Relay armature air gap measurement method based on rotating template matching

Abstract

The application relates to the technical field of relays, and discloses a relay armature air gap measurement method based on rotation template matching, which is used for improving the accuracy of relay armature air gap measurement. The method comprises the following steps: performing energization test on a target relay; based on an action cycle, performing image acquisition on the target relay through an industrial high-speed camera to obtain a target relay corresponding to a to-be-processed image set; performing pretreatment on the to-be-processed image set to obtain a candidate image set; performing identification tracking processing on the candidate image set through a rotation template matching algorithm to generate an armature image set corresponding to the target relay; performing iron core identification on each armature image in the armature image set through a feature identification algorithm to determine iron core coordinate data corresponding to each armature image; and performing armature air gap analysis through the iron core coordinate data corresponding to each armature image to determine an armature air gap data set.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of relays, in particular to a relay armature air gap measurement method based on rotating template matching. BACKGROUND

[0002] In a relay, the armature is an important part connecting the electromagnetic system and the contact spring system of the relay, and the air gap of the armature refers to the distance between the armature and the core, which greatly affects the electromagnetic attraction, so the air gap of the armature will directly affect the electrical characteristics of the relay.

[0003] At present, the measurement of the air gap of the relay armature adopts a manual measurement method: first, fix the armature, and then measure the distance using a vernier caliper, but the armature part and the placement of the vernier caliper cannot be relatively horizontal and fixed, and are affected by human subjectivity, resulting in a large difference between the measurement result and the actual value. The above measurement method is a static measurement, which can only measure the air gap after fixing the position of the armature, and the measured air gap is discontinuous. Therefore, the accuracy of the measurement of the air gap of the relay armature is low at present. SUMMARY

[0004] Therefore, the embodiments of the present application provide a relay armature air gap measurement method based on rotating template matching, which solves the technical problem of low accuracy in measuring the air gap of the relay armature.

[0005] The present application provides a relay armature air gap measurement method based on rotating template matching, comprising: conducting a power-on test on a target relay installed at a preset target placement position; based on a preset action period, image acquisition of the target relay is performed by an industrial high-speed camera to obtain a set of to-be-processed images corresponding to the target relay; the set of to-be-processed images is preprocessed to obtain a set of candidate images;

[0006] The set of candidate images is recognized and tracked by a rotating template matching algorithm to generate a set of armature images corresponding to the target relay; the core coordinates of each armature image are determined by core recognition of each armature image in the set of armature images through a feature recognition algorithm; and the air gap data set is determined by air gap analysis of the core coordinates of each armature image.

[0007] In the present application, before the step of conducting a power-on test on a target relay installed at a preset target placement position, it further comprises: pre-experimenting the target relay through a preset experimental condition, and simultaneously, image acquisition of the target relay is performed during the pre-experimenting process to obtain a set of to-be-analyzed images; the target placement position is determined by placement position analysis of the target relay through the set of to-be-analyzed images; and the target relay is installed at the target placement position.

[0008] In the present application, the pre-processing of the image set to be processed to obtain a candidate image set step includes: performing gray processing on each image to be processed in the image set to be processed to obtain a gray image set; performing binary processing on each gray image in the gray image set to obtain a binary image set; and performing filtering processing on each binary image in the binary image set to obtain a candidate image set.

[0009] In the present application, the step of identifying and tracking the candidate image set by the rotating template matching algorithm to generate the armature image set corresponding to the target relay includes: performing rotating processing on a preset armature template image to obtain a plurality of to-be-matched armature template images at different angles; and identifying and tracking the candidate image set by the rotating template matching algorithm based on the plurality of to-be-matched armature template images at different angles to generate the armature image set corresponding to the target relay.

[0010] In the present application, the step of identifying and tracking the candidate image set by the rotating template matching algorithm based on the plurality of to-be-matched armature template images at different angles to generate the armature image set corresponding to the target relay includes: performing similarity calculation on each candidate image in the candidate image set by the rotating template matching algorithm based on the plurality of to-be-matched armature template images at different angles to obtain a similarity calculation result; and identifying and tracking the candidate image set by the similarity calculation result to generate the armature image set corresponding to the target relay.

[0011] In the present application, the step of identifying the core of each armature image in the armature image set by the feature recognition algorithm to determine the core coordinate data corresponding to each armature image includes: performing edge detection on each armature image in the armature image set by the feature recognition algorithm to determine an edge detection result; performing gradient intensity analysis on each armature image in the armature image set based on the edge detection result to determine a gradient matrix corresponding to each armature image; performing image direction analysis on each armature image in the armature image set to determine an image direction corresponding to each armature image; and identifying the core of each armature image in the armature image set by the gradient matrix corresponding to each armature image and the image direction corresponding to each armature image to determine the core coordinate data corresponding to each armature image.

[0012] In the present application, the iron core recognition of each armature image in the armature image set is performed by the gradient matrix corresponding to each armature image and the image direction corresponding to each armature image, the iron core coordinate data corresponding to each armature image is determined, including: performing image point position traversal processing on the gradient matrix corresponding to each armature image to determine the image point position traversal result corresponding to each armature image; performing fuzzy boundary removal processing on each armature image respectively through the image point position traversal result corresponding to each armature image to obtain a target armature image set; performing iron core recognition on each armature image in the target armature image set through a double threshold algorithm to determine the iron core coordinate data corresponding to each armature image

[0013] In the present application, the target relay installed at the preset target placement position is tested by energization; based on a preset action period, an industrial high-speed camera is used to collect images of the target relay to obtain a set of to-be-processed images corresponding to the target relay; the set of to-be-processed images is preprocessed to obtain a set of candidate images; a rotation template matching algorithm is used to identify and track the set of candidate images to generate a set of armature images corresponding to the target relay; a feature recognition algorithm is used to identify the iron core of each armature image in the set of armature images to determine the iron core coordinate data corresponding to each armature image; and the iron core coordinate data corresponding to each armature image is used to analyze the air gap of the armature to determine a set of air gap data. In the present application, the image of a relay during one action process is collected, the rotation template matching algorithm is used to identify and track the armature of the relay to obtain air gap change data. The present application uses the advantages of fast processing speed, high accuracy and non-contact of the image processing technology, and fills the gap in the measurement method of the parameter by using the dynamic and static measurement method of the air gap of the armature of the relay, so that the parameter value can be quickly and accurately obtained. BRIEF DESCRIPTION OF DRAWINGS

[0014] In order to more clearly illustrate the specific embodiments of the present application or the technical solutions in the prior art, the following will briefly introduce the drawings needed to be used in the specific embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0015] Figure 1 The flow chart of the relay armature air gap measurement method based on the rotation template matching in the present application embodiment.

[0016] Figure 2 The flow chart of the iron core recognition of each armature image in the set of armature images by the feature recognition algorithm in the present application embodiment. DETAILED DESCRIPTION

[0017] The technical solutions of the present application will be described clearly and completely in combination with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0018] In the description of the present application, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer" and the like indicate the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present application and simplifying the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation of the present application. In addition, the terms "first", "second", "third" are only for the purpose of description, and cannot be understood as indicating or implying relative importance.

[0019] In addition, the technical features involved in the different embodiments of the present application described below can be combined with each other as long as there is no conflict.

[0020] For the convenience of understanding, the specific flow of the embodiments of the present application will be described below. Please refer to Figure 1 , Figure 1 is a flowchart of the relay armature air gap measurement method based on rotating template matching of the embodiments of the present application, as Figure 1 shown, the flowchart includes the following steps:

[0021] S101, energizing test is performed on the target relay installed at the preset target placement position;

[0022] S102, based on the preset action period, the target relay is image collected by an industrial high-speed camera to obtain a set of target relay corresponding images to be processed;

[0023] Specifically, first, the target relay is installed on the target relay experimental base, and the industrial high-speed camera is placed opposite the target relay experimental base to collect images on the side of the target relay. It should be noted that the annular light source is installed on the same side as the industrial high-speed camera, which can effectively save space, reduce installation cost, and customize the illumination angle and color of the light source to meet the requirements of low-angle lighting or high-angle lighting. At the same time, the lens is installed on the camera to solve the parallax and distortion caused by the lens. The present application selects a telecentric lens due to its unique design, and the collected light path has almost no angle change, which maximizes the restoration of the true situation of the object. Thus, the measurement accuracy is effectively improved, and then the power connected to the target relay experimental base device is turned on, the target relay is adjusted to have a motion cycle of 3s, the exposure rate and depth of field of the industrial high-speed camera are adjusted to make the target relay armature part and the background area have obvious contrast, and the shooting frequency is set to 2000 frames per second. At the same time, the industrial high-speed camera and the target relay motion signal are given, and the entire motion process of the target relay is completely photographed. When the camera parameters and the position of the target relay are fixed, the pixel value and the actual value parameter conversion are realized by shooting the optical scale, the circular optical scale is automatically recognized to obtain the pixel value of the radius R, and the actual radius of the optical scale is known, so the actual distance value represented by a single pixel point can be calculated. When the target relay completes a motion cycle, the shooting process is completed, and the corresponding target relay image set is obtained.

[0024] S103, pre-processing the target image set to obtain a candidate image set;

[0025] It should be noted that in order to reduce the calculation amount and improve the processing speed of the image, the collected relay motion image is grayed, the appropriate threshold value is selected for the binary threshold processing of the image, the image is filtered before recognition and tracking, that is, most of the noise is removed on the basis of retaining the outline and information of the image as much as possible, and finally the candidate image set is obtained.

[0026] S104, identifying and tracking the candidate image set by a rotating template matching algorithm to generate a target relay corresponding armature image set;

[0027] It should be noted that compared with the method of rotating the image to be searched, the size of the template image to be rotated is much smaller, and the template image to be rotated can be processed in the background without affecting the matching speed, and the interference noise on the template image is much less than that in the image to be searched, so the method of rotating the template image is selected. The matching principle is: the existing template image is rotated to obtain the template image under different angles. During matching, the template image under different angles is used to identify and track the candidate image set through the rotating template matching algorithm, to generate the armature image set corresponding to the target relay, the similarity measurement results corresponding to different angles can be calculated, the best template image is determined, and the identification area is finally determined. It should be noted that if the image is rotated at the center, the original coordinate system needs to be translated to the center point of the image.

[0028] S105, core identification is performed on each armature image in the armature image set through a feature recognition algorithm, to determine the core coordinate data corresponding to each armature image;

[0029] Specifically, a straight line feature recognition method is used to perform edge detection on each armature image in the armature image set, and then the gradient intensity and direction of each armature image in the armature image set are calculated. and direction The calculation formula of the gradient intensity is:

[0030]

[0031] wherein, is the gradient intensity, is the component of the gradient intensity in the x-axis, is the component of the gradient intensity in the y-axis;

[0032] The calculation formula of the direction is:

[0033]

[0034] Specifically, all points on the gradient intensity matrix are traversed, and the pixel points with maximum values in the edge direction are retained, and other values are deleted, so that the fuzzy boundary becomes clear. Finally, a double-threshold method is used to determine the boundary, the image area is subjected to straight line feature recognition and detection, the straight line of the core edge is obtained, the intersection points of the obtained straight line and its extension line are determined, and whether the four intersection points are rectangles is determined in sequence, the threshold is set to remove the rectangles that do not meet the requirements, and then the core coordinate data corresponding to each armature image is determined.

[0035] S106, armature air gap analysis is performed on the core coordinate data corresponding to each armature image, to determine the armature air gap data set.

[0036] ​Specifically, the iron core coordinate data corresponding to each armature image is used to analyze the armature air gap, the distance between the edge of the iron core center and the armature edge in the horizontal direction, i.e., the armature air gap, is calculated, and the data of the change of the armature air gap in the movement process is obtained, and the law of the change of the length of the armature air gap with time is drawn, so that the dynamic measurement of the armature air gap of the target relay is realized.

[0037] By performing the above steps, the target relay installed at the preset target placement position is subjected to energization test; based on the preset action period, the target relay is subjected to image acquisition by an industrial high-speed camera to obtain a set of to-be-processed images corresponding to the target relay; the set of to-be-processed images is subjected to preprocessing to obtain a set of candidate images; the set of candidate images is subjected to recognition and tracking processing by a rotation template matching algorithm to generate a set of armature images corresponding to the target relay; the iron core of each armature image in the set of armature images is identified by a feature recognition algorithm to determine the iron core coordinate data corresponding to each armature image; and the iron core coordinate data corresponding to each armature image is used to analyze the armature air gap to determine a set of armature air gap data. In the embodiment of the present application, the image acquisition of one action process of the relay is performed, the relay armature is recognized and tracked by using the rotation template matching algorithm, and the air gap change data is obtained. The present application uses the image processing technology which has the advantages of fast processing speed, high accuracy and non-contact, and fills the gap of the current measurement method of the parameter by using the dynamic and static measurement method of the relay armature air gap, so that the parameter value can be quickly and accurately obtained.

[0038] In a specific embodiment, before step S101 is performed, the following steps can also be specifically included:

[0039] (1) The target relay is subjected to pre-experiment processing under preset experimental conditions, and at the same time, the target relay is subjected to image acquisition during the pre-experiment processing to obtain a set of to-be-analyzed images;

[0040] (2) The target placement position is determined by analyzing the placement position of the target relay based on the set of to-be-analyzed images;

[0041] (3) The target relay is installed at the target placement position.

[0042] Under the preset experimental conditions, the target relay is subjected to pre-experiment processing. This can include applying voltage or current to test the working state or other characteristics of the relay. A set of to-be-analyzed images is recorded by an image acquisition device (such as a camera), which will be used in subsequent steps. The target placement position is determined by analyzing the placement position of the target relay based on the set of to-be-analyzed images. Specifically, the server detects and analyzes the object in the image by using computer vision technology, and determines the position and pose thereof. According to the determined target placement position, the target relay is installed at the target placement position.

[0043] In a specific embodiment, the process of performing step S103 can specifically include the following steps:

[0044] (1) performing gray-scale processing on each image in the image set to be processed to obtain a gray-scale image set;

[0045] (2) performing binaryzation processing on each gray-scale image in the gray-scale image set to obtain a binaryzation image set;

[0046] (3) performing filtering processing on each binaryzation image in the binaryzation image set to obtain a candidate image set.

[0047] It should be noted that the principle of binaryzation processing is to select a threshold value, and compare the threshold value with the gray-scale value of each pixel point of the image. If the gray-scale value of the pixel point is not less than the selected threshold value, the gray-scale value of the pixel point is set to 255, and the pixel point is determined as a target region. If the gray-scale value of the pixel point is less than the selected threshold value, the gray-scale value of the pixel point is set to 0, and the pixel point is determined as a background region. The principle of filtering is to select the median value of the pixel values near the pixel point to be processed to replace the pixel, and the gray-scale value of the pixel is close to the surrounding pixels, so that independent noise points can be removed. The median filtering can eliminate noise while protecting the details of the image as much as possible. In the embodiment of the present application, the server performs gray-scale processing on each image in the image set to be processed, converts the RGB color value of each pixel to a gray-scale value, and generates a gray-scale image set. The server performs binaryzation processing on each gray-scale image in the gray-scale image set, converts the gray-scale value to black and white binary, and generates a binaryzation image set. This helps to more easily detect and analyze the target object in the subsequent steps. The server performs filtering processing on each binaryzation image in the binaryzation image set, removes noise points using various filtering algorithms (such as Gaussian filtering), and smooths the image edges, thereby obtaining a candidate image set.

[0048] In a specific embodiment, the process of performing step S104 can specifically include the following steps:

[0049] (1) performing rotation processing on the preset armature template image to obtain a plurality of armature template images to be matched at different angles;

[0050] (2) based on the plurality of armature template images to be matched at different angles, performing recognition and tracking processing on the candidate image set by using a rotation template matching algorithm to generate an armature image set corresponding to the target relay.

[0051] Specifically, the armature template image is rotated at a step of 0.1° to obtain an initial set of template images, and the center point of the armature template image is respectively located in the to-be-matched image. The armature template image is sequentially moved in the full image from left to right and from top to bottom, and the similarity at each pixel point is calculated each time. The similarity result corresponding to each pixel point is saved in a matrix, and the point with the highest similarity in the matrix is found, which is the matching position and matching angle. Further, the server rotates the preset armature template image to obtain a plurality of to-be-matched armature template images at different angles. The plurality of to-be-matched armature template images at different angles are identified and tracked by the rotating template matching algorithm to generate an armature image set corresponding to the target relay.

[0052] It should be noted that when the step is selected as 0.2° and 0.5°, the matching accuracy is poor, and it is difficult to meet the requirements of image recognition and tracking. When the step is selected as 0.05°, the time required for matching will greatly increase, and it is difficult to meet the rapidity in image recognition and tracking. Therefore, 0.1° is selected as the step of rotating template matching in consideration of the matching accuracy, which can reach more than 99%. After the image acquisition system hardware is installed in a suitable position, the power connected to the relay experimental base device is turned on, the time relay is adjusted to have an action period of 3s, the exposure rate and depth of field parameters of the industrial high-speed camera are adjusted, the armature part of the relay has a clear contrast with the background area, and the shooting frequency is set to 2000 frames per second. At the same time, the industrial high-speed camera and the relay action signal are given, and the entire movement process of the relay is completely photographed.

[0053] In a specific embodiment, the process of performing the step of generating an armature image set corresponding to the target relay based on a plurality of to-be-matched armature template images at different angles through the rotating template matching algorithm to identify and track the candidate image set is specifically as follows:

[0054] (1) Based on a plurality of to-be-matched armature template images at different angles, the similarity of each candidate image in the candidate image set is calculated through the rotating template matching algorithm to obtain a similarity calculation result.

[0055] (2) The candidate image set is identified and tracked through the similarity calculation result to generate an armature image set corresponding to the target relay.

[0056] Specifically, the server calculates the similarity of multiple magnet armature template images at different angles through the rotation template matching algorithm. This algorithm matches each candidate image in the candidate image set with all template images and calculates the similarity scores between them. Through the rotation template matching algorithm, the server obtains the similarity scores of each candidate image in the candidate image set with all template images. These scores can be used to evaluate the matching degree of each candidate image with the template image to determine which images may contain the target relay. The server uses the similarity calculation results to perform recognition and tracking processing on the candidate image set. Specifically, the server sorts the candidate images according to the similarity scores from high to low and checks each image one by one to determine whether the target relay exists in it. If the target relay is found, the corresponding magnet armature image set is generated. After finding the target relay, the server generates the corresponding magnet armature image set.

[0057] In a specific embodiment, as shown in FIG. 1, the process of step S105 can specifically include the following steps: Figure 2

[0058] S201, performing edge detection on each magnet armature image in the magnet armature image set through a feature recognition algorithm to determine an edge detection result;

[0059] S202, performing gradient intensity analysis on each magnet armature image in the magnet armature image set based on the edge detection result to determine a gradient matrix corresponding to each magnet armature image;

[0060] S203, performing image direction analysis on each magnet armature image in the magnet armature image set to determine an image direction corresponding to each magnet armature image;

[0061] S204, performing core recognition on each magnet armature image in the magnet armature image set through the gradient matrix corresponding to each magnet armature image and the image direction corresponding to each magnet armature image to determine core coordinate data corresponding to each magnet armature image.

[0062] ​Specifically, the server performs edge detection on each armature image in the armature image set using a feature recognition algorithm. This can typically be achieved by using an edge detection algorithm such as Canny edge detection to obtain the edge detection result of the object contour in the image. Based on the edge detection result, the server performs gradient intensity analysis on each armature image in the armature image set to determine the gradient matrix corresponding to each image. This can typically be achieved using a gradient calculation method such as Sobel operator to calculate the gradient amplitude and direction of each pixel point in the image. The server performs image direction analysis on each armature image in the armature image set to determine the main image direction corresponding to each image. This can typically be achieved using techniques such as Hough transform to detect and determine the straight lines in the image, thereby finding the main direction of the image. The server performs core recognition according to the gradient matrix corresponding to each armature image and the main direction corresponding to each image. By analyzing the gradient matrix and the local direction, the core position in each armature image is detected and determined, and the corresponding core coordinate data is generated.

[0063] In a specific embodiment, the process of performing step S204 can specifically include the following steps:

[0064] (1) performing image point position traversal processing on the gradient matrix corresponding to each armature image to determine the image point position traversal result corresponding to each armature image;

[0065] (2) performing fuzzy boundary removal processing on each armature image respectively through the image point position traversal result corresponding to each armature image to obtain a target armature image set;

[0066] (3) performing core recognition on each armature image in the target armature image set through a double threshold algorithm to determine the core coordinate data corresponding to each armature image.

[0067] It should be noted that the server performs image point position traversal processing on the gradient matrix corresponding to each armature image to determine the image point position traversal result corresponding to each image. This process can usually be processed using a pixel-by-pixel manner to analyze the neighborhood information around each pixel, and based on the image point position traversal result corresponding to each armature image, the server performs fuzzy boundary removal processing to obtain the target armature image set. This usually includes using various image processing techniques (such as Gaussian filtering, morphological operations, etc.) to denoise and smooth the image to eliminate areas with weak gradients and unclear edges. This helps to improve the accuracy and reliability of the core recognition in the subsequent steps, and through a double threshold algorithm, the core of each armature image in the target armature image set is recognized to determine the core coordinate data corresponding to each image. The algorithm usually involves setting two thresholds to binarize the image, and using connected component analysis and other techniques to detect and segment different regions in the image. Finally, the core position in each image is determined according to the shape and size of the core and other characteristics, and the corresponding core coordinate data is generated.

[0068] The above embodiments are only used to illustrate the technical solutions of the present application, not to limit it. Although the present application has been described in detail with reference to the embodiments, it should be understood by those skilled in the art that the specific embodiments of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application. Any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the scope of the claims of the present application.

Claims

1. A method for measuring the air gap of a relay armature based on rotating template matching, characterized in that the method The method comprises the following steps: carrying out energization test on a target relay installed at a preset target placement position; based on a preset action cycle, carrying out image acquisition on the target relay by an industrial high-speed camera to obtain a set of to-be-processed images corresponding to the target relay; preprocessing the set of to-be-processed images to obtain a set of candidate images; carrying out recognition and tracking processing on the set of candidate images by a rotation template matching algorithm to generate a set of armature images corresponding to the target relay; carrying out rotation processing on a preset armature template image to obtain a plurality of to-be-matched armature template images at different angles; based on the plurality of to-be-matched armature template images at different angles, carrying out recognition and tracking processing on the set of candidate images by the rotation template matching algorithm to generate the set of armature images corresponding to the target relay; carrying out core recognition on each armature image in the set of armature images by a feature recognition algorithm to determine core coordinate data corresponding to each armature image; carrying out armature air gap analysis by using the core coordinate data corresponding to each armature image to determine a set of armature air gap data.

2. The rotating template matching based relay armature air gap measurement method according to claim 1, characterized in that, Before the step of carrying out energization test on a target relay installed at a preset target placement position, the method further comprises the following steps: carrying out pre-experiment processing on the target relay by using preset experimental conditions, and simultaneously, carrying out image acquisition on the target relay during the pre-experiment processing to obtain a set of to-be-analyzed images; carrying out placement position analysis on the target relay by using the set of to-be-analyzed images to determine a target placement position; installing the target relay to the target placement position.

3. The rotating template matching based relay armature air gap measurement method of claim 1, wherein, The step of preprocessing the set of to-be-processed images to obtain a set of candidate images comprises the following steps: carrying out grayscale processing on each to-be-processed image in the set of to-be-processed images to obtain a set of grayscale images; carrying out binarization processing on each grayscale image in the set of grayscale images to obtain a set of binarized images; carrying out filtering processing on each binarized image in the set of binarized images to obtain a set of candidate images.

4. The rotating template matching based relay armature air gap measurement method of claim 1, wherein, The step of carrying out recognition and tracking processing on the set of candidate images by the rotation template matching algorithm to generate the set of armature images corresponding to the target relay based on the plurality of to-be-matched armature template images at different angles comprises the following steps: based on the plurality of to-be-matched armature template images at different angles, carrying out similarity calculation on each candidate image in the set of candidate images by the rotation template matching algorithm to obtain a similarity calculation result; carrying out recognition and tracking processing on the set of candidate images by using the similarity calculation result to generate the set of armature images corresponding to the target relay.

5. The rotating template matching based relay armature air gap measurement method according to claim 1, wherein, The step of carrying out core recognition on each armature image in the set of armature images by a feature recognition algorithm to determine core coordinate data corresponding to each armature image comprises the following steps: carrying out edge detection on each armature image in the set of armature images by the feature recognition algorithm to determine an edge detection result; based on the edge detection result, carrying out gradient intensity analysis on each armature image in the set of armature images to determine a gradient matrix corresponding to each armature image; The image direction of each armature image in the armature image set is analyzed to determine the image direction corresponding to each armature image; The gradient matrix corresponding to each armature image and the image direction corresponding to each armature image are used to identify the core of each armature image in the armature image set to determine the core coordinate data corresponding to each armature image.

6. The rotating mask pattern based relay armature air gap measurement method according to claim 5, wherein, The step of identifying the core of each armature image in the armature image set by using the gradient matrix corresponding to each armature image and the image direction corresponding to each armature image to determine the core coordinate data corresponding to each armature image comprises: The gradient matrix corresponding to each armature image is subjected to image point position traversal processing to determine the image point position traversal result corresponding to each armature image; Each armature image is subjected to fuzzy boundary removal processing according to the image point position traversal result corresponding to each armature image to obtain a target armature image set; A double threshold algorithm is used to identify the core of each armature image in the target armature image set to determine the core coordinate data corresponding to each armature image.

Images