Intelligent positioning method for fault heating area of power equipment based on infrared image

Abstract

The present application relates to the technical field of image processing, more particularly, the present application relates to a power equipment fault heating area intelligent positioning method based on infrared image, the method comprises the following steps: collecting the infrared image of power equipment, obtaining the heat source pixel point of each candidate high temperature area, obtaining the centripetal weight of the heat source pixel point of each candidate high temperature area; according to the centripetal weight of the heat source pixel point of each candidate high temperature area and the laplace operator value of each candidate high temperature area, the heat source convergence degree of each candidate high temperature area is obtained, the surface reflection possibility of each candidate high temperature area is obtained, according to the surface reflection possibility and the heat source convergence degree of each candidate high temperature area, the fault possibility of each candidate high temperature area is obtained, according to the fault possibility of each candidate high temperature area, the fault heating area is obtained, the accuracy of the fault heating area extraction is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing. More particularly, the present application relates to an intelligent positioning method for fault heating areas of power equipment based on infrared images. BACKGROUND

[0002] In the power system, an infrared thermal imager has become an important technical means for realizing preventive maintenance and ensuring safe and stable operation of the power grid by capturing thermal distribution images of equipment. Equipment failure usually causes local temperature to abnormally rise, especially due to poor contact, material aging or equipment overload, etc. Therefore, by collecting infrared images through an infrared thermal imager, high-temperature areas in the infrared images are extracted for fault judgment.

[0003] However, the existing automatic diagnosis technology mainly relies on image processing algorithms to analyze infrared images, and the core process is to extract high-temperature areas through image segmentation algorithms. These high-temperature areas are usually considered as signs of equipment failure.

[0004] However, this method has significant limitations in practical application, especially in outdoor inspection scenarios. When sunlight shines on the smooth surface (metal fittings) of power equipment, reflection will cause high-brightness spot areas to appear on the infrared image. The brightness of these spot areas is often very similar to the fault heating areas caused by equipment failure. Therefore, the existing image segmentation algorithm can obtain high-temperature areas after segmenting the infrared image, but cannot effectively distinguish whether the high-temperature areas are fault heating areas generated by equipment heating or spot areas generated by sunlight reflection. This leads to the fact that the spot areas caused by sunlight reflection are often misjudged as fault heating areas, thereby increasing the frequency of false alarms, affecting the accuracy of the diagnosis system, greatly reducing the practicality and reliability of the automatic system, and increasing the burden on maintenance personnel. SUMMARY

[0005] In order to solve the problem that the brightness of the fault heating area of the equipment and the spot area generated by sunlight reflection is similar, causing the spot area caused by sunlight reflection to be misjudged as the equipment failure area, the present application proposes an intelligent positioning method for fault heating areas of power equipment based on infrared images. The method includes the following steps:

[0006] Collecting an infrared image of power equipment; segmenting the infrared image of the power equipment to obtain a plurality of candidate high-temperature areas; obtaining heat source pixel points of each candidate high-temperature area;

[0007] According to the gradient vector of the neighborhood pixel point of the heat source pixel point, the centripetal weight of the heat source pixel point of each candidate high temperature area is obtained;According to the centripetal weight of the heat source pixel point of the candidate high temperature area, the Laplace operator value of the heat source pixel point, and the difference value between the temperature value of the neighborhood pixel point of the heat source pixel point and the average temperature of the edge pixel point of the candidate high temperature area, the heat source convergence degree of each candidate high temperature area is obtained;

[0008] According to the curvature variance and gradient variance of the edge pixel point of the candidate high temperature area, the surface reflection possibility of each candidate high temperature area is obtained;According to the surface reflection possibility and the heat source convergence degree, the failure possibility of each candidate high temperature area is obtained;According to the failure possibility of each candidate high temperature area, the failure heating area is obtained.

[0009] The innovation of the application lies in that according to the characteristics of the failure heating area and the light spot area, the heat source convergence degree of each candidate high temperature area is obtained, then according to the curvature variance and gradient variance of the edge pixel point of the candidate high temperature area, the surface reflection possibility of each candidate high temperature area is obtained, and according to the surface reflection possibility and the heat source convergence degree, the failure possibility of each candidate high temperature area is obtained, which can distinguish the failure heating area from the light spot area, obtain the accurate failure heating area, improve the practicability and reliability of the automatic system, and reduce the burden of maintenance personnel.

[0010] Preferably, the power equipment infrared image is segmented to obtain a plurality of candidate high temperature areas, comprising:

[0011] The power equipment infrared image is segmented by using the maximum inter-class variance method to generate a binary image, and the binary image is subjected to morphological closing operation to obtain all connected high temperature areas, which are recorded as each candidate high temperature area.

[0012] Preferably, the heat source pixel point of each candidate high temperature area is obtained, comprising:

[0013] The pixel point with the highest temperature value in each candidate high temperature area is obtained as the heat source pixel point of each candidate high temperature area.

[0014] Preferably, the centripetal weight of the heat source pixel point of each candidate high temperature area is obtained, comprising:

[0015] The neighborhood pixel point of the heat source pixel point in each candidate high temperature area is obtained.

[0016] , The centripetal weight of the heat source pixel point in the i th candidate high temperature area is represented; The number of neighborhood pixel points of the heat source pixel point in the i th candidate high temperature area is represented; a gradient vector of an nth neighborhood pixel point representing a heat source pixel point in the ith candidate high-temperature region; a direction vector of the nth neighborhood pixel point representing the heat source pixel point in the ith candidate high-temperature region points to the heat source pixel point; max() represents a maximum value function; a modulus of the vector.

[0017] Preferably, the heat source convergence degree of each candidate high-temperature region is obtained by:

[0018] ;

[0019] In the formula, the heat source convergence degree of the ith candidate high-temperature region; a temperature value of a jth neighborhood pixel point of a heat source pixel point of the ith candidate high-temperature region; a temperature mean value of all edge pixel points of the ith candidate high-temperature region; a number of neighborhood pixel points of the heat source pixel point of the ith candidate high-temperature region; a centripetal weight of the heat source pixel point of the ith candidate high-temperature region; a Laplace operator value of the heat source pixel point of the ith candidate high-temperature region; a maximum temperature value of all pixel points in the ith candidate high-temperature region; max() represents a maximum value function.

[0020] According to the centripetal weight of the heat source pixel point of the candidate high-temperature region, the Laplace operator value of the heat source pixel point, and the difference between the temperature value of the neighborhood pixel point of the heat source pixel point and the temperature mean value of the edge pixel point of the candidate high-temperature region, the judgment of the real fault heat generation region is no longer dependent on a single temperature value, but is based on a multidimensional comprehensive evaluation of the heat conduction physical law, which greatly improves the recognition ability of the real fault heat point.

[0021] Preferably, the surface reflection possibility of each candidate high-temperature region is obtained by:

[0022] , a surface reflection possibility of the ith candidate high-temperature region; a variance of curvatures of all edge pixel points of the ith candidate high-temperature region; a number of all edge pixel points of the ith candidate high-temperature region; a variance of gradients of all edge pixel points of the ith candidate high-temperature region; exp() represents an exponential function with a natural constant as a base number.

[0023] The curvature variance of all edge pixel points of the specular reflection light spot is small and the gradient variance is large, so that the concentrated light spot can be filtered out in subsequent preparation.

[0024] Preferably, the fault possibility of each candidate high-temperature region is obtained according to the surface reflection possibility and the heat source convergence degree, and the fault possibility comprises:

[0025] ;

[0026] In the formula, represents the fault possibility of the i th candidate high-temperature region; represents the heat source convergence degree of the i th candidate high-temperature region; represents the surface reflection possibility of the i th candidate high-temperature region; and norm() represents a normalization function.

[0027] The fault heat region can be accurately identified.

[0028] Preferably, the fault heat region is obtained according to the fault possibility of each candidate high-temperature region, and the obtaining comprises:

[0029] A preset fault possibility threshold T is set, and if the fault possibility of any candidate high-temperature region is greater than or equal to the fault possibility T, the candidate high-temperature region is recorded as the fault heat region.

[0030] The accuracy of identifying the fault heat region is improved.

[0031] Preferably, the infrared image of the power equipment is collected, and the collecting comprises:

[0032] When the power equipment is running, an infrared thermal imager is used to shoot the power equipment at different angles and distances to obtain the infrared image of the power equipment.

[0033] Preferably, the neighborhood pixel points of the heat source pixel points in each candidate high-temperature region are obtained, and the obtaining comprises:

[0034] A neighborhood window is constructed around the heat source pixel points of each candidate high-temperature region to obtain all neighborhood pixel points of the heat source pixel points of each candidate high-temperature region.

[0035] ​The present application has the following beneficial effects: the present application firstly obtains heat source convergence of each candidate high-temperature region according to the centripetal weight of the heat source pixel point of the candidate high-temperature region, the Laplacian value of the heat source pixel point and the difference between the temperature value of the neighborhood pixel point of the heat source pixel point and the temperature mean value of the edge pixel point of the candidate high-temperature region, so that the judgment of the real fault heat region is no longer dependent on a single temperature value, but a multidimensional comprehensive evaluation based on the heat conduction physical law, which greatly improves the identification ability of the real fault hot spot, then according to the curvature variance and gradient variance of the edge pixel point of the candidate high-temperature region, the surface reflection possibility of each candidate high-temperature region is obtained, according to the surface reflection possibility and the heat source convergence degree, the fault possibility of each candidate high-temperature region is obtained, which can distinguish the fault heat region from the light spot region, obtain the accurate fault heat region, improve the practicability and reliability of the automatic system, and reduce the burden of the maintenance personnel. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a step flow chart of the intelligent positioning method for the fault heat region of the power equipment based on the infrared image according to the embodiment of the present application. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be clearly and completely described in combination with the drawings in the embodiments of the present application.

[0038] Please refer to Figure 1 which shows the step flow chart of the intelligent positioning method for the fault heat region of the power equipment based on the infrared image according to an embodiment of the present application, and the method comprises the following steps:

[0039] S001, collecting the infrared image of the power equipment.

[0040] In the embodiment of the present application, when the power equipment is running, an infrared thermal imager is used to shoot the power equipment at different angles and distances to ensure that the main parts of the equipment (the high-voltage end, the connection end, the contact point or the wire of the equipment, etc.) are covered, and the infrared image of the power equipment is obtained.

[0041] The maximum inter-class variance method is used for adaptive threshold segmentation of the infrared image of the power equipment to generate a binary image, morphological closing operation is performed on the binary image to fill the holes in the region and remove isolated noise points, and all connected high-temperature regions are obtained, which are recorded as each candidate high-temperature region.

[0042] It should be noted that the core component of the infrared thermal imager is a sensor specially used to detect the intensity of infrared radiation emitted by the surface of an object, and according to the intensity, the instrument generates an infrared image, and each pixel point in the image represents the radiation intensity of the corresponding position, and the intensity value is accurately converted into a temperature value, so the value of each pixel point in the infrared image is a temperature value.

[0043] S002, the heat source pixel point of each candidate high temperature area is obtained, the centripetal weight of the heat source pixel point of each candidate high temperature area is obtained, and the heat source convergence degree of each candidate high temperature area is obtained according to the centripetal weight of the heat source pixel point of each candidate high temperature area and the Laplacian value of each candidate high temperature area.

[0044] It should be noted that the physical heat source of the fault point will form a heat source point at the center of the temperature field, and the temperature will spread to the surrounding area from the heat source point and gradually decrease, forming a certain heat concentration area, and the heat concentration area is the fault heating area, and the temperature distribution in the bright spot area formed by the reflection of sunlight is uniform and the noise point also does not have the above characteristics, so the heat source pixel point (the pixel point with the highest temperature) of each candidate high temperature area is obtained first, if the gradients of the neighborhood pixel points of the heat source pixel point of the candidate high temperature area all point to the heat source pixel point, then the candidate high temperature area is more likely to belong to the fault heating area, so the centripetal weight of the heat source pixel point of each candidate high temperature area is obtained first according to the characteristics, and the greater the value, the more likely the candidate high temperature area belongs to the fault heating area.

[0045] In the embodiment of the application, the pixel point with the highest temperature value in each candidate high temperature area is obtained as the heat source pixel point of each candidate high temperature area;

[0046] A neighborhood window is constructed with the heat source pixel point of each candidate high temperature area as the center , so as to obtain all neighborhood pixel points of the heat source pixel point of each candidate high temperature area;

[0047] The centripetal weight of the heat source pixel point of each candidate high temperature area is obtained:

[0048] ;

[0049] In the formula, represents the centripetal weight of the heat source pixel point in the i th candidate high temperature area; represents the number of neighborhood pixel points of the heat source pixel point in the i th candidate high temperature area; represents the gradient vector of the n th neighborhood pixel point of the heat source pixel point in the i th candidate high temperature area; represents the direction vector of the n th neighborhood pixel point of the heat source pixel point in the i th candidate high temperature area. represents the length of the representative vector; max() represents the maximum value function, in order to map the value range of to 0-1; The greater the value of the greater, the i-th candidate high temperature area, the direction vector of the neighborhood pixel point of the heat source pixel point in the i-th candidate high temperature area is similar to the gradient vector of the neighborhood pixel point, at this time it is indicated that the gradient vector of the neighborhood pixel point of the heat source pixel point points to the heat source pixel point, so the i-th candidate high temperature area is more likely to belong to the fault heating area.

[0050] It should be noted that the physical heat source of the known fault point will form a heat source point at the center of the temperature field, and the temperature will spread to the surrounding and gradually decrease from the heat source point, forming a certain heat concentration area, the heat concentration area is the fault heating area, therefore the temperature of the neighborhood pixel point of the heat source pixel point of the fault heating area is higher relative to the edge point of the fault heating area, which is more important for judging the fault heating area, and the temperature distribution inside the bright spot area is uniform, so if the temperature of the neighborhood pixel point of the heat source pixel point of the candidate high temperature area is higher relative to the edge point of the candidate high temperature area, then the candidate high temperature area is more likely to belong to the fault heating area.

[0051] And since the Laplace operator value is used to represent the change of the temperature field of a point, if the Laplace operator value of a point is negative, it indicates that the temperature of the point is higher relative to the surrounding area, which means that the point is the center of heat convergence, and heat spreads or concentrates to the surrounding, so if any candidate high temperature area is a fault heating area, the Laplace operator value of the heat source pixel point of the candidate high temperature area is negative, and the bright spot area formed by sunlight reflection is different from the fault heating area, the bright spot area formed by sunlight reflection will produce different performances in the two-dimensional temperature field, such as the temperature distribution inside the bright spot area is usually smooth and uniform, so the Laplace operator of the heat source pixel point of the bright spot area is close to zero, indicating that there is no temperature change inside the region (i.e. no heat concentration or diffusion trend); therefore, the present application combines the temperature difference between the neighborhood pixel point and the edge pixel point of the heat source pixel point of the candidate high temperature area, the Laplace operator value of the heat source pixel point of the candidate high temperature area, and the centripetal weight of the heat source pixel point, to obtain the heat source convergence degree of each candidate high temperature area.

[0052] In the embodiment of the present application, the heat source convergence degree of each candidate high temperature area is obtained as follows:

[0053] ;

[0054] In the formula, represents the heat source convergence degree of the i-th candidate high temperature area; represents the temperature value of the j-th neighborhood pixel point of the heat source pixel point of the i-th candidate high temperature area; The average temperature of all edge pixels representing the i-th candidate high-temperature region; The number of neighboring pixels of the heat source pixel in the i-th candidate high-temperature region; The centripetal weight of the heat source pixel representing the i-th candidate high-temperature region; The Laplacian operator value representing the heat source pixel of the i-th candidate high-temperature region; `max()` represents the maximum temperature of all pixels in the i-th candidate high-temperature region, used for normalization; `max()` represents the maximum value function.

[0055] It is known that the neighboring pixels of the heat source pixel in the faulty heating area have a higher temperature than the edge pixels of the faulty heating area, making them more important for identifying the faulty area. The larger the value, the greater the heat source concentration of the i-th candidate high temperature region, and the more likely the i-th candidate high temperature region is to be a fault heating region.

[0056] It is known that the physical heat source in the known fault heating area will form a heat source pixel in the temperature field, and the Laplacian operator value of the heat source pixel is significantly negative. Therefore, using... This makes the Laplace operator value at the heat source point positive, that is... The larger the value, the more likely the i-th candidate high-temperature region is to be a fault heating region. The value is more accurate; The larger the value, the more likely the gradient vectors of the neighboring pixels of the heat source pixel are to point to the heat source pixel. Therefore, the candidate high-temperature region is more likely to belong to the faulty heat source region. The value is more accurate.

[0057] S003. Obtain the surface reflection probability of each candidate high-temperature region, and obtain the failure probability of each candidate high-temperature region based on the surface reflection probability and the heat source concentration degree of each candidate high-temperature region.

[0058] It should be noted that the reflection phenomenon can be divided into diffuse reflection and specular reflection, when the reflected light irradiates the rough or uneven surface, the light will be scattered in multiple directions, usually in the infrared image, a uniform temperature distribution of the light spot area will be presented, but if the specular reflection occurs when the light irradiates the smooth surface, the light will be reflected at the equal angle of the incident angle, at this time, if the sunlight is reflected into the lens of the infrared thermal imager at the appropriate angle, a bright and concentrated light spot will appear on the infrared image, this light spot is focused, and a very small and high brightness point is formed on the infrared image, which is called a focused light spot; although the focused light spot is caused by reflection rather than a real fault heating area, but its temperature distribution may be very similar to the fault heating area, so the heat source convergence degree of the focused light spot is close to that of the fault heating area, so only relying on the heat source convergence degree index, the focused light spot may be misjudged as the fault heating area;

[0059] It is known that the profile of the fault heating area is determined by heat conduction and the physical structure of the object itself, and is usually irregular and not smooth, because the heat conduction speed is different in different materials or different thicknesses, resulting in irregular boundary of the fault heating area, so the curvature variance of the edge pixel points of the fault heating area will be large; and the specular reflection is most likely to occur on the smooth surface, and the profile formed on the infrared image of an ideal specular reflection is smooth in geometry, so the curvature variance of the edge pixel points of the focused light spot tends to zero;

[0060] It is known that the temperature distribution feature of the fault heating area is smooth and gradually downward from the center to the outside, so the gradient value of the edge pixel points of the fault heating area will be small and relatively uniform on the whole profile; on the contrary, the specular reflection only produces high temperature in a specific area, and the temperature of the surrounding area is not affected, so the temperature of the focused light spot will sharply decrease from the high temperature area to the surrounding low temperature area, resulting in large gradient difference of the edge pixel points of the focused light spot;

[0061] Therefore, the curvature variance and the gradient variance of the edge pixel points of the candidate high temperature area are combined to obtain the surface reflection possibility of each candidate high temperature area, and the greater the value is, the more likely the candidate high temperature area is the focused light spot area.

[0062] In the embodiment of the present application, the surface reflection possibility of each candidate high temperature area is obtained as follows:

[0063] ;

[0064] In the formula, represents the surface reflection possibility of the i th candidate high temperature area; represents the variance of the curvature of all edge pixel points of the i th candidate high temperature area; This represents the number of all edge pixels in the i-th candidate high-temperature region; represents the variance of the gradient of all edge pixels of the i-th candidate high-temperature region; exp() represents an exponential function with the natural constant as the base.

[0065] It should be noted that when the heat source convergence of the candidate high-temperature region is higher and the surface reflection probability of the candidate high-temperature region is greater, it indicates that the high heat source convergence of the candidate high-temperature region is caused by the concentrated light spot region. At this time, the heat source convergence should be reduced according to the surface reflection probability of the candidate high-temperature region to obtain a smaller probability of failure in the candidate high-temperature region and avoid subsequent misjudgment.

[0066] In this embodiment of the invention, the failure probability of each candidate high-temperature region is obtained:

[0067] ;

[0068] In the formula, This represents the probability of failure for the i-th candidate high-temperature region. This represents the heat source convergence degree of the i-th candidate high-temperature region; represents the surface reflection probability of the i-th candidate high-temperature region; norm() represents the normalization function;

[0069] The larger the value and The smaller the value, the more likely the i-th candidate high-temperature region is to be a fault heating region, and the greater its probability of failure. The larger the value and A larger value indicates that the i-th candidate high-temperature region is more likely to be a focused light spot generated by specular reflection. At this point, according to... The value will Decreasing the value of makes the probability of failure in the i-th candidate high-temperature region smaller;

[0070] The smaller the value, the more likely the i-th candidate high-temperature region is to be a light spot region (a light spot region generated by diffuse reflection). Whether the value is large or small, the result is The smaller the probability of failure, the lower the probability of failure in the i-th candidate high-temperature region.

[0071] S004. Based on the probability of failure of each candidate high-temperature area, obtain the faulty heating area.

[0072] In the embodiment of the present application, a preset failure possibility threshold T is provided, and if the failure possibility of any candidate high-temperature region is greater than or equal to the failure possibility T, the candidate high-temperature region is recorded as a failure heating region. It should be noted that the preset failure possibility threshold T=0.73, and in other embodiments, the implementer can preset the value of T according to the specific implementation.

[0073] The above merely describes the preferred embodiments of the present application and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. An intelligent positioning method for fault heating area of power equipment based on infrared image, characterized in that, The method comprises the steps of: collecting an infrared image of a power equipment; segmenting the infrared image of the power equipment to obtain a plurality of candidate high-temperature regions; According to the gradient vector of the neighbor pixel point of the heat source pixel point, a centripetal weight of the heat source pixel point of each candidate high-temperature region is obtained, including: obtaining the neighbor pixel point of the heat source pixel point in each candidate high-temperature region; , , is the centripetal weight of the heat source pixel point of the i th candidate high-temperature region, the number of neighbor pixel points, is the gradient vector of the n th neighbor pixel point of the heat source pixel point of the i th candidate high-temperature region, is the direction vector of the n th neighbor pixel point of the heat source pixel point of the i th candidate high-temperature region pointing to the heat source pixel point, and max() is a maximum function, is the module length of the vector; According to the centripetal weight of the heat source pixel point of the candidate high temperature area, the Laplacian value of the heat source pixel point, and the difference between the temperature value of the neighborhood pixel point of the heat source pixel point and the temperature mean value of the edge pixel point of the candidate high temperature area, the heat source convergence degree of each candidate high temperature area is obtained, comprising: , is the heat source convergence degree of the i th candidate high temperature area, is the temperature value of the j th neighborhood pixel point of the heat source pixel point of the i th candidate high temperature area, is the temperature mean value of all edge pixel points of the i th candidate high temperature area, is the Laplacian value of the heat source pixel point of the i th candidate high temperature area, is the maximum temperature value of all pixel points in the i th candidate high temperature area; According to the curvature variance and the gradient variance of the candidate high-temperature region edge pixel points, a surface reflection possibility of each candidate high-temperature region is obtained, including: , is a surface reflection possibility of the i th candidate high-temperature region, , , is the number of all edge pixel points, the curvature variance, and the gradient variance of the i th candidate high-temperature region, respectively, and exp() is an exponential function with a natural constant as a base number. According to the surface reflection possibility and the heat source convergence degree, the failure possibility of each candidate high-temperature region is obtained, including: , is the failure possibility of the i th candidate high-temperature region, and norm() is a normalization function; according to the failure possibility of each candidate high-temperature region, the failure heat-emitting region is obtained. 2.The infrared image-based power equipment fault heat area intelligent positioning method according to claim 1, characterized in that, obtaining a heat source pixel point of each candidate high-temperature region; the step of segmenting the infrared image of the power equipment to obtain a plurality of candidate high-temperature regions comprises the steps of: 3.The infrared image-based power equipment fault heat area intelligent positioning method according to claim 1, characterized in that, performing adaptive threshold segmentation on the infrared image of the power equipment using the maximum inter-class variance method to generate a binary image, and performing morphological closing operation on the binary image to obtain all connected high-temperature regions, which are recorded as each candidate high-temperature region. the step of obtaining a heat source pixel point of each candidate high-temperature region comprises the steps of: 4.The infrared image-based power equipment fault heat area intelligent positioning method according to claim 1, characterized in that, obtaining a pixel point with the highest temperature value in each candidate high-temperature region as the heat source pixel point of each candidate high-temperature region. the step of obtaining a fault heating region according to the fault possibility of each candidate high-temperature region comprises the steps of:

5. The method for intelligent positioning of the fault heating area of the power equipment based on infrared images according to claim 1, characterized in that, presetting a fault possibility threshold T, and if the fault possibility of any candidate high-temperature region is greater than or equal to the fault possibility T, the candidate high-temperature region is recorded as the fault heating region. the step of collecting the infrared image of the power equipment comprises the steps of:

6. The method of claim 1, wherein, when the power equipment is running, an infrared thermal imager is used to shoot the power equipment at different angles and distances to obtain the infrared image of the power equipment. the step of obtaining the neighborhood pixel points of the heat source pixel point in each candidate high-temperature region comprises the steps of: A neighborhood window is constructed with the heat source pixel point of each candidate high temperature region as the center, to obtain all neighborhood pixel points of the heat source pixel point of each candidate high temperature region. A neighborhood window is constructed with the heat source pixel point of each candidate high temperature region as the center, to obtain all neighborhood pixel points of the heat source pixel point of each candidate high temperature region.

Images