A power distribution cabinet operation abnormality monitoring system

Abstract

The present application relates to the technical field of electronic power, in particular to a power distribution cabinet operation abnormality monitoring system.The system comprises a data acquisition module for acquiring a power distribution cabinet infrared image; a first data processing module for determining a heat source window and acquiring the change consistency of each parallel line of the optimal straight line in each heat source window; a second data processing module for determining a mutation window and determining a mutation degree; a region division module for correcting an initial corner point threshold value in combination with the mutation degree and the change consistency to obtain a corrected corner point threshold value; performing corner point detection on the power distribution cabinet infrared image to obtain a corner point, and obtaining a monitoring region according to the corner point; and an early warning module for feeding back an early warning signal based on the monitoring region of the power distribution cabinet infrared image.The present application improves the monitoring accuracy of the power distribution cabinet and realizes monitoring and early warning of the power distribution cabinet.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electronic power, in particular to a power distribution cabinet operation abnormality monitoring system. BACKGROUND

[0002] The power distribution cabinet of a transformer area is the core of power utilization of a region, and when the power distribution cabinet is abnormal, it will affect the users of a region, and needs to be monitored safely, and the temperature information of the power distribution box is analyzed to monitor whether there is a fire hazard in the power distribution cabinet. The common method for monitoring the power distribution cabinet at present is to collect the infrared image of the power distribution cabinet, and further analyze the infrared image to detect the temperature of the power distribution cabinet. Compared with the temperature monitoring by the sensor, the infrared image has the advantages of non-contact temperature measurement, real-time monitoring, high accuracy, wide monitoring range and good application in complex situations. However, the imaging principle of the infrared image will result in low image resolution, which will result in that the details are not obvious when the regional temperature of the infrared image is analyzed due to the low resolution, and the accuracy of monitoring the power distribution cabinet is low. SUMMARY

[0003] In order to solve the technical problem that the accuracy is low when the corner points are obtained by the corner point detection algorithm to segment the infrared image of the power distribution cabinet, resulting in low accuracy of monitoring the power distribution cabinet, the purpose of the present application is to provide a power distribution cabinet operation abnormality monitoring system, and the technical scheme adopted is as follows:

[0004] A data acquisition module is used to acquire the infrared image of the power distribution cabinet.

[0005] A first data processing module is used to determine the heat source window according to the gray scale distribution of the infrared image of the power distribution cabinet, draw a straight line in different directions through the center pixel point of each heat source window, screen the straight lines in different directions to obtain the optimal straight line of each heat source window, and obtain the consistency of change of each parallel line according to the consistency of gray scale between the parallel lines parallel to the optimal straight line.

[0006] A second data processing module is used to obtain the temperature change trend of the optimal straight line by fusing the gray scale values of the pixel points in the optimal straight line and the corresponding parallel line based on the consistency of change, determine the mutation window according to the temperature change trend and the consistency of change, and determine the mutation degree according to the gray scale distribution in the mutation window.

[0007] A region division module is used to correct the initial corner point threshold value in combination with the mutation degree and the consistency of change to obtain a corrected corner point threshold value, detect the corner points of the infrared image of the power distribution cabinet according to the corrected corner point threshold value, divide the infrared image of the power distribution cabinet into regions according to the corner points, and obtain a monitoring region.

[0008] The early warning module is configured to feed back an early warning signal based on a monitoring area of the power distribution cabinet infrared image.

[0009] Preferably, the straight lines in different directions are screened to obtain the optimal straight line of each heat source window, including:

[0010] The pixel points in the heat source window are taken as heat source pixel points.

[0011] For any heat source window, the number of heat source pixel points on the straight line in different directions passing through the center pixel point of the heat source window is obtained as a judgment number, and the straight line with the largest judgment number is taken as the optimal straight line of the heat source window.

[0012] Preferably, the change consistency of each parallel line is obtained according to the gray scale consistency between parallel lines parallel to the optimal straight line, including:

[0013] For each parallel line of the optimal straight line, a parallel line sequence is constructed by using the gray scale values of the pixel points on the parallel line.

[0014] For the parallel line sequences corresponding to any two parallel lines corresponding to the optimal straight line, the similarity in the two parallel line sequences is calculated.

[0015] For any parallel line corresponding to the optimal straight line, the average of the similarities of the parallel line and other parallel lines corresponding to the optimal straight line is obtained as a similarity average, the minimum value of the similarities of the parallel line and other parallel lines corresponding to the optimal straight line is taken as a minimum similarity, and the normalized value of the product of the similarity average and the minimum similarity is taken as the change consistency of the parallel line.

[0016] Preferably, the temperature change trend of the optimal straight line is obtained by fusing the gray scale values of the pixel points in the optimal straight line and the corresponding parallel lines based on the change consistency, including:

[0017] The parallel line with the largest change consistency among the parallel lines corresponding to the optimal straight line is taken as a main line, the change consistency is taken as a weight, the gray scale values of the pixel points on all the parallel lines corresponding to the optimal straight line are weighted and fused to obtain a fused pixel value, and the fused pixel value is assigned to each pixel point on the main line; a sequence composed of the pixel points on the main line is taken as the temperature change trend of the optimal straight line.

[0018] Preferably, the mutation window is determined according to the temperature change trend and the change consistency, including:

[0019] When the change consistency of the main line is less than a preset consistency threshold, the fused pixel values of the pixel points on the main line are curve fitted to obtain a fused fitting curve.

[0020] The average of the slopes of the points on the fusion fitting curve is calculated as a slope average, and when the slopes of a preset number of continuous points on the fusion fitting curve are greater than the slope average, the heat source window corresponding to the fusion fitting curve is taken as a prominent window, wherein the preset number is a positive integer.

[0021] Preferably, the determining of the mutation degree according to the gray scale distribution in the mutation window comprises:

[0022] The gray scale values of the pixel points in the mutation window are fitted to obtain a window fitting curve, and a normalized value of the variance of the slopes of the points corresponding to the gray scale values of all the pixel points on the window fitting curve is taken as the mutation degree.

[0023] Preferably, the correcting of the initial corner point threshold value in combination with the mutation degree and the change consistency to obtain a corrected corner point threshold value comprises:

[0024] When the heat source window is a non-mutation window, the product of the initial corner point threshold value and the change consistency of the optimal straight line of the heat source window is taken as the corrected corner point threshold value of the optimal straight line of the heat source window.

[0025] When the heat source window is a mutation window, a difference between a preset threshold value and the mutation degree is taken as an adjustment degree, and the product of the initial corner point threshold value, the change consistency of the optimal straight line of the heat source window and the adjustment degree of the heat source window is taken as the corrected corner point threshold value of the optimal straight line of the heat source window, wherein the preset threshold value is greater than the mutation degree in value.

[0026] Preferably, the determining of the heat source window according to the gray scale distribution of the power distribution cabinet infrared image comprises:

[0027] The gray scale mean values of the pixel points in the sliding window are calculated, and the area corresponding to the sliding window in which the gray scale mean value is greater than the gray scale mean value of the power distribution cabinet infrared image is taken as a high-heat area.

[0028] The heat source window is determined according to the gray scale distribution in the high-heat area.

[0029] Preferably, the determining of the heat source window according to the gray scale distribution in the high-heat area comprises:

[0030] For any high-heat area, the difference between the gray scale mean value in the high-heat area and the gray scale mean value in all the high-heat areas is calculated as the gray scale mean difference of the high-heat area.

[0031] The variance of the gray scale values of the pixel points in the high-heat area is inversely proportional normalized to obtain an initial probability of the high-heat area.

[0032] The normalized value of the product of the gray scale mean difference and the initial probability of the high-heat area is taken as the heat generation probability of the high-heat area.

[0033] According to the heat generation probability, the heat source window is screened out from the high-heat area.

[0034] Preferably, the monitoring area based on the power distribution cabinet infrared image feeds back a warning signal, comprising:

[0035] The temperature of each monitoring area is acquired, and when the temperature of the monitoring area does not belong to the normal range of the corresponding actual area, the power distribution cabinet is warned.

[0036] The embodiment of the application has at least the following beneficial effects:

[0037] The system comprises a data acquisition module for acquiring a power distribution cabinet infrared image; a first data processing module for determining a heat source window, the heat source of the infrared image being a relatively stable high-heat area, the influence of this area on the surrounding being a diverging area, the closer the distance, the greater the influence degree and the more intense the gray scale change, the heat source window in the image being acquired according to the gray scale information of the image, and subsequent analysis of the heat source window only being required; since the heat source window may be affected by multiple heat sources, leading to a low credibility of the calculated area consistency, when the heat generation area appears an overlapping phenomenon, the consistency calculation of the heat source window is affected, therefore, the application first screens out the optimal straight line corresponding to each heat source window, and the change consistency is obtained according to the gray scale consistency degree between the parallel lines parallel to the optimal straight line, the worse the change consistency, the more the special situation existing in the heat source window, and the more the edge area or the overlapping area; a second data processing module for determining a mutation window; according to the gray scale distribution in the mutation window, the mutation degree is determined, the mutation degree reflecting the intense change degree of the mutation window, and the initial corner point threshold is corrected in combination with the mutation degree; a region division module for correcting the initial corner point threshold in combination with the mutation degree and the change consistency to obtain a corrected corner point threshold, realizing dynamic adjustment of the threshold selection of the corner point detection according to the change consistency and the mutation degree, avoiding the influence of the setting of a unified threshold on the identification of the heat generation area in the power distribution cabinet infrared image due to too many or too few corner points; according to the corrected corner point threshold, corner point detection is performed on the power distribution cabinet infrared image to obtain a corner point, region division is performed on the power distribution cabinet infrared image according to the corner point to obtain a monitoring area; and a warning module for feeding back a warning signal based on the monitoring area of the power distribution cabinet infrared image. The application acquires the change consistency and the mutation degree by analyzing the power distribution cabinet infrared image, and then dynamically adjusts the corner point threshold, improves the acquisition accuracy of the corner point, and further improves the monitoring accuracy of the power distribution cabinet. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, below will briefly introduce the drawings needed in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative effort.

[0039] Figure 1 A system block diagram of an abnormal operation monitoring system of a power distribution cabinet according to an embodiment of the present application;

[0040] Figure 2 A system block diagram of an abnormal operation monitoring area division system of a power distribution cabinet according to an embodiment of the present application. DETAILED DESCRIPTION

[0041] In order to further illustrate the technical means and effects adopted by the present application to achieve the predetermined object, the specific implementation, structure, features and effects of the abnormal operation monitoring system of a power distribution cabinet according to the present application are described in detail below in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0043] An embodiment of an abnormal operation monitoring system of a power distribution cabinet:

[0044] The embodiment of the present application provides a specific implementation method of an abnormal operation monitoring system of a power distribution cabinet, which is suitable for a power distribution cabinet monitoring scene. In this scene, an infrared image of the power distribution cabinet is obtained by an infrared camera, and subsequent analysis is performed on the infrared image of the power distribution cabinet. In order to solve the technical problem that the accuracy is low when the infrared image is segmented by a corner point detection algorithm, resulting in low accuracy of monitoring the power distribution cabinet. The present application distinguishes the heat source window of the power distribution cabinet by the gray value, calculates the consistency of change of the parallel lines corresponding to the optimal straight line of each window and the mutation degree of the mutation window according to the trend of temperature change in the heat source window, and dynamically adjusts the threshold selection of the corner point detection according to the consistency of change and the mutation degree. Avoid setting a uniform threshold to cause too many or too few corner points in the infrared image detection, which affects the recognition of the heat area in the image, and improve the monitoring accuracy of the power distribution cabinet.

[0045] The specific scheme of the abnormal operation monitoring system of a power distribution cabinet according to the present application is described in detail below in combination with the drawings.

[0046] Please refer toFigure 1 Figure 1 shows a system block diagram of an abnormal operation monitoring system for a power distribution cabinet according to an embodiment of the present application, which comprises the following modules:

[0047] A data acquisition module 10 is configured to acquire an infrared image of the power distribution cabinet.

[0048] The infrared camera is used to capture an infrared video of the working power distribution cabinet, and a frame of infrared image is intercepted from the infrared video as a to-be-processed image, which is denoted as the infrared image of the power distribution cabinet. It should be noted that the infrared image of the power distribution cabinet is a grayscale image.

[0049] A first data processing module 20 is configured to determine a heat source window according to the grayscale distribution of the infrared image of the power distribution cabinet, draw a straight line in different directions through the center pixel point of each heat source window, screen the straight lines in different directions to obtain an optimal straight line of each heat source window, and obtain the consistency of change of each parallel line according to the consistency of the grayscale between the parallel lines parallel to the optimal straight line.

[0050] The resolution of the infrared image of the low-voltage transformer distribution box is low, which may result in unclear resolution of details in the infrared image. In addition, the extraction of the temperature information feature points in the infrared image is inaccurate. The temperature in different regions in the infrared image may divide the same object into two different grayscale information, or the influence of the heat source on the surrounding environment, such as the phenomenon of halo, may result in problems in the heat source region detected by the corner point. The present application needs the heat source region rather than the region affected by the heat source.

[0051] A sliding window is slid on the infrared image of the power distribution cabinet, and the size of the sliding window is 3*3 in the embodiment of the present application. The sliding window regions in the infrared image of the power distribution cabinet are screened according to the similarity and connectivity of the grayscale distribution in the sliding window.

[0052] The similarity of the grayscale between adjacent sliding window regions is analyzed. The heat source in the infrared image is a relatively stable region, which can be denoted as the heat source region. The influence of the heat source region on the surrounding environment is a divergent region. The closer to the heat source region, the greater the degree of influence and the more intense the change in grayscale. In an ideal condition, the windows at the same distance from the heat source are affected by the heat source in the same way.

[0053] The heat source region is determined according to the region with stable grayscale change and the grayscale value in the region in the infrared image of the power distribution cabinet. According to the grayscale information of the infrared image of the power distribution cabinet, the high-heat region in the infrared image of the power distribution cabinet is acquired first. The high-heat region is a region with a large grayscale value. After the high-heat region is obtained, the heat source region is obtained by analyzing the high-heat region. The subsequent analysis of the influence of the temperature change of the power distribution cabinet only needs to analyze the heat source region in the infrared image.

[0054] According to the mean value of the gray scale in the sliding window and the mean value of the gray scale of the whole power distribution cabinet infrared image, the heat source region in the infrared image is obtained, that is, according to the gray scale distribution of the power distribution cabinet infrared image, the heat source window is determined, specifically:

[0055] On the power distribution cabinet infrared image, the sliding window is slid, and the region corresponding to the sliding window in which the mean value of the gray scale of the pixel points in the sliding window is greater than the mean value of the gray scale of the power distribution cabinet infrared image is taken as the high-heat region. The mean value of the gray scale of the pixel points in the sliding window is denoted as the mean value of the gray scale values of the pixel points in the sliding window, and the mean value of the gray scale of the power distribution cabinet infrared image is the mean value of the gray scale values of all pixel points in the power distribution cabinet infrared image.

[0056] Further, according to the gray scale distribution in the high-heat region, the heat source window is determined, specifically: for any high-heat region, the difference between the mean value of the gray scale in the high-heat region and the mean value of the gray scale in all high-heat regions is calculated as the gray scale mean difference of the high-heat region. The variance of the gray scale values of the pixel points in the high-heat region is inversely proportional to the normalization to obtain the initial probability of the high-heat region. The normalized value of the product of the gray scale mean difference and the initial probability of the high-heat region is taken as the heat probability of the high-heat region. According to the heat probability, the heat source window is selected from the high-heat region, that is, the high-heat region with a heat probability greater than a preset probability threshold is taken as the heat source window. In the embodiment of the present application, the value of the preset probability threshold is 0.78, and in other embodiments, it can be set by the implementer according to the actual situation.

[0057] The calculation formula of the heat probability is:

[0058]

[0059] wherein, is the heat probability of the i-th high-heat region; is the mean value of the gray scale of the i-th high-heat region; is the mean value of the gray scale of all high-heat regions; is the variance of the gray scale values of the pixel points in the i-th high-heat region; is the initial probability of the i-th high-heat region; is the gray scale mean difference of the i-th high-heat region; and norm is a normalization function.

[0060] The heat source region should be the region with the largest gray scale value in the power distribution cabinet infrared image, and the region affected by the heat source is relatively stable. The mean value of the gray scale of the window in the high-heat region and the variance of the gray scale in the window are calculated, and the probability that the high-heat region is the heat source region is obtained according to the mean value and the variance. The greater the mean value and the smaller the variance, the higher the probability that the window is the heat source.

[0061] wherein, the gray scale mean difference The high heat distribution of the i-th high heat area is reflected, the greater the gray level difference, the more the high heat area presents the high heat distribution. The variance of the gray level value of the pixel point in the high heat area reflects the intensity of the gray level change in the high heat area, the more intense the gray level change in the heat area, the smaller the probability that the high heat area is a heat source. In the embodiment of the application, the exponential function with the natural constant as the base and the variance of the gray level value of the pixel point in the negative high heat area as the index is used , to realize the inverse proportional normalization mapping of the variance of the gray level value of the pixel point in the high heat area. In the embodiment of the application, the normalization function adopts a linear normalization function, and in other embodiments, the selection of the normalization function can be adjusted by the implementer according to the actual situation.

[0062] For example, the linear normalization function in the embodiment of the application is specifically a minimum-maximum normalization.

[0063] Therefore, according to the heat generation probability, a plurality of heat source windows are screened out from the high heat area, and specifically, the high heat area with a heat generation probability greater than a preset probability threshold is taken as a heat source window.

[0064] Further, feature extraction is performed on the heat source window in the infrared image. The corner points detected in the affected area are not needed, but the gray level values in the overlapping area of the two heat sources are similar, and the two heat sources are obviously distinguished but the overlapping area should have obvious edges, but the effect of corner point detection in this area is not obvious.

[0065] According to prior knowledge, it is known that the propagation efficiency of temperature under different media is different. The influence of the heat source area on the surrounding high heat area is analyzed, and the consistency of the heat source and the high heat area is quantified according to the different influences of the heat sources under the same distance. Since the high heat area may be affected by multiple heat sources, the reliability of the calculated area consistency is low. When the heat source areas overlap, the temperature of the two heat sources is similar, resulting in a high area consistency of the two objects or an inability to calculate the area consistency. However, there is a slight gray difference in the overlapping area of the heat sources. According to the gray difference, the image can be divided into physical overlap and visual overlap. The gray value of the visual overlap area is slightly lower than that of the heat source area, and the gray value of the physical overlap area is slightly greater than that of the heat source area.

[0066] In the calculation of the obvious heat source edge in the overlapping area, the extension and gray level change consistency of the non-overlapping area can be used to distinguish the edge information of the overlapping area. However, there are multiple heat sources overlapping in the overlapping area. According to the extension of the non-overlapping heat source edge, the overlapping area is divided. Since the gray level change in the overlapping area is not obvious, the extension of the edge in many directions is good, which leads to inaccurate division of the overlapping area and affects the subsequent division of the heat source area.

[0067] Passing through the center pixel point of each heat source window, straight lines in different directions are made, and the straight lines in different directions are screened to obtain the optimal straight line of each heat source window. Specifically, straight lines are made in different directions passing through the center pixel point of each heat source window. In the embodiment of the present application, the range of the angle of different directions is [0, 180), and the angle is an integer. That is, an arbitrary window in the heat source window is selected, a straight line is made in any direction passing through the center pixel point in the window.

[0068] The pixel points in the heat source window are taken as heat source pixel points; for any heat source window, the number of heat source pixel points on the straight line in different directions passing through the center pixel point of the heat source window is obtained as the judgment number, and the straight line with the largest judgment number is taken as the optimal straight line of the heat source window. By analyzing the change of the gray value on the straight line, because there is a certain difference between the boundary region and the internal region, the difference is amplified according to the overall change of the subtle difference, so that the accurate overlapping edge can be better obtained, and therefore the optimal straight line is screened out.

[0069] A perpendicular line is made passing through the optimal straight line, and a parallel line parallel to the optimal straight line is made passing through the points on the perpendicular line in turn to obtain a plurality of parallel lines corresponding to the optimal straight line. Further, the gray value change of the pixel points on the parallel line is analyzed.

[0070] According to the consistency degree of the gray values between the parallel lines parallel to the optimal straight line, the change consistency of each parallel line is obtained. Specifically:

[0071] For each parallel line of the optimal straight line, a parallel line sequence is constructed by the gray values of the pixel points on the parallel line; due to the limitation of the power distribution cabinet infrared image, the number of pixel points on each parallel line may not be the same, so it cannot be calculated by one-to-one correspondence of the pixel points at the same position. Therefore, for the parallel line sequences corresponding to any two parallel lines corresponding to the optimal straight line, the similarity between the two parallel line sequences is calculated, specifically: the similarity between the two parallel line sequences is calculated by using a dynamic time warping algorithm (Dynamic Time Warping, DTW). The similarity between two sequences of different orders of magnitude is calculated by the DTW algorithm. The element values in the parallel line sequences corresponding to the two parallel lines are matched, and there may be one-to-many cases in the element values in the sequence, but the matching between the element values has a sequence, and a curve p is established according to the Euclidean distance of the pixel points corresponding to the element values in the two parallel line sequences and the order of the element pairs. The element pair is composed of two element values that match successfully. The change stability of the curve p is analyzed, and the more intense the change of the curve p, the smaller the similarity between the corresponding two parallel line sequences. The ordinate of the curve p is the Euclidean distance between the element pairs, and the abscissa is the order value of the element pairs. The order values of the element pairs are arranged in sequence according to the left-to-right order of the parallel line sequences, and when there are multiple element values corresponding to an element value, the element pairs are sorted according to the left-to-right order of the corresponding multiple element values. The change consistency of the multiple parallel lines corresponding to the optimal straight line of each heat source window.

[0072] The calculation formula of the similarity between the two parallel line sequences is:

[0073]

[0074] Where D is the similarity between the two parallel line sequences; exp is an exponential function with a natural constant as the base number; is the variance of the slope of all element pairs corresponding to the two parallel line sequences.

[0075] The variance is greater, the more intense the fluctuation of the curve, and the smaller the similarity between the two parallel line sequences, so here the is realized to realize the negative correlation normalization mapping of the variance , so that the variance

[0076] According to the calculated similarity between the two parallel line sequences, the similarity of the temperature change on the parallel line corresponding to the parallel line sequence is reflected, and according to the calculated similarity, the consistency of the parallel line and other parallel lines is obtained. The worse the consistency, the greater the difference between the temperature change curve on the parallel line and other parallel lines, and according to the characteristics of the corner point, the greater the difference between the parallel lines, the greater the probability of the existence of the corner point on the parallel line.

[0077] Furthermore, for any parallel line corresponding to the optimal line, the mean similarity between the parallel line and other parallel lines corresponding to the optimal line is obtained as the mean similarity; the minimum similarity between the parallel line and other parallel lines corresponding to the optimal line is taken as the minimum similarity; and the normalized value of the product of the mean similarity and the minimum similarity is taken as the consistency of the change of the parallel line.

[0078] The formula for calculating the consistency of this change is:

[0079]

[0080] in, To ensure consistency in the changes of the i-th parallel line; This is the normalization function; It is a minimum value function; Let be the minimum similarity corresponding to the i-th parallel line; Let be the mean similarity value corresponding to the i-th parallel line.

[0081] This consistency reflects the consistency of grayscale changes among the multiple parallel lines corresponding to the optimal line. The more consistent the changes among the multiple parallel lines, the greater the consistency, indicating a lower probability of special cases existing in the heat source window corresponding to the optimal line.

[0082] When determining the consistency of changes, the similarity between the parallel line to be detected and other parallel lines is calculated to obtain the consistency of temperature changes on the parallel line to be detected and other parallel lines. When the similarity detected on the parallel line to be detected and other parallel lines is low, it reflects that the consistency of the parallel line to be detected is poor. If only one or two lines have low similarity, it may be due to the poor consistency of other line segments.

[0083] Indicates the first Gray-scale changes on parallel lines and removal of the first line The consistency of grayscale variation among all parallel lines; Indicates the first Gray-scale changes on parallel lines and removal of the first line The minimum similarity between all parallel lines is used to determine the consistency. If the minimum similarity is consistently high, it can reflect the consistency of the line being detected. The parallel lines are relatively consistent with the other parallel lines. Let be the average of the grayscale changes on the i-th parallel line and the similarity between it and all other parallel lines except the i-th parallel line. The larger this average value, the higher the similarity between the other parallel lines and the i-th parallel line, and the higher the corresponding consistency. Minimum similarity. By mean similarity To correct prevent the occurrence of the first The similarity of the first The consistency of the first

[0084] According to the variation consistency obtained above, the probability of the existence of a corner point on the parallel line can be determined, and the worse the variation consistency is, the more special conditions exist on the straight line, which can be edge regions or overlapping regions, and when the corner point detection is performed on these special conditions, the threshold of the corner point detection needs to be adjusted.

[0085] The second data processing module 30 is configured to obtain a temperature variation trend of the optimal straight line by fusing the gray values of the pixels in the optimal straight line and the corresponding parallel lines based on the variation consistency, determine a mutation window according to the temperature variation trend and the variation consistency, and determine a mutation degree according to the gray distribution in the mutation window.

[0086] Further, the temperature variation section on the parallel line causing the poor variation consistency is analyzed, the parallel lines are fused by taking the variation consistency as a weight, the temperature variation trend in the direction corresponding to the parallel lines is obtained, and the gray values on the parallel lines are fused directly. Since the infrared image of the power distribution cabinet is an infrared image, the gray values on the image reflect the temperature data, and therefore the temperature data on the parallel lines is fused by fusing the gray values on the parallel lines.

[0087] Based on the variation consistency, the temperature variation trend of the optimal straight line is obtained by fusing the gray values of the pixels in the optimal straight line and the corresponding parallel lines, specifically: the parallel line with the maximum variation consistency among the parallel lines corresponding to the optimal straight line is taken as a main line; the variation consistency is taken as a weight, the gray values of the pixels on all the parallel lines corresponding to the optimal straight line are fused by weighting, a fused pixel value is obtained, and the fused pixel value is assigned to each pixel on the main line; and a sequence composed of the pixels on the main line is taken as the temperature variation trend of the optimal straight line. In the embodiment of the application, the consistency is taken as a weight, the other parallel lines except the main line are taken as auxiliary lines, and the auxiliary lines are fused by polynomial interpolation to the data of the main line. According to the weighted fusion between different parallel lines, the temperature variation value of the main line is obtained.

[0088] According to the temperature change trend, a main line with low change consistency and a region with a relatively intense change trend are obtained. When the change consistency of the main line is less than a preset consistency threshold, a fusion pixel value of each pixel point on the main line is subjected to curve fitting to obtain a fusion fitting curve, wherein the preset consistency threshold has a value range of (0, 1). In the embodiment of the present application, the value of the preset consistency threshold is 0.3, and in other embodiments, the value can be adjusted by the implementer according to the actual situation. A mean value of the slope of each point on the fusion fitting curve is calculated as a slope mean value. When the slope of a preset number of consecutive points on the fusion fitting curve is greater than the slope mean value, a heat source window corresponding to the fusion fitting curve is taken as a prominent window, wherein the preset number is a positive integer. In the embodiment of the present application, the value of the preset number is 3, and in other embodiments, the value can be adjusted by the implementer according to the actual situation. It should be noted that each fusion fitting curve has a corresponding unique main line, each main line has a corresponding unique optimal straight line, and each optimal straight line has a corresponding unique heat source window, so the fusion fitting curve has a corresponding heat source window.

[0089] For the mutation window, the threshold for detecting the corner point obtained according to the temperature change consistency of the straight line needs to be adjusted according to the degree of change, so the degree of mutation of the mutation window is further calculated, which is also the degree of change. Each mutation window corresponds to a degree of mutation.

[0090] According to the gray scale distribution in the mutation window, the degree of mutation is determined, specifically: the gray scale values of the pixel points in the mutation window are fitted to obtain a window fitting curve; the normalized value of the variance of the slope of the points corresponding to the gray scale values of all the pixel points on the window fitting curve is taken as the degree of mutation.

[0091] The calculation formula of the degree of mutation is:

[0092]

[0093] wherein, is the degree of mutation of the kth mutation window; norm is a normalization function; is the variance of the slope of the points corresponding to the gray scale values of all the pixel points on the window fitting curve corresponding to the mutation window.

[0094] wherein, the variance of the slope of the points corresponding to the gray scale values of all the pixel points on the window fitting curve corresponding to the mutation window The greater the value is, the more intense the change of the mutation window is.

[0095] The region division module 40 is configured to correct the initial corner point threshold value in combination with the mutation degree and the change consistency, to obtain a corrected corner point threshold value; perform corner point detection on the power distribution cabinet infrared image according to the corrected corner point threshold value, to obtain a corner point; and perform region division on the power distribution cabinet infrared image according to the corner point, to obtain a monitoring region.

[0096] The mutation degree and the change consistency both reflect the change degree of a region, and therefore the initial corner point threshold value is further corrected in combination with the mutation degree and the change consistency, specifically as follows.

[0097] When the heat source window is a non-mutation window, the product of the initial corner point threshold value and the change consistency of the optimal straight line of the heat source window is taken as the corrected corner point threshold value of the optimal straight line of the heat source window, wherein the initial corner point threshold value ranges from 0 to 1. In the embodiment of the present application, the initial corner point threshold value is 0.1, and in other embodiments, the value can be adjusted by the implementer according to the actual situation.

[0098] When the heat source window is a mutation window, the initial corner point threshold value is corrected according to the mutation degree and the change consistency, to obtain the corrected corner point threshold value of the optimal straight line of the heat source window. It should be noted that the non-optimal straight line refers to the straight line other than the optimal straight line.

[0099] When the heat source window is a mutation window, the method for obtaining the corrected corner point threshold value of the optimal straight line of the heat source window is as follows: taking the difference between a preset threshold value and the mutation degree as an adjustment degree; and taking the product of the initial corner point threshold value, the change consistency of the optimal straight line of the heat source window and the adjustment degree of the heat source window as the corrected corner point threshold value of the optimal straight line of the heat source window, wherein the preset threshold value is greater than the mutation degree. In the embodiment of the present application, the preset threshold value is 1, and in other embodiments, the value can be adjusted by the implementer according to the actual situation.

[0100] When the heat source window is a non-mutation window, the calculation formula of the corrected corner point threshold value of the optimal straight line of the heat source window is as follows:

[0101]

[0102] wherein, is the corrected corner point threshold value; is the initial corner point threshold value; is the change consistency.

[0103] When the heat source window is a non-mutation window, the change consistency is taken as a weight to adjust the corrected corner point threshold value in the calculation formula of the corrected corner point threshold value of the optimal straight line of the heat source window, and the change consistency and the corrected corner point threshold value are in a positive correlation.

[0104] When the heat source window is a mutation window, further adjustment of the correction corner point threshold is needed in combination with the mutation degree of the mutation window, and the calculation formula of the correction corner point threshold of the optimal straight line of the heat source window is as follows:

[0105]

[0106] wherein, is the correction corner point threshold; is the initial corner point threshold; is the change consistency; is the mutation degree; is the adjustment degree.

[0107] When the straight line is a non-optimal straight line, the calculation formula of the correction corner point threshold of the non-optimal straight line not only takes the change consistency as the weight of the correction corner point threshold, but also negatively correlates the mutation degree to obtain the corresponding adjustment degree. The change consistency and the adjustment degree are combined to adjust the correction corner point threshold, wherein the change consistency and the adjustment degree are positively correlated with the correction corner point threshold.

[0108] According to the correction corner point threshold, the corner point of the power distribution cabinet infrared image is detected to obtain the corner point. It should be noted that the corner point is obtained by the corner point detection based on the corner point threshold, which is a known technology to those skilled in the art and will not be described here.

[0109] Further, the power distribution cabinet infrared image is divided into regions according to the obtained corner point to obtain a monitoring region with obvious region distribution. The power distribution cabinet infrared image is divided into regions according to the corner point to obtain a monitoring region, specifically: the edge of the power distribution cabinet infrared image is detected to obtain a plurality of edge lines, and a plurality of closed regions are formed by the edge lines. The closed region on the edge line containing the corner point is taken as the monitoring region on the power distribution cabinet infrared image.

[0110] The warning module 50 is configured to feed back a warning signal based on the monitoring region of the power distribution cabinet infrared image.

[0111] The power distribution cabinet infrared image containing the monitoring region obtained above is input into the trained region judgment neural network to obtain an actual region corresponding to each monitoring region, wherein the actual region includes an internal component region, a connection region and a grounding region. It should be noted that the method of training the neural network is a known technology to those skilled in the art.

[0112] The temperature of each monitoring area is measured to obtain the temperature of each monitoring area, and when the temperature of the monitoring area does not belong to the normal range of the corresponding actual area, the power distribution cabinet is warned, that is, when the temperature of the monitoring area does not belong to the normal range of the corresponding actual area, it is determined that the power distribution cabinet is abnormal, and an abnormal signal is sent to the monitoring personnel for alarm. It should be noted that according to the national standard, the normal range corresponding to different areas of the power distribution cabinet is different, and the normal range of the internal component area is 70-90℃ under normal circumstances, the normal range of the connection area is 40-60℃, and the normal range of the grounding area is 60-80℃.

[0113] In summary, the present application relates to the field of electronic power technology. The system includes the following modules: a data acquisition module for acquiring power distribution cabinet infrared images; a first data processing module for determining heat source windows according to the gray scale distribution of the power distribution cabinet infrared images; passing through the center pixel point of each heat source window, making straight lines in different directions, and selecting different straight lines to obtain the optimal straight line of each heat source window; obtaining the change consistency of each parallel line according to the gray scale consistency between parallel lines parallel to the optimal straight line; a second data processing module for fusing the gray scale values of the pixels in the optimal straight line and the corresponding parallel lines based on the change consistency to obtain the temperature change trend of the optimal straight line; determining the mutation window according to the temperature change trend and the change consistency; determining the mutation degree according to the gray scale distribution in the mutation window; a region division module for correcting the initial corner point threshold value in combination with the mutation degree and the change consistency to obtain a corrected corner point threshold value; performing corner point detection on the power distribution cabinet infrared image according to the corrected corner point threshold value to obtain a corner point, and dividing the power distribution cabinet infrared image into regions according to the corner point to obtain a monitoring area; a warning module for feeding back a warning signal based on the monitoring area of the power distribution cabinet infrared image. The present application distinguishes the heat source windows of the power distribution cabinet by gray scale value, calculates the change consistency of the parallel lines corresponding to the optimal straight line of each window and the mutation degree of the mutation window according to the temperature change trend in the heat source window, and dynamically adjusts the threshold value selection of the corner point detection according to the change consistency and the mutation degree. Avoid setting a uniform threshold value to cause too many or too few corner points in the infrared image detection to affect the recognition of the heat generating area in the image, and improve the monitoring accuracy of the power distribution cabinet.

[0114] A power distribution cabinet operation abnormality monitoring area division system embodiment:

[0115] The existing region temperature analysis on the power distribution cabinet infrared image is to analyze the whole power distribution cabinet infrared image, and the calculation amount is large and the analysis process is complex. Therefore, the corner point detection is usually used to obtain the corner points in the image to realize the segmentation of the image to obtain the monitoring region. Only the monitoring region is analyzed, and the calculation amount can be reduced. However, the details are not obvious due to the low resolution, and the corner points are obtained by the corner point detection algorithm to realize the segmentation of the power distribution cabinet infrared image. The monitoring region in the power distribution cabinet infrared image is not accurate.

[0116] In order to solve the problem that the corner points are obtained by the corner point detection algorithm to realize the segmentation of the power distribution cabinet infrared image, and the monitoring region in the power distribution cabinet infrared image is not accurate, the embodiment provides a power distribution cabinet operation abnormality monitoring region division system. Please refer to Figure 2 which shows a system block diagram of a power distribution cabinet operation abnormality monitoring region division system. The system includes the following modules:

[0117] The data acquisition module 10 is used to acquire the power distribution cabinet infrared image.

[0118] The first data processing module 20 is used to determine the heat source window according to the gray scale distribution of the power distribution cabinet infrared image. The center pixel point of each heat source window is used to draw straight lines in different directions. The straight lines in different directions are screened to obtain the optimal straight line of each heat source window. The gray scale consistency between the parallel lines parallel to the optimal straight line is used to obtain the change consistency of each parallel line.

[0119] The second data processing module 30 is used to fuse the gray scale values of the pixel points in the optimal straight line and the corresponding parallel line based on the change consistency to obtain the temperature change trend of the optimal straight line. The temperature change trend and the change consistency are used to determine the mutation window. The gray scale distribution in the mutation window is used to determine the mutation degree.

[0120] The region division module 40 is used to correct the initial corner point threshold value in combination with the mutation degree and the change consistency to obtain the corrected corner point threshold value. The corner point detection is performed on the power distribution cabinet infrared image according to the corrected corner point threshold value to obtain the corner points. The region division is performed on the power distribution cabinet infrared image according to the corner points to obtain the monitoring region.

[0121] The data acquisition module 10 to the region division module 40 have been described in detail in the above-mentioned power distribution cabinet operation abnormality monitoring system embodiment, and will not be repeated here.

[0122] The power distribution cabinet operation abnormality monitoring area division system relates to the technical field of electronic power, and comprises a data acquisition module, which is used to acquire a power distribution cabinet infrared image; a first data processing module, which is used to determine a heat source window; the heat source of the infrared image is a relatively stable high-heat area, and the area has an influence on the surrounding area, which is a diverging area, and the closer the distance, the greater the influence degree and the more intense the gray scale change; the heat source window in the image is acquired according to the gray scale information of the image, and subsequent analysis of the heat source window is only needed; since the heat source window may be affected by multiple heat sources, the consistency of the calculated area is low, and when the heat area overlaps, the consistency calculation of the heat source window is affected, so the application first screens the optimal straight line corresponding to each heat source window, and the change consistency is obtained according to the gray scale consistency between the parallel lines parallel to the optimal straight line; the worse the change consistency, the more the special situation in the heat source window, which may be an edge area or an overlapping area; a second data processing module is used to determine a mutation window; according to the gray scale distribution in the mutation window, the mutation degree is determined, the mutation degree reflects the degree of change of the mutation window, and the initial screening corner point threshold is corrected in combination with the mutation degree; an area division module is used to correct the initial corner point threshold in combination with the mutation degree and the change consistency to obtain a corrected corner point threshold, so that the threshold selection of the corner point detection is dynamically adjusted according to the change consistency and the mutation degree, the influence of the excessive or insufficient corner points of the power distribution cabinet infrared image detection on the recognition of the heat area in the image is avoided, the corner point of the power distribution cabinet infrared image is detected according to the corrected corner point threshold, the corner point is obtained, the area of the power distribution cabinet infrared image is divided according to the corner point, the monitoring area is obtained, and the division accuracy of the monitoring area in the power distribution cabinet infrared image is improved.

[0123] It should be noted that the above-mentioned sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0124] Each embodiment in the specification is described in a progressive manner, and the same or similar parts of each embodiment can be referred to each other, and each embodiment mainly describes the differences from other embodiments.

Claims

1. A power distribution cabinet operation anomaly monitoring system, characterized in that, The system includes the following modules: The data acquisition module is used to acquire infrared images of the power distribution cabinet; The first data processing module is used to determine the heat source window based on the grayscale distribution of the infrared image of the power distribution cabinet; draw straight lines in different directions through the center pixel of each heat source window, filter the straight lines in different directions to obtain the optimal straight line for each heat source window; and obtain the consistency of change of each parallel line based on the grayscale consistency between parallel lines parallel to the optimal straight line, including: for each parallel line of the optimal straight line, constructing a parallel line sequence based on the grayscale values ​​of the pixels on the parallel lines; for any two parallel lines corresponding to the optimal straight line, calculating the similarity between the two parallel line sequences; for any parallel line corresponding to the optimal straight line, obtaining the mean similarity between the parallel line and other parallel lines corresponding to the optimal straight line as the mean similarity; obtaining the minimum similarity between the parallel line and other parallel lines corresponding to the optimal straight line as the minimum similarity; and obtaining the normalized value of the product of the mean similarity and the minimum similarity as the consistency of change of the parallel line. The second data processing module is used to obtain the temperature change trend of the optimal straight line by fusing the gray values ​​of pixels in the optimal straight line and the corresponding parallel lines based on the consistency of change; determine the mutation window based on the temperature change trend and the consistency of change; and determine the degree of mutation based on the gray distribution in the mutation window. The region division module is used to correct the initial corner threshold by combining the degree of mutation and the consistency of change to obtain the corrected corner threshold; according to the corrected corner threshold, corner detection is performed on the infrared image of the power distribution cabinet to obtain corners; and the infrared image of the power distribution cabinet is divided into regions according to the corners to obtain the monitoring area. The early warning module is used to provide early warning signals to the monitored area based on the infrared image of the power distribution cabinet.

2. The power distribution cabinet operation abnormality monitoring system according to claim 1, characterized by, The process of filtering straight lines in different directions to obtain the optimal straight line for each heat source window includes: The pixels within the heat source window are used as heat source pixels; For any heat source window, obtain the number of heat source pixels on straight lines passing through the center pixel of the heat source window in different directions, and use this as the discrimination number. The straight line with the largest discrimination number is taken as the optimal straight line of the heat source window.

3. The power distribution cabinet operation abnormality monitoring system according to claim 1, characterized by, The method of obtaining the temperature change trend of the optimal straight line by fusing the grayscale values ​​of pixels in the optimal straight line and its corresponding parallel lines based on change consistency includes: Among the parallel lines corresponding to the optimal line, the one with the greatest consistency of change is taken as the main line; using the consistency of change as the weight, the gray values ​​of each pixel on all parallel lines corresponding to the optimal line are weighted and fused to obtain the fused pixel value, which is then assigned to each pixel on the main line; the sequence of pixels on the main line is taken as the temperature change trend of the optimal line.

4. The power distribution cabinet operation abnormality monitoring system according to claim 3, characterized by, The step of determining the abrupt change window based on the temperature change trend and consistency includes: When the consistency of the main line changes is less than the preset consistency threshold, curve fitting is performed on the fused pixel values ​​of each pixel on the main line to obtain the fused fitting curve. The average of the slopes of the points on the fusion fitting curve is calculated as a slope average, and when the slopes of a preset number of consecutive points on the fusion fitting curve are greater than the slope average, the heat source window corresponding to the fusion fitting curve is taken as a prominent window, where the preset number is a positive integer.

5. The power distribution cabinet operation abnormality monitoring system according to claim 1, characterized by, The mutation degree is determined according to the gray distribution in the mutation window, and the mutation degree includes: The gray values of the pixel points in the mutation window are fitted to obtain a window fitting curve, and a normalized value of the variance of the slopes of the points corresponding to the gray values of all the pixel points on the window fitting curve is taken as the mutation degree.

6. The power distribution cabinet operation abnormality monitoring system according to claim 1, characterized by, The initial corner point threshold is corrected in combination with the mutation degree and the change consistency to obtain a corrected corner point threshold, and the correction includes: When the heat source window is a non-mutation window, the product of the initial corner point threshold and the change consistency of the optimal straight line of the heat source window is taken as the corrected corner point threshold of the optimal straight line of the heat source window; When the heat source window is a mutation window, a difference between a preset threshold and the mutation degree is taken as an adjustment degree, and the product of the initial corner point threshold, the change consistency of the optimal straight line of the heat source window and the adjustment degree of the heat source window is taken as the corrected corner point threshold of the optimal straight line of the heat source window, where the preset threshold has a value greater than the mutation degree.

7. The power distribution cabinet operation abnormality monitoring system according to claim 1, characterized by, The heat source window is determined according to the gray distribution of the power distribution cabinet infrared image, and the determination includes: A sliding window is slid on the power distribution cabinet infrared image, and the region corresponding to the sliding window in which the gray average of the pixel points in the sliding window is greater than the gray average of the power distribution cabinet infrared image is taken as a high-heat region; The heat source window is determined according to the gray distribution in the high-heat region.

8. The power distribution cabinet operation abnormality monitoring system according to claim 7, characterized by, The heat source window is determined according to the gray distribution in the high-heat region, and the determination includes: For any high-heat region, the difference between the gray average in the high-heat region and the gray average in all the high-heat regions is calculated as the gray average difference of the high-heat region; The variance of the gray values of the pixel points in the high-heat region is inversely proportional to the initial probability of the high-heat region; The product of the gray average difference and the initial probability of the high-heat region is normalized to obtain the heat generation probability of the high-heat region; The heat source window is selected from the high-heat region according to the heat generation probability.

9. The power distribution cabinet operation abnormality monitoring system according to claim 1, characterized by, The monitoring area feedback early warning signal based on the power distribution cabinet infrared image includes: The temperature of each monitoring area is obtained, and when the temperature of the monitoring area does not belong to the normal range of the corresponding actual area, the power distribution cabinet is warned.

Images