Power equipment defect detection method and system based on machine vision

Abstract

The invention relates to the technical field of electrical equipment defect detection, in particular to an electrical equipment defect detection method and system based on machine vision, and the method comprises a modeling module which is used for uploading structural parameters of electrical equipment and constructing a three-dimensional model of the electrical equipment based on the structural parameters of the electrical equipment; the recognition module is used for recognizing the output posture of the power equipment on the production line; the coordination module is used for receiving the output posture of the power equipment recognized in the recognition module; according to the method, the appearance three-dimensional model of the power equipment is constructed through the structural parameters of the power equipment, and the appearance three-dimensional model of the power equipment is coordinated based on appearance image acquisition of the power equipment produced by the production line, so that the model is consistent with the posture of the power equipment produced by the production line; image data consistent with the visual angle of the appearance image of the power equipment produced by the production line is obtained from the model and used for similarity comparison, and finally defect judgment is conducted on the appearance of the power equipment based on the similarity comparison result.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment defect detection, and in particular to a power equipment defect detection method and system based on machine vision. Background Art

[0002] Defect detection of power equipment is extremely important. It uses advanced technologies such as infrared temperature measurement and partial discharge detection to conduct a comprehensive "physical examination" of the equipment. It can promptly detect defects such as overheating and insulation damage of the equipment, prevent failures from occurring, ensure the stable operation of the power system, and provide reliable power supply for production and life.

[0003] The invention patent with application number 202111312290.X discloses a method for detecting defects in power equipment based on a mixed reality device, which is characterized in that it includes the following steps: Step 1, the mixed reality device is connected to the monitoring background, and the preset power equipment inspection plan is imported into the mixed reality device, and the mixed reality device displays the power equipment inspection plan to the staff; Step 2, the staff determines the power equipment to be inspected according to the metering inspection plan, and uses the camera of the mixed reality device to capture the overall appearance image of the power equipment, and performs image quality verification on the overall appearance image of the power equipment. If the image quality verification passes, execute step 3; if the image quality verification fails, re-capture the overall appearance image of the power equipment through the camera of the mixed reality device; Step 3, perform device recognition based on the overall appearance image of the power equipment, and select the device from the image based on the recognition result. The monitoring background obtains the power equipment archive, and retrieves the equipment information of the power equipment in the power equipment archive, and displays the equipment information on the mixed reality device. The mixed reality device displays the corresponding operation steps of the power equipment in the power equipment inspection plan, and the staff conducts inspections according to the equipment information displayed by the mixed reality device and the operation steps of the power equipment inspection plan; Step 4, during the inspection process, the mixed reality device collects the image of the power equipment in real time through the camera, and performs basic operation detection of the power equipment according to the power equipment archive and the power equipment image, and extracts the display area image of the power equipment image at the same time, obtains the metering data in the display area image, compares the metering data with the data records in the power equipment archive, and classifies the defects of the power equipment according to the basic operation detection results and the comparison results of the metering data, and uploads the defect classification results to the monitoring background.

[0004] The application aims to solve the problem that "in the field of power equipment inspection technology, the existing applications of mixed reality devices mostly only stay at the stage of collecting and displaying on-site data. It is difficult to effectively judge the operating status of on-site equipment and equipment failures or defects. Operators are still required to make their own judgments and perform troubleshooting and repairs, resulting in low inspection efficiency."

[0005] However, in the production process of power equipment, its appearance inspection is also particularly important. A good power equipment casing can provide protection for its internal electrical components and provide a long and stable service life for the power equipment.

[0006] To this end, we proposed a power equipment defect detection system based on machine vision. Summary of the invention

[0007] In view of the above-mentioned shortcomings of the prior art, the present invention provides a method and system for detecting defects in electric power equipment based on machine vision, which solves the technical problems raised in the above-mentioned background technology.

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0009] In a first aspect, a power equipment defect detection system based on machine vision comprises:

[0010] A modeling module is used to upload the structural parameters of the power equipment and build a three-dimensional model of the power equipment based on the structural parameters of the power equipment; an identification module is used to identify the posture of the power equipment output on its production line; a coordination module is used to receive the output posture of the power equipment identified in the identification module, and coordinate the three-dimensional model of the power equipment constructed in the modeling module based on the output posture of the power equipment, so that the posture of the three-dimensional model of the power equipment is consistent with the output posture of the power equipment; a camera module is used to collect image data of the power equipment when the power equipment is output on its production line; a retrieval module is used to retrieve a model image on the three-dimensional model of the power equipment that has completed posture coordination and has the same perspective as the image data of the power equipment collected by the camera module; a judgment module is used to receive the image data and model image of the power equipment collected by the camera module, and judge whether the power equipment from which the image data of the power equipment is derived has defects based on the similarity comparison between the image data of the power equipment and the model image.

[0011] Furthermore, the structural parameters of the electric equipment used in the modeling module when constructing the three-dimensional model of the electric equipment, that is, the size of all the parts constituting the electric equipment and the relative position information of the joint assembly;

[0012] The identification module is integrated by a plurality of groups of ranging sensors, which are arranged in a matrix on the same plane, and the plane on which the ranging sensors are arranged is deployed at the output end of the production line, perpendicular to the upper surface of the transmission surface of the production line. When the output end of the production line outputs power equipment, all ranging modules on the plane run to perform ranging operations;

[0013] Among them, the distance between the distance measuring sensors arranged in a matrix on the same plane is customized by the system end user, and it obeys the setting logic that the higher the requirement for power equipment defect detection accuracy, the smaller the distance, and the lower the power equipment defect detection accuracy, the larger the distance.

[0014] Furthermore, the ranging ends of the ranging modules are all arranged opposite to the transmission surface of the power equipment production line. When the power equipment is output on the production line, all ranging modules synchronously perform ranging operations, and further create a set of planes in the three-dimensional space. Line segments perpendicular to the planes are drawn on the created planes based on the ranging results, and the position distribution state of each set of drawn line segments is consistent with the position distribution state of the ranging sensor on the plane.

[0015] Based on the adjacent connection between the endpoints of each drawn line segment on the plane in the three-dimensional space that are far away from the plane, a group of combined surfaces obtained by splicing a plurality of groups of finite surfaces are constructed;

[0016] Among them, a set of planes created in the three-dimensional space are consistent with the deployment surface of the ranging sensor in size, direction and angle.

[0017] Furthermore, during the operation phase of the coordination module, the combined surface constructed based on the operation of the ranging module in the identification module is placed in the three-dimensional space for constructing the three-dimensional model of the power equipment, and the three-dimensional model of the power equipment is continuously rotated and translated until several groups of surfaces on the three-dimensional model of the power equipment can completely overlap with all surfaces on the combined surface, and the coordination module completes the posture coordination of the three-dimensional model of the power equipment;

[0018] When the combined surface is placed in the three-dimensional space for constructing the three-dimensional model of the electric power equipment, the posture of the combined surface relative to the three-dimensional coordinate axis system of the three-dimensional space does not change.

[0019] Furthermore, the camera module is provided with three groups of industrial cameras, which are respectively deployed on the left and right sides of the production line and on the side opposite to the output end of the production line, and the camera ends of the industrial cameras deployed on the left and right sides of the production line are opposite to each other, and the camera end of the industrial camera deployed on the side opposite to the output end of the production line is centered based on the production line;

[0020] The camera module is provided with submodules at the lower level, including:

[0021] A storage unit, used for receiving the image data of the electric power equipment collected by the camera module and storing the image data of the electric power equipment;

[0022] An extraction module, used for traversing the power equipment image data stored in the storage unit, and extracting the power equipment contour image from the power equipment image data;

[0023] After the extraction module runs to extract the contour image of the power equipment image data, it synchronously iterates to the corresponding power equipment image data stored in the storage unit.

[0024] Furthermore, when the industrial camera collects the image data of the power equipment, the collection frequency is not less than 60 groups / second. After completing the collection of the image data of the power equipment, the quality of each group of collected image data of the power equipment is further identified, and a group of the power equipment image data with the highest quality is retained for further processing by the storage unit and the extraction unit;

[0025] The recognition logic of the power equipment image data is:

[0026]

[0027] Where: Q is the quality of the power equipment image data; C, K, N are the clarity, contrast and noise level of the power equipment image data; ω C ,ω K ,ω A ,ω N is the weight; M and N are the number of rows and columns of pixels in the power equipment image data; d(x, y) is the color difference between the pixel at the position (x, y) in the power equipment image data and the pixel at the same position in the reference image;

[0028] Among them, when the extraction module extracts the contour image of the power equipment, it first converts the power equipment image data into a grayscale image, and then sets the pixel grayscale value to extract the target threshold. The image presented by the pixels that meet the threshold in the grayscale image determined by the pixel grayscale value extraction target threshold is the contour image of the power equipment.

[0029] Furthermore, the reference image is the first group of collected power equipment image data among the collected power equipment image data, and the weight ω C ,ω K ,ω A ,ω N The values ​​of are 0.4, 0.3, 0.2, and 0.1, and the calculation logic of the color difference d(x, y) is:

[0030]

[0031] Where: R(x,y), G(x,y), B(x,y) are the values ​​of the pixel at position (x,y) of the power equipment image data based on the three channels of R, G, and B; R ref (x,y),G ref (x,y),B ref (x, y) is the value of the pixel at position (x, y) of the reference image based on the R, G, and B channels.

[0032] Furthermore, a similarity determination threshold is set in the determination module, and the determination module determines whether the power equipment from which the power equipment image data comes is defective based on the similarity determination threshold;

[0033]

[0034] Where: sim(a,b) is the similarity between the power equipment image data and the model image; w and h are the width and height of the image; C(x,y) is the pixel value of the pixel with coordinates (x,y) in the power equipment image data; I(x,y) is the pixel value of the pixel with coordinates (x,y) in the model image; S is the defined matching condition range; is the weight; d is the center coordinate position deviation between the power equipment image data and the model image; d max is the diagonal length of the power equipment image data;

[0035] The defined matching condition range S is defined by the system user, [·] is the condition judgment, if it is established, then [·] = 1, otherwise, then [·] = 0, that is, when C(x, y) = 1 and I(x, y)∈S, [·] = 1, otherwise, then [·] = 0, weight The sum is 1. Based on the above formula, each group of power equipment image data and its corresponding model image are judged. When sim(a, b) of any group of power equipment image data and its corresponding model image does not meet the similarity judgment threshold, it is judged that the power equipment from which the power equipment image data comes is defective.

[0036] Furthermore, the modeling module is interactively connected to the recognition module, the coordination module and the camera module through a wireless network, the camera module is interactively connected to the storage unit and the extraction unit through a wireless network, and the camera module is interactively connected to the retrieval module and the determination module through a wireless network.

[0037] In a second aspect, a method for detecting defects in power equipment based on machine vision comprises:

[0038] Determine whether the electric equipment has defects based on the electric equipment defect detection system, and when the determination result is yes, obtain the electric equipment image data of the determination source and its corresponding model image;

[0039] Segment the acquired power equipment image data and its corresponding model image to obtain u×u groups of sub-power equipment image data and u×u groups of sub-model images, so that the size of each group of sub-power equipment image data and the sub-model image is equal;

[0040] The similarity between each group of sub-power equipment image data and its corresponding sub-model image is calculated and compared with the similarity determination threshold, and the defective area on the defective power equipment is determined based on the comparison result.

[0041] Compared with the known public technology, the technical solution provided by the present invention has the following beneficial effects:

[0042] The present invention provides an electric equipment defect detection system based on machine vision. During operation, the system constructs a three-dimensional model of the appearance of the electric equipment through the structural parameters of the electric equipment, further coordinates the three-dimensional model of the appearance of the electric equipment based on the acquisition of the appearance image of the electric equipment produced by the production line, so that the model is consistent with the posture of the electric equipment produced by the production line, and then obtains image data consistent with the perspective of the appearance image of the electric equipment produced by the production line from the model for similarity comparison, and finally determines the defects of the appearance of the electric equipment based on the similarity comparison result, and finally captures the area with defects on the surface of the electric equipment based on the electric equipment defect detection method, so as to provide quality inspection services for the production of electric equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 It is a structural schematic diagram of a power equipment defect detection system based on machine vision;

[0045] Figure 2 The figure is a flow chart of a method for defect detection of power equipment based on machine vision. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] The present invention will be further described below in conjunction with the embodiments.

[0048] Embodiment 1:

[0049] A power equipment defect detection system based on machine vision in this embodiment, such as Figure 1 As shown, including:

[0050] A modeling module, used for uploading structural parameters of power equipment and building a three-dimensional model of the power equipment based on the structural parameters of the power equipment;

[0051] An identification module, used to identify the output posture of the power equipment on its production line;

[0052] A coordination module, used to receive the output posture of the power equipment identified in the identification module, and coordinate the three-dimensional model of the power equipment constructed in the modeling module based on the output posture of the power equipment, so that the posture of the three-dimensional model of the power equipment is consistent with the output posture of the power equipment;

[0053] A camera module for collecting image data of the power equipment when it is output on its production line;

[0054] The camera module is equipped with three groups of industrial cameras, which are respectively deployed on the left and right sides of the production line and on the opposite side of the production line output end. The camera ends of the industrial cameras deployed on the left and right sides of the production line are arranged oppositely, and the camera end of the industrial camera deployed on the opposite side of the production line output end is centered based on the production line.

[0055] The camera module is equipped with sub-modules, including:

[0056] A storage unit, used for receiving the image data of the electric power equipment collected by the camera module and storing the image data of the electric power equipment;

[0057] An extraction module, used for traversing the power equipment image data stored in the storage unit, and extracting the power equipment contour image from the power equipment image data;

[0058] Wherein, after the extraction module runs to extract the contour image of the power equipment image data, it synchronously iterates to the corresponding power equipment image data stored in the storage unit;

[0059] When the industrial camera collects the image data of the power equipment, the collection frequency is not less than 60 groups / second. After completing the collection of the image data of the power equipment, the quality of each group of collected image data of the power equipment is further identified, and a group of the highest quality image data of the power equipment is retained for further processing by the storage unit and the extraction unit;

[0060] The recognition logic of power equipment image data is:

[0061]

[0062] Where: Q is the quality of the power equipment image data; C, K, N are the clarity, contrast and noise level of the power equipment image data; ω C ,ω K ,ω A ,ω Nis the weight; M and N are the number of rows and columns of pixels in the power equipment image data; d(x, y) is the color difference between the pixel at the position (x, y) in the power equipment image data and the pixel at the same position in the reference image;

[0063] When the extraction module extracts the contour image of the power equipment, it first converts the image data of the power equipment into a grayscale image, then sets the pixel grayscale value to extract the target threshold, and extracts the image presented by the pixels meeting the threshold in the grayscale image determined by the pixel grayscale value to extract the target threshold, i.e., the contour image of the power equipment;

[0064] The reference image is the first group of collected power equipment image data among the collected power equipment image data, and the weight ω C ,ω K ,ω A ,ω N The values ​​of are 0.4, 0.3, 0.2, and 0.1, and the calculation logic of color difference d(x,y) is:

[0065]

[0066] Where: R(x,y), G(x,y), B(x,y) are the values ​​of the pixel at position (x,y) of the power equipment image data based on the three channels of R, G, and B; R ref (x,y),G ref (x,y),B ref (x, y) is the value of the pixel at position (x, y) of the reference image based on the three channels of R, G, and B;

[0067] A calling module, used to call a model image on the three-dimensional model of the power equipment that has completed posture coordination and has the same viewing angle as the power equipment image data collected by the camera module;

[0068] A determination module, used for receiving the power equipment image data and the model image collected by the camera module, and determining whether the power equipment from which the power equipment image data comes is defective based on a similarity comparison between the power equipment image data and the model image;

[0069] A similarity determination threshold is set in the determination module, and the determination module determines whether the power equipment from which the power equipment image data comes is defective based on the similarity determination threshold;

[0070]

[0071] Where: sim(a,b) is the similarity between the power equipment image data and the model image; w and h are the width and height of the image; C(x,y) is the pixel value of the pixel with coordinates (x,y) in the power equipment image data; I(x,y) is the pixel value of the pixel with coordinates (x,y) in the model image; S is the defined matching condition range; is the weight; d is the center coordinate position deviation between the power equipment image data and the model image; d max is the diagonal length of the power equipment image data;

[0072] The defined matching condition range S is defined by the system user, [·] is the condition judgment, if it is established, then [·] = 1, otherwise, then [·] = 0, that is, when C(x, y) = 1 and I(x, y)∈S, [·] = 1, otherwise, then [·] = 0, weight The sum is 1, and each set of power equipment image data and its corresponding model image are judged based on the above formula. When sim(a, b) of any set of power equipment image data and its corresponding model image does not meet the similarity judgment threshold, it is judged that the power equipment from which the power equipment image data comes is defective;

[0073] The modeling module is interactively connected to the recognition module, the coordination module and the camera module through a wireless network. The camera module is interactively connected to the storage unit and the extraction unit through a wireless network. The camera module is interactively connected to the retrieval module and the determination module through a wireless network.

[0074] In this embodiment, the modeling module runs to upload the structural parameters of the power equipment, and constructs the three-dimensional model of the power equipment based on the structural parameters of the power equipment. The recognition module recognizes the posture of the power equipment output on its production line in real time. The coordination module then receives the output posture of the power equipment recognized in the recognition module, and coordinates the three-dimensional model of the power equipment constructed in the modeling module based on the output posture of the power equipment, so that the posture of the three-dimensional model of the power equipment is consistent with the output posture of the power equipment. The camera module further collects the image data of the power equipment when the power equipment is output on its production line. The storage unit synchronously receives the image data of the power equipment collected by the camera module and stores the image data of the power equipment. The extraction module traverses the image data of the power equipment stored in the storage unit in real time, extracts the contour image of the power equipment from the image data of the power equipment, and the retrieval module runs to retrieve the model image of the three-dimensional model of the power equipment with the same perspective as the image data of the power equipment collected by the camera module. Finally, the judgment module receives the image data of the power equipment collected by the camera module and the model image, and determines whether the power equipment from which the image data of the power equipment is derived has defects based on the similarity comparison between the image data of the power equipment and the model image.

[0075] Through the operation of the system in the above embodiment, appearance inspection is provided for the production of power equipment, the quality of the power equipment is ensured, and the power distribution service is stably completed in the power circuit system.

[0076] Embodiment 2:

[0077] The power equipment structural parameters used when constructing the three-dimensional model of the power equipment in the modeling module, that is, the size of all the parts and components constituting the power equipment and the relative position information of their joint assembly;

[0078] The identification module is integrated by several groups of ranging sensors, which are arranged in a matrix on the same plane. The plane where the ranging sensors are arranged is deployed at the output end of the production line, perpendicular to the upper surface of the transmission surface of the production line. When the output end of the production line outputs power equipment, all ranging modules on the plane run to perform ranging operations.

[0079] Among them, the distance between the distance measuring sensors arranged in a matrix on the same plane is customized by the system end user, which is subject to the setting logic that the higher the requirement for power equipment defect detection accuracy, the smaller the distance, and the lower the power equipment defect detection accuracy, the larger the distance;

[0080] The ranging ends of the ranging modules are all opposite to the transmission surface of the power equipment production line. When the power equipment is output on the production line, all ranging modules synchronously perform ranging operations, further create a set of planes in three-dimensional space, and draw line segments perpendicular to the planes on the created planes based on the ranging results. The position distribution state of each set of drawn line segments is consistent with the position distribution state of the ranging sensor on the plane.

[0081] Based on the three-dimensional space, the endpoints of each drawn line segment on the plane far away from the plane are adjacently connected to each other to construct a group of combined surfaces obtained by splicing a number of groups of finite surfaces;

[0082] A set of planes are created in three-dimensional space, which are consistent with the deployment surface of the ranging sensor in size, direction and angle.

[0083] During the operation phase of the coordination module, the combined surface constructed based on the operation of the ranging module in the identification module is placed in the three-dimensional space for constructing the three-dimensional model of the power equipment, and the three-dimensional model of the power equipment is continuously rotated and translated until several groups of surfaces on the three-dimensional model of the power equipment can completely overlap with all surfaces on the combined surface. The coordination module then completes the posture coordination of the three-dimensional model of the power equipment.

[0084] When the combined surface is placed in the three-dimensional space for constructing the three-dimensional model of the electric power equipment, the posture of the combined surface relative to the three-dimensional coordinate axis system of the three-dimensional space does not change.

[0085] In this embodiment, through the above settings, further module operation data and logic support are provided for the system in Example 1 to ensure stable operation of the system in Example 1 and detect defects in power equipment.

[0086] Embodiment 3:

[0087] A method for detecting defects in electric power equipment based on machine vision, comprising:

[0088] Determine whether the electric equipment has defects based on the electric equipment defect detection system, and when the determination result is yes, obtain the electric equipment image data of the determination source and its corresponding model image;

[0089] Segment the acquired power equipment image data and its corresponding model image to obtain u×u groups of sub-power equipment image data and u×u groups of sub-model images, so that the size of each group of sub-power equipment image data and the sub-model image is equal;

[0090] The similarity between each group of sub-power equipment image data and its corresponding sub-model image is calculated and compared with the similarity determination threshold, and the defective area on the defective power equipment is determined based on the comparison result.

[0091] In summary, in the above embodiment, the system constructs a three-dimensional model of the appearance of the power equipment through the structural parameters of the power equipment, and further coordinates the three-dimensional model of the appearance of the power equipment based on the acquisition of the appearance image of the power equipment produced by the production line, so that the model is consistent with the posture of the power equipment produced by the production line, and then obtains image data from the model that is consistent with the perspective of the appearance image of the power equipment produced by the production line for similarity comparison, and finally determines the defects of the appearance of the power equipment based on the similarity comparison results, and finally captures the areas with defects on the surface of the power equipment based on the power equipment defect detection method, so as to provide quality inspection services for the production of power equipment.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power equipment defect detection system based on machine vision, characterized in that: include: A modeling module, used to upload the structural parameters of the power equipment and build a three-dimensional model of the power equipment based on the structural parameters of the power equipment; An identification module, used to identify the output posture of the power equipment on its production line; A coordination module, used to receive the output posture of the power equipment identified in the identification module, and coordinate the three-dimensional model of the power equipment constructed in the modeling module based on the output posture of the power equipment, so that the posture of the three-dimensional model of the power equipment is consistent with the output posture of the power equipment; A camera module for collecting image data of the power equipment when it is output on its production line; A calling module, used to call a model image on the three-dimensional model of the power equipment that has completed posture coordination and has the same viewing angle as the power equipment image data collected by the camera module; The determination module is used to receive the power equipment image data and the model image collected by the camera module, and determine whether there is a defect in the power equipment from which the power equipment image data comes based on the similarity comparison between the power equipment image data and the model image.

2. The power equipment defect detection system based on machine vision according to claim 1 is characterized in that: The structural parameters of the electric equipment used when constructing the three-dimensional model of the electric equipment in the modeling module, that is, the dimensions of all the parts constituting the electric equipment and the relative position information of their joint assembly; The identification module is integrated by a plurality of groups of ranging sensors, which are arranged in a matrix on the same plane, and the plane on which the ranging sensors are arranged is deployed at the output end of the production line, perpendicular to the upper surface of the transmission surface of the production line. When the output end of the production line outputs power equipment, all ranging modules on the plane run to perform ranging operations; Among them, the distance between the distance measuring sensors arranged in a matrix on the same plane is customized by the system end user, and it obeys the setting logic: the higher the requirement for power equipment defect detection accuracy, the smaller the distance, and the lower the power equipment defect detection accuracy, the larger the distance.

3. The power equipment defect detection system based on machine vision according to claim 2 is characterized in that: The ranging ends of the ranging modules are all arranged opposite to the transmission surface of the power equipment production line. When the power equipment is output on the production line, all ranging modules synchronously perform ranging operations, further create a set of planes in the three-dimensional space, and draw line segments perpendicular to the planes on the created planes based on the ranging results, and the position distribution state of each set of drawn line segments is consistent with the position distribution state of the ranging sensor on the plane; Based on the adjacent connection between the endpoints of each drawn line segment on the plane in the three-dimensional space that are far away from the plane, a group of combined surfaces obtained by splicing a plurality of groups of finite surfaces are constructed; Among them, a set of planes created in the three-dimensional space are consistent with the deployment surface of the ranging sensor in size, direction and angle.

4. The power equipment defect detection system based on machine vision according to claim 3 is characterized in that: During the operation phase of the coordination module, the combined surface constructed based on the operation of the ranging module in the identification module is placed in the three-dimensional space for constructing the three-dimensional model of the power equipment, and the three-dimensional model of the power equipment is continuously rotated and translated until several groups of surfaces on the three-dimensional model of the power equipment can completely overlap with all surfaces on the combined surface, and the coordination module completes the posture coordination of the three-dimensional model of the power equipment; When the combined surface is placed in the three-dimensional space for constructing the three-dimensional model of the electric power equipment, the posture of the combined surface relative to the three-dimensional coordinate axis system of the three-dimensional space does not change.

5. The power equipment defect detection system based on machine vision according to claim 1 is characterized in that: The camera module is provided with three groups of industrial cameras, which are respectively deployed on the left and right sides of the production line and on the side opposite to the output end of the production line, and the camera ends of the industrial cameras deployed on the left and right sides of the production line are opposite to each other, and the camera end of the industrial camera deployed on the side opposite to the output end of the production line is centered based on the production line; The camera module is provided with submodules at the lower level, including: A storage unit, used for receiving the image data of the electric power equipment collected by the camera module and storing the image data of the electric power equipment; An extraction module, used for traversing the power equipment image data stored in the storage unit, and extracting the power equipment contour image from the power equipment image data; After the extraction module runs to extract the contour image of the power equipment image data, it synchronously iterates to the corresponding power equipment image data stored in the storage unit.

6. The power equipment defect detection system based on machine vision according to claim 5, characterized in that: When the industrial camera collects the image data of the power equipment, the collection frequency is not less than 60 groups / second. After completing the collection of the image data of the power equipment, the quality of each group of collected image data of the power equipment is further identified, and a group of the power equipment image data with the highest quality is retained for further processing by the storage unit and the extraction unit; The recognition logic of the power equipment image data is: Where: Q is the quality of the power equipment image data; C, K, N are the clarity, contrast and noise level of the power equipment image data; ω C ,ω K ,ω A ,ω N is the weight; M and N are the number of rows and columns of pixels in the power equipment image data; d(x, y) is the color difference between the pixel at the position (x, y) in the power equipment image data and the pixel at the same position in the reference image; Among them, when the extraction module extracts the contour image of the power equipment, it first converts the power equipment image data into a grayscale image, and then sets the pixel grayscale value to extract the target threshold. The image presented by the pixels that meet the threshold in the grayscale image determined by the pixel grayscale value extraction target threshold is the contour image of the power equipment.

7. The power equipment defect detection system based on machine vision according to claim 6, characterized in that: The reference image is the first group of collected power equipment image data in the collected power equipment image data, and the weight ω C ,ω K ,ω A ,ω N The values ​​of are 0.4, 0.3, 0.2, and 0.1, and the calculation logic of the color difference d(x, y) is: Where: R(x,y), G(x,y), B(x,y) are the values ​​of the pixel at position (x,y) of the power equipment image data based on the three channels of R, G, and B; R ref (x,y),G ref (x,y),B ref (x, y) is the value of the pixel at position (x, y) of the reference image based on the R, G, and B channels.

8. The power equipment defect detection system based on machine vision according to claim 1, characterized in that: A similarity determination threshold is set in the determination module, and the determination module determines whether the power equipment from which the power equipment image data comes is defective based on the similarity determination threshold; Where: sim(a,b) is the similarity between the power equipment image data and the model image; w and h are the width and height of the image; C(x,y) is the pixel value of the pixel with coordinates (x,y) in the power equipment image data; I(x,y) is the pixel value of the pixel with coordinates (x,y) in the model image; S is the defined matching condition range; is the weight; d is the center coordinate position deviation between the power equipment image data and the model image; d max is the diagonal length of the power equipment image data; The defined matching condition range S is defined by the system user, [·] is the condition judgment, if it is established, then [·] = 1, otherwise, then [·] = 0, that is, when C(x, y) = 1 and I(x, y)∈S, [·] = 1, otherwise, then [·] = 0, weight The sum is 1. Based on the above formula, each group of power equipment image data and its corresponding model image are judged. When sim(a, b) of any group of power equipment image data and its corresponding model image does not meet the similarity judgment threshold, it is judged that the power equipment from which the power equipment image data comes is defective.

9. The power equipment defect detection system based on machine vision according to claim 1, characterized in that: The modeling module is interactively connected to the recognition module, the coordination module and the camera module through a wireless network. The camera module is interactively connected to the storage unit and the extraction unit through a wireless network. The camera module is interactively connected to the retrieval module and the determination module through a wireless network.

10. A method for detecting defects in electric equipment based on machine vision, the method being an implementation method of a system for detecting defects in electric equipment based on machine vision as claimed in any one of claims 1 to 9, characterized in that: include: Determine whether the electric equipment has defects based on the electric equipment defect detection system, and when the determination result is yes, obtain the electric equipment image data of the determination source and its corresponding model image; Segment the acquired power equipment image data and its corresponding model image to obtain u×u groups of sub-power equipment image data and u×u groups of sub-model images, so that the size of each group of sub-power equipment image data and the sub-model image is equal; The similarity between each group of sub-power equipment image data and its corresponding sub-model image is calculated and compared with the similarity determination threshold, and the defective area on the defective power equipment is determined based on the comparison result.

Images