered infrared image and the pixel value in its preset window, the temperature difference characteristic value of each pixel in the line ice-covered infrared image is obtained; thereafter, the melting characteristic value of each pixel in the line ice-covered image is obtained by combining the line fluctuation degree of each pixel in the line ice-covered image and the temperature difference characteristic value of the pixel at the same position of the pixel in the line ice-covered infrared image; the temperature difference characteristic value and the line fluctuation degree are both positively correlated with the melting characteristic value.

[0080] The specific process of obtaining the fusion feature value of the pixel points in the line ice image is as follows: Figure 3 As shown, it includes steps 301, 302 and 303:

[0081] 301, obtaining the line fluctuation degree of each pixel point in the ice-covered line image.

[0082] Preferably, the specific calculation method of the line fluctuation is: according to the number of edge pixels in a preset window of each pixel in the line ice-covered image, the edge feature pixels in the line ice-covered image are screened out; for any edge feature pixel in the line ice-covered image, curve fitting is performed on all edge pixels in the preset window of the edge feature pixel to obtain the edge curve segment of the edge feature pixel, and the absolute value of the difference between the horizontal coordinates of the pixel coordinates of the two end points of the edge curve segment is used as the edge horizontal difference of the edge feature pixel; the absolute value of the difference between the vertical coordinates of the pixel coordinates of the two end points of the edge curve segment is used as the edge vertical difference of the edge feature pixel; the line fluctuation of the edge feature pixel in the line ice-covered image is obtained by combining the edge horizontal difference, the edge vertical difference and the length of the preset window of the edge feature pixel; the line fluctuation of the pixel points that are not edge feature pixels in the line ice-covered image is set to 0.

[0083] The number of edge pixels in the preset window of pixels in the line ice-covered image affects the accuracy of the edge contour information around the pixels. Edge feature pixels, i.e. pixels with more surrounding edge contour information, are screened out. The contour information around these pixels is further analyzed to obtain the line fluctuation of the pixels.

[0084] Preferably, the specific method for obtaining edge feature pixel points is: performing edge detection on the line ice-covered image to obtain edge pixel points in the line ice-covered image; taking the pixel point whose ratio of the number of edge pixel points in a preset window of each pixel point in the line ice-covered image to the total number of pixel points in the preset window is greater than a preset edge threshold as the edge feature pixel point in the line ice-covered image.

[0085] The Canny edge detection operator is used to obtain edge pixels in the line ice-covered image; when the number of edge pixels in the preset window of the pixel in the line ice-covered image is more, it means that there is more contour information of the ice-covered area in the local area of ​​the pixel, and the edge line contour in the local area of ​​the pixel is more valuable for analysis. The edge feature points are screened out according to the number of edge line pixels in the preset window of the pixel, and the preset window of the edge feature pixel contains more contour information. Among them, the Canny edge detection operator is a well-known technology for those skilled in the art and will not be repeated here.

[0086] In the embodiment of the present application, the size of the preset window is , L takes the empirical value of 5, and the preset edge threshold takes the empirical value of 0.7. The implementer can set it according to the specific situation. The preset window of the pixel point is centered on the pixel point and has a size of window.

[0087] The ice in the ice-covered area on the transmission line is solid, and the lines are mostly straight lines; water is liquid, and the lines are relatively soft and mostly arcs. Therefore, a curve fitting is performed on the edge pixels in the preset window of each edge feature pixel to obtain the edge curve segment of each edge feature pixel; the degree of fluctuation of the contour line in the preset window of the edge feature pixel is presented according to the curvature of the edge curve segment, and then the possibility of water droplets or ice around the edge feature pixel is judged.

[0088] In the embodiment of the present application, the least square method is used to perform curve fitting on edge pixel points within a preset window of edge feature pixel points.

[0089] Take the i-th edge feature pixel in the line ice image Take the analysis as an example, Figure 2 A schematic diagram of edge curve segment analysis of edge feature pixel points provided in an embodiment of the present application is shown in FIG. Figure 2 As shown, the curve segment L1 is the edge feature pixel point The edge curve segment L1 has two endpoints, point a and point b, and point c is the midpoint of the edge curve segment L1. This application uses the size of ∠acb to measure the edge feature pixel point The fluctuation degree of the contour line in the preset window of acb; when acb is larger, it means that the edge curve segment L1 is relatively gentle, and it is more likely to present the ice contour; when acb is smaller, it means that the fluctuation and change degree of the edge curve segment L1 is larger, and it is more likely to present the water drop contour. It should be noted that acb can be obtained by using geometric operations.

[0090] As an example, the following method is provided to obtain ∠acb. The specific process is as follows:

[0091] Get the pixel coordinates of point a and point b in the line ice image, and take the absolute value of the difference between the horizontal coordinates of the pixel coordinates of point a and point b as the edge feature pixel point The edge cross difference , the absolute value of the difference between the vertical coordinates of the pixel coordinates of point a and point b is used as the edge feature pixel point The edge vertical difference , then the length of line segment ab is .

[0092] To facilitate the calculation of ∠acb, since the edge curve segment L1 is within the preset window, this application uses As the radius, draw circles with point a and point b as the center, and point z is the intersection of the two circles; with point z as the center, A circle is drawn with radius , and circle Z is obtained. The edge curve segment L1 is a partial arc segment of circle Z. The edge curve segment L1 and the dotted curve segment L2 constitute the circumference of circle Z.

[0093] It should be noted that the default window size is , to prevent the number of edge pixels in the preset window from being too large, so as to ensure that the edge curve segment formed by the edge pixels in the preset window of pixels is the local area of ​​the water drop, so that the edge curve segment cannot exceed half of the circumference of the circle, and thus the value range of ∠acb is .

[0094] Since point c is the midpoint of edge curve segment L1, line segment cd is perpendicular to line segment ab, that is, △bez is a right triangle. Take the length of line segment be and the length of line segment bz The arc sine value of the ratio, that is, ∠bzc is equal to Since the central angle of a circle is equal to twice the circumferential angle of the same arc, ∠azb and ∠adb both correspond to the edge curve segment L1, so ∠azb=2∠adb, and since ∠azb=2∠bzc, ​​∠bzc=∠adb. The angles of ∠cbd and ∠cad are , then ∠acb= It should be noted that the line segment cd is the diameter of the circle Z, the line segments zc, zb, za and zd are the radii of the circle Z, and the point e is the midpoint of the line segment ab.

[0095] Since the value range of ∠acb is To ensure that the larger the ∠acb is, the smaller the line fluctuation of the edge curve segment L1 is, the calculation formula for the line fluctuation of the edge feature pixel point in the line ice image is as follows:

[0096] ;

[0097] In the formula, is the line fluctuation degree of the i-th edge feature pixel point in the line ice-covered image; is the edge lateral difference of the i-th edge feature pixel point in the line ice-covered image; is the edge longitudinal difference value of the ith edge feature pixel point in the line ice image; L is the length of the preset window; arcsin() is the inverse sine function; cos() is the cosine function.

[0098] It should be noted that when When the value is larger, it means that the edge curve segment in the preset window of the edge feature pixel is relatively gentle, and the possibility of ice contour near the edge feature pixel is relatively high. The smaller; when The smaller the value, the greater the fluctuation of the edge curve segment in the preset window of the edge feature pixel point, and the greater the possibility of a water droplet outline near the edge feature pixel point. The bigger.

[0099] For pixels other than edge feature pixels in the line ice-covered image, there is little edge contour information in the preset window of such pixels and they have no reference value. Therefore, the line fluctuation degree of pixels other than edge feature pixels in the line ice-covered image is set to 0.

[0100] 302, obtaining a temperature difference characteristic value of each pixel in the line ice-covered image.

[0101] Since there is a temperature difference between water and ice, the difference between the pixel values ​​of a pixel point in the line ice-covered infrared image and its surrounding pixels reflects the temperature difference, and thus reflects the water and ice conditions around the pixel point.

[0102] Preferably, the specific calculation method of the temperature difference characteristic value is: obtain the ambient temperature at the moment when the infrared image of the line ice coverage is obtained; select any pixel point in the infrared image of the line ice coverage as the analysis pixel point, and use the difference between the pixel value of the analysis pixel point and each pixel point in its preset window as the temperature difference value of each pixel point in the preset window of the analysis pixel point; when the pixel value of the analysis pixel point is equal to the preset first value and the ambient temperature is greater than the preset second value, mark the pixels in the preset window of the analysis pixel point whose temperature difference value is less than the preset temperature threshold; use the ratio of the number of marked pixels in the preset window of the analysis pixel point to the total number of pixels in its preset window as the temperature difference characteristic value of the analysis pixel point.

[0103] In terms of temperature, the temperature of ice is almost less than or equal to , the water temperature is greater than . Analyze in combination with ambient temperature: When the ambient temperature is greater than When the ambient temperature is less than When the ambient temperature is equal to When the temperature of ice water is lower than that of ice water, the state of ice water remains unchanged. Since the thermal conductivity of water is relatively high, water droplets will show different temperature characteristics on the ice surface. Therefore, in the embodiment of the present application, the first value is preset to take the empirical value of 0, the second value is preset to take the empirical value of 0 degrees Celsius, and the temperature threshold is preset to take the empirical value of 2. The implementer can set it according to the specific situation.

[0104] Take the mth pixel in the infrared image of ice on the line With the k-th pixel in its preset window Take the pixel point as an example for analysis. With pixels The difference between the pixel values ​​is taken as the pixel point There are two situations where the pixel The edge points that may be the water droplet formation area are:

[0105] The first case: when the pixel The pixel value is large, and the pixel point When the temperature difference is less than 2, the pixel is the edge point of the area where water droplets may form; the second case: when the pixel point The pixel value is equal to 0, and the ambient temperature at the time of obtaining the infrared image of the line ice coverage is greater than When the ice on the line melts, water droplets may appear, then the pixel It may be the edge point of the water droplet formation area; in this case, the pixel point mark.

[0106] Pixels The number of marked pixels in the preset window can represent the melting of ice near the pixel. The number of marked pixels in the preset window; the larger the number, the The greater the degree of melting of nearby ice.

[0107] 303, obtaining a melting characteristic value of each pixel in the line ice-covered image.

[0108] When the line fluctuation of the pixel point in the line ice-covered image is greater, the possibility of the existence of water droplet formation area around the pixel point is greater; when the temperature difference characteristic value of the pixel point in the line ice-covered infrared image is greater, the possibility of the existence of edge points of the water droplet formation area around the pixel point is greater; in summary, when there is a certain degree of fluctuation in the lines around the pixel point and the probability of temperature difference is greater, the degree of melting of the ice-covered area is greater.

[0109] In an embodiment of the present application, the product of the line fluctuation of a pixel point in the line ice-covered image and the temperature difference characteristic value of the pixel point at the corresponding position in the line ice-covered infrared image is used as the melting characteristic value of the pixel point in the line ice-covered image; in other embodiments of the present application, the sum of the line fluctuation and the temperature difference characteristic value can also be used as the melting characteristic value.

[0110] The calculation formula for the melting characteristic value of each pixel in the line ice image is as follows:

[0111] ;

[0112] In the formula, is the melting characteristic value of the jth pixel in the line ice image; is the line fluctuation degree of the jth pixel in the line ice image; It is the temperature difference characteristic value of the pixel point at the corresponding position in the infrared image of the line ice coverage at the jth pixel point in the line ice coverage image.

[0113] It should be noted that when the line fluctuation The larger the value, the more likely it is that the pixels around the pixel are the edge contour of the water droplet, and the melting eigenvalue The larger the temperature difference characteristic value is, the When the value is larger, it means that the difference between the temperature value of the pixel and its surrounding pixels is greater, and the possibility of melting of the ice-covered area at the pixel location is higher. The bigger.

[0114] At this point, the melting characteristic value of each pixel in the line ice-covered image is obtained.

[0115] Step S003: obtaining an initial cost weight of each pixel in the line ice-covered image according to the degree of confusion of the melting feature value within a preset window of each pixel in the line ice-covered image.

[0116] When the ice in the ice-covered area on the transmission line has not melted, the outline of the ice-covered area is mostly lines. When the ice melts and produces water, the texture of the outline of the ice-covered area changes. The chaos of the melting feature values ​​of the pixels in the preset window of the pixels presents the texture of the melted ice. Based on the initial cost weight that represents the ice texture, the matching status of the left grayscale image of the line ice and the right grayscale image of the line ice is adaptively adjusted to improve the matching accuracy.

[0117] Preferably, the specific method for obtaining the initial cost weight is: obtaining the information entropy of the melting characteristic value of each pixel in the preset window of each pixel in the line ice-covered image as the temperature difference disorder of each pixel in the line ice-covered image; obtaining the initial cost weight of each pixel in the line ice-covered image according to the melting characteristic value and the temperature difference disorder of each pixel in the line ice-covered image; the melting characteristic value and the temperature difference disorder are both positively correlated with the initial cost weight. Among them, the calculation formula of the information entropy is a well-known technology for those skilled in the art and will not be repeated here.

[0118] When the temperature difference chaos is greater, the texture richness of the area around the pixel is greater; when the melting eigenvalue is greater, the probability of water droplets appearing around the pixel is higher, resulting in an uneven surface of the ice-covered line; therefore, a larger matching weight is required. In order to ensure that the melting eigenvalue and the temperature difference chaos are both positively correlated with the initial cost weight in the embodiment of the present application, the normalized result of the product of the fusion eigenvalue and the temperature difference chaos is used as the initial cost weight; in another embodiment of the present application, the sum of the melting eigenvalue and the temperature difference chaos can be used as the initial cost weight, and other basic mathematical operations can be selected in other embodiments of the present application to form a positive correlation, which is not limited here.

[0119] The calculation formula of the initial cost weight of each pixel in the line ice image is as follows:

[0120] ;

[0121] In the formula, is the initial cost weight of the jth pixel in the line ice-covered image; is the melting characteristic value of the jth pixel in the line ice image; is the temperature difference disorder of the jth pixel in the line ice image; Norm is the normalization function.

[0122] It should be noted that when the melting characteristic value When the value is larger, the probability of water droplets melting around the pixel is higher, which makes the ice surface of the line uneven and requires a larger initial cost weight. The larger the temperature difference, the more chaotic the When the value is larger, the texture of the area around the pixel is richer, and the initial cost weight The bigger.

[0123] Step S004: Use the SAD matching algorithm and the Census matching algorithm to match the left grayscale image of the line ice coverage with the right grayscale image of the line ice coverage, respectively, obtain the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of the line ice coverage, and combine the initial cost weight to obtain the ice coverage depth map.

[0124] When matching the left grayscale image of the line ice cover with the right grayscale image of the line ice cover, the SAD algorithm is used for fast matching speed and strong real-time performance in the edge areas of images with rich grayscale information and texture, but it is sensitive to the uniformity and noise of light intensity and has poor robustness; the improved Census matching algorithm has higher robustness and matching accuracy, but it is prone to mismatching in areas of discontinuous disparity. Therefore, it is difficult to obtain a disparity map with high accuracy in actual scenes by using only one of the matching algorithms. The present invention combines the SAD matching algorithm with the improved Census matching algorithm for analysis to obtain a disparity map with high accuracy. Among them, the SAD matching algorithm and the Census matching algorithm are both well-known technologies to those skilled in the art and will not be described in detail here.

[0125] The specific method for obtaining the line ice depth map is as follows: first, any pixel point in the left grayscale image of the line ice is taken as the left pixel point, and the left pixel point and each pixel point in the right grayscale image of the line ice form a pair to be matched; the maximum value of the initial cost weights of the two pixels in the pair to be matched is taken as the matching cost weight when matching the two pixels of the pair to be matched; according to the matching cost weights of the pixels in the left grayscale image of the line ice and the right grayscale image of the line ice; secondly, the fused disparity cost of each pixel point in the left grayscale image of the line ice under the disparity value is obtained; there are at least two fused disparity costs for each pixel point in the left grayscale image of the line ice, including the SAD matching cost and the Census matching cost; then, The line ice-covered image is filtered to obtain a filtered ice-covered image, which includes: a filtered ice-covered left image and a filtered ice-covered right image; the filtered ice-covered left image is matched with the filtered ice-covered right image to obtain the fusion disparity cost of each pixel in the filtered ice-covered left image under the disparity value; the disparity value corresponding to the minimum value of the fusion disparity cost of the pixel at the same position in the line ice-covered left grayscale image and the filtered ice-covered left image is used as the final disparity value of the pixel in the line ice-covered left grayscale image; the final disparity value of each pixel in the line ice-covered left grayscale image constitutes an initial disparity map; the initial disparity map is optimized to obtain an optimized disparity map; and the optimized disparity map is converted into a line ice-covered depth map by using a conversion formula between the disparity map and the depth map.

[0126] As an example, take the pixel in the grayscale image of the line covered with ice Take the example of analysis, when the left grayscale image of line ice coverage is stereo matched with the right grayscale image of line ice coverage, the pixel point Each pixel in the grayscale image of the line ice is matched with the pixel in the grayscale image of the line ice. Use SAD algorithm and Census algorithm to obtain pixel points With pixels The SAD matching cost and Census matching cost between them; Since the line ice image has rich texture, the SAD matching algorithm is the main algorithm for matching the line ice image. Therefore, the pixel point With pixels The maximum value of the initial cost matching weight is used as the matching cost weight when two pixels are matched, and is used as the weight of the SAD matching cost, and combined with the Census matching cost to obtain the pixel point The fusion disparity cost under the disparity value.

[0127] It should be noted that due to the pixel There is a fusion disparity cost under the disparity value between each pixel in the right grayscale image of the line ice.

[0128] The calculation formula of the fusion disparity cost is as follows:

[0129] ;

[0130] ;

[0131] ;

[0132] In the formula, is the fusion disparity cost of the j-th pixel in the left grayscale image of the line ice; d is the disparity value of the j-th pixel in the left grayscale image of the line ice; is the matching cost weight of the jth pixel in the left grayscale image of the line ice cover; is the SAD matching cost of the jth pixel in the left grayscale image of the line ice cover; is the Census matching cost of the jth pixel in the left grayscale image of the line ice cover; It is a normal number, and its empirical value is 1; is a normal number with an empirical value of 1; exp() is an exponential function with the natural constant e as the base.

[0133] It should be noted that is a normalization function used to control the matching cost calculated by the SAD matching algorithm and the Census matching algorithm. ; When the pixel point is located in the edge area of ​​the image with rich texture, the SAD matching algorithm is given a higher matching cost weight. When the image is smooth and the texture is repeated, the Census matching algorithm is given a higher matching cost weight.

[0134] Since the cost calculation only considers local correlation and is very sensitive to noise, it cannot be used directly to calculate the optimal disparity value. The guided filtering algorithm is used to denoise the left grayscale image of the line ice-covered and the right grayscale image of the line ice-covered to obtain the filtered ice-covered left image and the filtered ice-covered right image, collectively referred to as filtered ice-covered images. The filtered ice-covered left image and the filtered ice-covered right image are stereo matched to obtain the fused disparity value of each pixel in the filtered ice-covered left image under the disparity value. It should be noted that the method for obtaining the fused disparity value of each pixel in the filtered ice-covered left image under the disparity value is the same as the method for obtaining the fused disparity value of each pixel in the line ice-covered left grayscale image under the disparity value, which will not be introduced in detail here.

[0135] Since the line ice image corresponds to the pixels in the filtered ice image one by one, the pixel points in the left grayscale image of the line ice are Analysis, assuming pixel There are 100 fusion disparity costs between the filtered ice image and the pixel point There are 100 fusion disparity costs for the corresponding pixel point, and the disparity value corresponding to the minimum value of these 200 fusion disparity costs is taken as the pixel point According to the above method, the final disparity value of each pixel in the left grayscale image of the line ice is obtained.

[0136] The initial disparity map is formed by the final disparity value of each pixel in the left grayscale image of the line ice cover. Since the initial disparity map result is relatively rough and may contain occlusion points and mismatched points, the left-right consistency detection is used to reallocate the disparity values, and the effective difference of the detected occlusion points is calculated by disparity filling. Finally, sub-pixel fitting is used to make the initial disparity result map more continuous to obtain an optimized disparity map. Among them, left-right consistency detection, disparity filling, and sub-pixel fitting are well-known technologies, and the specific content will not be repeated.

[0137] According to the camera parameters obtained in the binocular vision calibration, combined with the conversion formula of the disparity map and the depth map, the optimized disparity map is converted into the line ice depth map. Among them, the conversion formula of the disparity map and the depth map is a well-known technology, and the specific content is not repeated here.

[0138] Step S005: Measure the ice thickness of the transmission line based on the line ice depth map.

[0139] Use a binocular camera to obtain an image of ice-covered transmission lines and obtain a depth map of the line without ice. It should be noted that the line without ice depth map corresponds to the pixels in the line ice depth map one by one. The absolute value of the difference between the depth value of each pixel in the line ice depth map and the depth value of the corresponding pixel in the line without ice depth map is used as the ice thickness of each pixel. When the ice thickness of each pixel in the line ice depth map is greater than the preset thickness, it is considered that the ice thickness of the line area has reached the warning value and needs to be cleaned in time.

[0140] In the embodiment of the present application, the preset thickness is an empirical value of 30 mm, and the implementer can set it specifically according to factors such as the surrounding environment and temperature.

[0141] This application is now complete.

[0142] In summary, in the embodiment of the present application, the position distribution of edge pixels around pixels in the line ice image is combined with the distribution of pixel values ​​around the pixels at the same position in the line ice infrared image to obtain the melting characteristic value; the initial cost weight of the pixel is obtained according to the degree of distribution disorder of the melting characteristic value around the pixel; the SAD matching algorithm and the Census algorithm are combined to match the pixel points in the left grayscale image of the line ice and the right grayscale image of the line ice, as well as the initial cost weight, to obtain the line ice depth map to measure the ice thickness. The present application obtains the initial cost weight through the contour line fluctuations and temperature differences of ice and water, and combines the SAD algorithm with the Census algorithm for analysis, so that the ice depth map is more accurate and the accuracy of ice thickness measurement of transmission lines is improved.

[0143] Transmission line ice depth conversion method embodiment based on stereo matching algorithm:

[0144] A binocular camera is used to capture ice-covered lines. The existing technology uses the Sum of Absolute Differences (SAD) algorithm to perform stereo matching on the obtained left grayscale image of the ice-covered line and the right grayscale image of the ice-covered line. Since the SAD algorithm only measures the degree of difference by the sum of the absolute differences of the pixel values ​​in the two images, it does not consider the feature information of the image and is easily affected by factors such as image noise and texture, resulting in a high possibility of mismatching, which makes the accuracy of the obtained disparity map low.

[0145] In order to solve the technical problem that the SAD algorithm is easily affected by image noise and texture, resulting in errors in image matching and low accuracy of the disparity map, the purpose of this application is to provide a monitoring method with line ice monitoring function, and the technical solution adopted is as follows:

[0146] Step S001: Processing the original image acquired by the monitoring device to obtain a line ice-covered image and a line ice-covered infrared image;

[0147] Step S002: Obtain the melting characteristic value of each pixel in the line ice-covered image according to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image and the difference between the pixel value of the pixel at the same position in the line ice-covered infrared image and the pixel value of the pixel in the preset window;

[0148] Step S003: obtaining an initial cost weight of each pixel in the line ice-covered image according to the degree of confusion of the melting characteristic value in a preset window of each pixel in the line ice-covered image;

[0149] Step S004: Use the SAD matching algorithm and the Census matching algorithm to match the left grayscale image of the line ice coverage with the right grayscale image of the line ice coverage, respectively, obtain the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of the line ice coverage, and combine the initial cost weight to obtain the line ice coverage depth map.

[0150] Step S005: Measure the ice thickness of the transmission line based on the line ice depth map.

[0151] The present application provides a monitoring method with line icing monitoring function, which has the following technical effects:

[0152] In the embodiment of the present application, the melting degree of ice in the ice-covered area is analyzed from two aspects: the difference in line fluctuation degree and temperature difference between liquid water and solid ice in the ice-covered area. The position distribution of edge pixels in a preset window of pixels in the ice-covered line image presents the line fluctuation degree around the pixels. The difference in pixel values ​​between pixels and surrounding pixels in the infrared image of the line ice reflects the temperature difference distribution around the pixels. Comprehensive analysis makes the melting characteristic value present the melting degree of ice more accurate. The distribution of melting characteristic values ​​presents the contour texture of the ice-covered area after melting, so that the initial cost weight is more in line with the ice melting scene. The SAD algorithm is easily affected by factors such as image noise and texture, and the Census algorithm can better make up for these shortcomings. The SAD algorithm and the Census algorithm are combined to match the left grayscale image of the line ice with the right grayscale image of the line ice, so as to improve the accuracy of image matching. Combined with the initial cost weight, the obtained line ice depth map is more in line with the actual situation of the line ice and has a higher accuracy.

[0153] Among them, steps S001-S005 have been described in detail in the above-mentioned embodiment of a monitoring method with a line icing monitoring function, and will not be repeated here.

[0154] On the other hand, the embodiment of the present application provides a monitoring device with a line icing monitoring function, such as Figure 4 As shown, the device 400 includes:

[0155] The data acquisition module 401 is used to process the original image acquired by the monitoring device to obtain a line ice-covered image and a line ice-covered infrared image respectively;

[0156] The image analysis module 402 is used to obtain the melting characteristic value of each pixel in the line ice-covered image according to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image and the difference between the pixel value of the pixel at the same position in the line ice-covered infrared image and the pixel value of the pixel in the preset window;

[0157] The weight calculation module 403 is used to obtain the initial cost weight of each pixel in the line ice-covered image according to the degree of confusion of the melting characteristic value in the preset window of each pixel in the line ice-covered image;

[0158] The image matching module 404 is used to match the left grayscale image of the line ice cover with the right grayscale image of the line ice cover by using the SAD matching algorithm and the Census matching algorithm respectively, obtain the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of the line ice cover, and obtain the line ice cover depth map by combining the initial cost weight;

[0159] The ice thickness monitoring module 405 is used to measure the ice thickness of the transmission line based on the line ice depth map.

[0160] It should be noted that the above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

[0161] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A monitoring method with line icing monitoring function, characterized in that: The method includes: The original image acquired by the monitoring device is processed to obtain a line ice-covered image and a line ice-covered infrared image respectively; According to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image, and the difference between the pixel values ​​of the pixel at the same position in the line ice-covered infrared image and the pixel in the preset window, the melting characteristic value of each pixel in the line ice-covered image is obtained; According to the degree of disorder of the melting characteristic value in the preset window of each pixel in the line ice-covered image, an initial cost weight of each pixel in the line ice-covered image is obtained; The left grayscale image of the line ice cover and the right grayscale image of the line ice cover are matched by using the SAD matching algorithm and the Census matching algorithm respectively, and the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of the line ice cover are obtained, and the line ice cover depth map is obtained by combining the initial cost weights; Measure the ice thickness of transmission lines based on line ice depth map; The method for obtaining the ice depth map comprises: Any pixel point in the left grayscale image of the line ice is used as the left pixel point, and the left pixel point and each pixel point in the right grayscale image of the line ice form a pair to be matched; the maximum value of the initial cost weights of the two pixels in the pair to be matched is used as the matching cost weight when matching the two pixels of the pair to be matched; According to the matching cost weights of the pixels in the left grayscale image of the line ice and the right grayscale image of the line ice, the SAD matching cost and the Census matching cost, the fused disparity cost of each pixel in the left grayscale image of the line ice under the disparity value is obtained; each pixel in the left grayscale image of the line ice has at least two fused disparity costs; The line ice-covered image is filtered to obtain a filtered ice-covered image, wherein the filtered ice-covered image includes: a filtered ice-covered left image and a filtered ice-covered right image; the filtered ice-covered left image is matched with the filtered ice-covered right image to obtain a fusion disparity cost of each pixel point in the filtered ice-covered left image under a disparity value; The disparity value corresponding to the minimum value of the fusion disparity cost of the pixel point at the same position in the left grayscale image of the line ice cover and the left grayscale image of the filtered ice cover is used as the final disparity value of the pixel point in the left grayscale image of the line ice cover; The final disparity value of each pixel in the left grayscale image of the line ice coverage constitutes an initial disparity map; the initial disparity map is optimized to obtain an optimized disparity map; and the optimized disparity map is converted into a line ice coverage depth map using a conversion formula between the disparity map and the depth map.

2. A monitoring method with line icing monitoring function according to claim 1, characterized in that: The method for obtaining the melting characteristic value includes: The monitoring device is fixed on the high-voltage line tower, and the original image of the ice-covered area on the transmission line is collected by using the built-in binocular camera of the monitoring device to obtain the line ice-covered RGB left image and line ice-covered RGB right image; Grayscale the left RGB image of ice on the line and the right RGB image of ice on the line respectively to obtain the left grayscale image of ice on the line and the right grayscale image of ice on the line, and use a thermal imaging camera to obtain the left infrared image of ice on the line and the right infrared image of ice on the line corresponding to the left grayscale image of ice on the line and the right grayscale image of ice on the line respectively; The line ice-covered image includes a line ice-covered left grayscale image and a line ice-covered right grayscale image; The line ice-covered infrared image includes a line ice-covered left infrared image and a line ice-covered right infrared image; According to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image, the line fluctuation degree of each pixel in the line ice-covered image is obtained; According to the difference between the pixel values ​​of each pixel in the infrared image of the ice-covered line and the pixel values ​​in the preset window, the temperature difference characteristic value of each pixel in the infrared image of the ice-covered line is obtained; The melting characteristic value of each pixel in the line ice-covered image is obtained by combining the line fluctuation degree of each pixel in the line ice-covered image and the temperature difference characteristic value of the pixel at the same position in the line ice-covered infrared image; the temperature difference characteristic value and the line fluctuation degree are both positively correlated with the melting characteristic value.

3. A monitoring method with line icing monitoring function according to claim 2, characterized in that: The method for obtaining the line fluctuation degree comprises: According to the number of edge pixels in a preset window of each pixel in the line ice-covered image, edge feature pixels in the line ice-covered image are screened out; For any edge feature pixel point in the line ice-covered image, curve fitting is performed on all edge pixel points in a preset window of the edge feature pixel point to obtain an edge curve segment of the edge feature pixel point, and the absolute value of the difference between the horizontal coordinates of the pixel coordinates of the two end points of the edge curve segment is used as the edge horizontal difference of the edge feature pixel point; the absolute value of the difference between the vertical coordinates of the pixel coordinates of the two end points of the edge curve segment is used as the edge vertical difference of the edge feature pixel point; The line fluctuation degree of the edge feature pixel points in the line ice-covered image is obtained by combining the edge transverse difference value, the edge longitudinal difference value and the length of the preset window of the edge feature pixel points; The line fluctuation of pixels other than edge feature pixels in the line ice-covered image is set to 0.

4. A monitoring method with line icing monitoring function according to claim 3, characterized in that: The method for acquiring edge feature pixel points comprises: Perform edge detection on the line ice-covered image to obtain edge pixel points in the line ice-covered image; The pixel points whose ratio of the number of edge pixels in a preset window of each pixel point in the line ice-covered image to the total number of pixels in the preset window is greater than a preset edge threshold are taken as edge feature pixels in the line ice-covered image.

5. The monitoring method with line icing monitoring function according to claim 3 is characterized in that: The calculation formula of the line fluctuation degree of the edge feature pixel points in the line ice-covered image is as follows: In the formula, is the line fluctuation degree of the i-th edge feature pixel point in the line ice-covered image; is the edge lateral difference of the i-th edge feature pixel point in the line ice-covered image; is the edge longitudinal difference value of the ith edge feature pixel point in the line ice image; L is the length of the preset window; arcsin() is the inverse sine function; cos() is the cosine function.

6. A monitoring method with line icing monitoring function according to claim 2, characterized in that: The method for obtaining the temperature difference characteristic value comprises: Obtain the ambient temperature at the time when the infrared image of the line ice coverage is obtained; Select any pixel point in the infrared image of ice covering the line as the analysis pixel point, and use the difference between the pixel value of the analysis pixel point and the pixel value of each pixel point in its preset window as the temperature difference value of each pixel point in the preset window of the analysis pixel point; When the pixel value of the analyzed pixel is greater than a preset first value, or when the pixel value of the analyzed pixel is equal to the preset first value and the ambient temperature is greater than a preset second value, the label value of the pixel whose temperature difference value is less than the preset temperature threshold within the preset window of the analyzed pixel is set to the first label value; The ratio of the number of pixels whose label values ​​are the first label value in the preset window of the analysis pixel to the total number of pixels in the preset window is used as the temperature difference characteristic value of the analysis pixel.

7. The monitoring method with line icing monitoring function according to claim 1 is characterized in that: The method for obtaining the initial cost weight includes: The information entropy of the melting characteristic value of each pixel in the preset window of the line ice-covered image is obtained as the temperature difference disorder degree of each pixel in the line ice-covered image; According to the melting characteristic value and the temperature difference chaos of each pixel in the line ice-covered image, the initial cost weight of each pixel in the line ice-covered image is obtained; the melting characteristic value and the temperature difference chaos are both positively correlated with the initial cost weight.

8. The monitoring method with line icing monitoring function according to claim 1 is characterized in that: The method of obtaining the fusion disparity cost of each pixel in the left grayscale image of the line ice coverage under the disparity value according to the matching cost weights of the pixels in the left grayscale image of the line ice coverage and the right grayscale image of the line ice coverage, the SAD matching cost and the Census matching cost, includes: The calculation formula of the fusion disparity cost is as follows: In the formula, is the fusion disparity cost of the j-th pixel in the left grayscale image of the line ice; d is the disparity value of the j-th pixel in the left grayscale image of the line ice; is the matching cost weight of the jth pixel in the left grayscale image of the line ice cover; is the SAD matching cost of the jth pixel in the left grayscale image of the line ice cover; is the Census matching cost of the jth pixel in the left grayscale image of the line ice cover; is a positive constant; is a normal number; exp() is an exponential function with the natural constant e as the base.

9. A monitoring device with line icing monitoring function, characterized in that: include: The data acquisition module is used to process the original image acquired by the monitoring device to obtain the line ice-covered image and the line ice-covered infrared image respectively; An image analysis module is used to obtain a melting characteristic value of each pixel in the line ice-covered image according to the position distribution of edge pixels in a preset window of each pixel in the line ice-covered image and the difference between the pixel values ​​of the pixel at the same position in the line ice-covered infrared image and the pixel in the preset window; A weight calculation module, used for obtaining an initial cost weight of each pixel in the line ice-covered image according to the degree of disorder of the melting characteristic value in a preset window of each pixel in the line ice-covered image; An image matching module is used to match the left grayscale image of line ice coverage with the right grayscale image of line ice coverage using the SAD matching algorithm and the Census matching algorithm respectively, obtain the SAD matching cost and the Census matching cost of each pixel in the left grayscale image of line ice coverage, and obtain the line ice coverage depth map in combination with the initial cost weight; An ice thickness monitoring module is used to measure the ice thickness of the transmission line based on the line ice depth map; The method for obtaining the ice depth map comprises: Any pixel point in the left grayscale image of the line ice is used as the left pixel point, and the left pixel point and each pixel point in the right grayscale image of the line ice form a pair to be matched; the maximum value of the initial cost weights of the two pixels in the pair to be matched is used as the matching cost weight when matching the two pixels of the pair to be matched; According to the matching cost weights of the pixels in the left grayscale image of the line ice and the right grayscale image of the line ice, the SAD matching cost and the Census matching cost, the fused disparity cost of each pixel in the left grayscale image of the line ice under the disparity value is obtained; each pixel in the left grayscale image of the line ice has at least two fused disparity costs; The line ice-covered image is filtered to obtain a filtered ice-covered image, wherein the filtered ice-covered image includes: a filtered ice-covered left image and a filtered ice-covered right image; the filtered ice-covered left image is matched with the filtered ice-covered right image to obtain a fusion disparity cost of each pixel point in the filtered ice-covered left image under a disparity value; The disparity value corresponding to the minimum value of the fusion disparity cost of the pixel point at the same position in the left grayscale image of the line ice cover and the left grayscale image of the filtered ice cover is used as the final disparity value of the pixel point in the left grayscale image of the line ice cover; The final disparity value of each pixel in the left grayscale image of the line ice coverage constitutes an initial disparity map; the initial disparity map is optimized to obtain an optimized disparity map; and the optimized disparity map is converted into a line ice coverage depth map using a conversion formula between the disparity map and the depth map.

Images