Disconnector posture determination method, device and storage medium

Abstract

The present application discloses a method, device and storage medium for determining the posture of an isolating switch. The method comprises: obtaining a target isolating switch image of a target isolating switch whose posture is to be determined; determining a target mark point and a first feature vector of the target isolating switch image; performing mark point pairing on the target mark point with an opening mark point of a target opening posture image based on the first feature vector to obtain a first pairing point number, and performing mark point pairing on the target mark point with a closing mark point of a target closing posture image based on the first feature vector to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points at which the target mark point is successfully paired with the closing mark point; determining the target posture of the target isolating switch based on the first pairing point number and the second pairing point number. The present application solves the technical problems of slow operation speed and poor recognition accuracy in the related art.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method, device and storage medium for determining the posture of an isolating switch. Background Art

[0002] An isolating switch is a switch used to completely isolate electrical equipment or circuits from power supplies. It can cut off the connection between electrical equipment and power supplies to ensure safe operation of equipment during maintenance, overhaul or emergency situations. Isolating switches are commonly used in industrial and power systems to ensure the safety of workers and the normal operation of equipment. Among them, the scissor-type isolating switch is an isolating switch with a scissor-like appearance. This type of isolating switch is flexible and convenient to operate and is widely used in 220kV substations. By optimizing the size of the scissor-type conductive arm, the reliability and stability of the operation can be ensured. Its biggest advantage is that there is no need to deactivate the busbar when repairing the isolating switch, which is conducive to the safe and reliable operation of the power grid.

[0003] During the power grid operation, the scissor-type disconnector needs to be opened and closed in order to proceed to the next step. Therefore, it is necessary to judge the posture of the scissor-type disconnector. This process requires manual work at the power grid operation area to judge the current posture of the scissor-type disconnector and whether the opening and closing operations are performed accurately, which has the disadvantages of low efficiency and low reliability.

[0004] Based on this technical problem, we can use machine vision technology related to deep learning to automatically judge the posture of the scissor-type disconnector through photos of the power grid operation area where the scissor-type disconnector is located, which can more efficiently identify the current posture of the scissor-type disconnector and improve processing efficiency and accuracy. However, the current state detection of the disconnector is slow: some methods use complex network training or high vector dimensions, resulting in high running time costs; low accuracy: some methods run faster, but the final accuracy is not ideal.

[0005] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0006] The embodiments of the present application provide a method, device and storage medium for determining the posture of an isolating switch, so as to at least solve the technical problems of slow operation speed and poor recognition accuracy in the related art.

[0007] According to one embodiment of the present application, a method for determining the posture of an isolating switch is provided. The method may include: obtaining a target isolating switch image of a target isolating switch whose posture is to be determined; determining a target marker point and a first feature vector of the target isolating switch image, wherein the target marker point is used to record the boundary information of the target isolating switch image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target isolating switch image; based on the first feature vector, the target marker point is paired with the opening marker point of the target opening posture image to obtain a first pairing point number, and based on the first feature vector, the target marker point is paired with the closing marker point of the target closing posture image to obtain a second pairing point number, wherein the first pairing point number is the number of points where the target marker point is successfully paired with the opening marker point, and the second pairing point number is the number of points where the target marker point is successfully paired with the closing marker point; and determining the target posture of the target isolating switch based on the first pairing point number and the second pairing point number.

[0008] Optionally, determining the target landmark point and the first eigenvector of the target isolating switch image includes: converting the target isolating switch image to a target scale space to obtain a first isolating switch image, wherein the target scale space is used to simulate the viewing angle of observing the target isolating switch image; determining the target landmark point of the first isolating switch image; determining the direction of the target landmark point based on the gradient direction of the target landmark point; and determining the first eigenvector based on the target landmark point and the target landmark point direction.

[0009] Optionally, the method further includes: establishing a target scale space based on a preset Gaussian blur scale.

[0010] Optionally, converting the target isolating switch image to a target scale space to obtain a first isolating switch image includes: converting the target isolating switch image to a target scale space to obtain a second isolating switch image; and performing a convolution operation on the second isolating switch image based on a Gaussian function to obtain the first isolating switch image.

[0011] Optionally, determining the target landmark point of the first isolating switch image includes: determining the extreme points of the first isolating switch image in the target scale space, wherein the extreme points are used to represent points where the pixel values ​​of the first isolating switch image are maximum or minimum; and determining the target landmark point based on the extreme points.

[0012] Optionally, determining the direction of the target marker point based on the gradient direction of the target marker point includes: determining a preset neighborhood based on the target marker point, and dividing the preset neighborhood into multiple directions; determining multiple first gradient magnitudes and multiple first gradient directions of the target marker point in multiple directions; determining a histogram based on the multiple first gradient magnitudes and the multiple first gradient directions; and determining the direction of the target marker point based on the peak value of the histogram.

[0013] Optionally, determining the first eigenvector based on the target marker point and the target marker point direction includes: dividing a preset neighborhood into multiple sub-regions, wherein the multiple sub-regions include regions of various shapes; determining a target coordinate system based on the target marker point direction; in the target coordinate system, determining multiple second gradient magnitudes and multiple second gradient directions of the target marker point in the multiple sub-regions; determining the second eigenvector based on the multiple second gradient magnitudes, the multiple second gradient directions, the grayscale value means of the pixels of the multiple sub-regions, and the grayscale value variance of the pixels of the multiple sub-regions; and normalizing the second eigenvector to obtain the first eigenvector.

[0014] Optionally, based on the first eigenvector, the target landmark points are paired with the opening landmark points of the target opening posture image to obtain a first pairing point number, and based on the first eigenvector, the target landmark points are paired with the closing landmark points of the target closing posture image to obtain a second pairing point number, including: determining a third eigenvector and a fourth eigenvector, wherein the third eigenvector is used to describe the shape and structure of the target opening posture image, and the fourth eigenvector is used to describe the shape and structure of the target closed posture image; based on the third eigenvector and the first eigenvector, the target landmark points are screened to obtain a first pairing point; based on the fourth eigenvector and the first eigenvector, the target landmark points are screened to obtain a second pairing point.

[0015] Optionally, screening the target landmark points based on the third eigenvector and the first eigenvector to obtain the first pairing point includes: determining a first Euclidean distance between the first eigenvector and the third eigenvector; screening landmark points in the first Euclidean distance that are less than a preset threshold to obtain the first pairing point.

[0016] Optionally, screening the target landmark points based on the fourth eigenvector and the first eigenvector to obtain the second pairing point includes: determining a second Euclidean distance between the first eigenvector and the fourth eigenvector; and screening landmark points in the second Euclidean distance that are less than a preset threshold to obtain the second pairing point.

[0017] Optionally, the method further includes: in response to the presence of a first special marker point that meets a first special filtering condition among the target marker points, determining a first degree of association between first special feature vectors corresponding to the first special marker point, wherein the first special filtering condition includes points paired with multiple open marker points and points paired with special points among the open marker points, and the special points among the open marker points are points paired with multiple target marker points; updating the first pairing point based on the first special marker point; or, in response to the presence of a second special marker point that meets a second special filtering condition among the target marker points, determining a second degree of association between second special feature vectors corresponding to the second special marker point, wherein the second special filtering condition includes points paired with multiple closed marker points and points paired with special points among the closed marker points, and the special points among the closed marker points are points paired with multiple target marker points; updating the second pairing point based on the second special marker point.

[0018] Optionally, determining the target posture of the target isolating switch based on the first pairing points and the second pairing points includes: determining a first pairing rate based on the first pairing points, wherein the first pairing rate is used to indicate the probability that the target isolating switch is in an open posture; determining a second pairing rate based on the second pairing points, wherein the second pairing rate is used to indicate the probability that the target isolating switch is in a closed posture; determining the target posture of the target isolating switch based on the first pairing rate and the second pairing rate.

[0019] Optionally, determining the target posture of the target isolating switch based on the first matching rate and the second matching rate includes: determining a matching rate ratio based on the first matching rate and the second matching rate, wherein the matching rate ratio is the ratio of the first matching rate to the second matching rate; in response to the matching rate ratio being greater than 1, determining that the target posture of the target isolating switch is an open posture; in response to the matching rate ratio being less than or equal to 1, determining that the target posture of the target isolating switch is a closed posture.

[0020] Optionally, the method also includes: determining a target open posture image and a target closed posture image, wherein the model of the isolating switch in the target open posture image and the model of the isolating switch in the target closed posture image are the same as the model of the target isolating switch, and both the target open posture image and the target closed posture image include a main switch and a static contact.

[0021] According to one embodiment of the present application, a device for determining the posture of an isolating switch is also provided. The device includes: an acquisition module, which is used to acquire a target isolating switch image of a target isolating switch whose posture is to be determined; a first determination module, which is used to determine a target mark point and a first feature vector of the target isolating switch image, wherein the target mark point is used to record the boundary information of the target isolating switch image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target isolating switch image; a pairing module, which is used to perform mark point pairing of the target mark point with the opening mark point of the target opening posture image based on the first feature vector to obtain a first pairing point number, and perform mark point pairing of the target mark point with the closing mark point of the target closing posture image based on the first feature vector to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points at which the target mark point is successfully paired with the closing mark point; a second determination module, which is used to determine the target posture of the target isolating switch based on the first pairing point number and the second pairing point number.

[0022] Optionally, the first determination module is also used to convert the target isolating switch image into a target scale space to obtain a first isolating switch image, wherein the target scale space is used to simulate the viewing angle of observing the target isolating switch image; determine the target landmark point of the first isolating switch image; determine the direction of the target landmark point based on the gradient direction of the target landmark point; and determine the first eigenvector based on the target landmark point and the target landmark point direction.

[0023] Optionally, the device further includes: a scale establishment module, configured to establish a target scale space based on a preset Gaussian blur scale.

[0024] Optionally, the first determination module is further used to convert the target isolation switch image into a target scale space to obtain a second isolation switch image; and perform a convolution operation on the second isolation switch image based on a Gaussian function to obtain the first isolation switch image.

[0025] Optionally, the first determination module is further used to determine extreme points of the first isolating switch image in the target scale space, wherein the extreme points are used to represent points where the pixel values ​​of the first isolating switch image are maximum or minimum; and determine the target landmark points based on the extreme points.

[0026] Optionally, the first determination module is also used to determine a preset neighborhood based on the target marker point, and divide the preset neighborhood into multiple directions; determine multiple first gradient magnitudes and multiple first gradient directions of the target marker point in multiple directions; determine a histogram based on the multiple first gradient magnitudes and the multiple first gradient directions; and determine the direction of the target marker point based on the peak value of the histogram.

[0027] Optionally, the first determination module is also used to divide a preset neighborhood into multiple sub-regions, wherein the multiple sub-regions include regions of various shapes; determine a target coordinate system based on the direction of a target marker point; in the target coordinate system, determine multiple second gradient magnitudes and multiple second gradient directions of the target marker point in the multiple sub-regions; determine a second eigenvector based on multiple second gradient magnitudes, multiple second gradient directions, grayscale value means of pixels in multiple sub-regions, and grayscale value variances of pixels in multiple sub-regions; and normalize the second eigenvector to obtain a first eigenvector.

[0028] Optionally, the pairing module is also used to determine a third eigenvector and a fourth eigenvector, wherein the third eigenvector is used to describe the shape and structure of the target open posture image, and the fourth eigenvector is used to describe the shape and structure of the target closed posture image; based on the third eigenvector and the first eigenvector, the target landmark points are screened to obtain a first pairing point; based on the fourth eigenvector and the first eigenvector, the target landmark points are screened to obtain a second pairing point.

[0029] Optionally, the pairing module is further used to determine a first Euclidean distance between the first feature vector and the third feature vector; and filter out landmark points in the first Euclidean distance that are smaller than a preset threshold to obtain first pairing points.

[0030] Optionally, the pairing module is further used to determine a second Euclidean distance between the first eigenvector and the fourth eigenvector; and filter out landmark points in the second Euclidean distance that are less than a preset threshold to obtain second pairing points.

[0031] Optionally, the pairing module is also used to determine a first degree of association between first special feature vectors corresponding to the first special marker point in response to the presence of a first special marker point that meets a first special filtering condition in the target marker points, wherein the first special filtering condition includes points paired with multiple open marker points and points paired with special points in the open marker points, and the special points in the open marker points are points paired with multiple target marker points; update the first pairing point based on the first special marker point; or, in response to the presence of a second special marker point that meets a second special filtering condition in the target marker points, determine a second degree of association between second special feature vectors corresponding to the second special marker point, wherein the second special filtering condition includes points paired with multiple closed marker points and points paired with special points in the closed marker points, and the special points in the closed marker points are points paired with multiple target marker points; update the second pairing point based on the second special marker point.

[0032] Optionally, the second determination module is also used to determine a first matching rate based on the first matching points, wherein the first matching rate is used to indicate the probability that the target isolating switch is in an open posture; determine a second matching rate based on the second matching points, wherein the second matching rate is used to indicate the probability that the target isolating switch is in a closed posture; and determine the target posture of the target isolating switch based on the first matching rate and the second matching rate.

[0033] Optionally, the second determination module is also used to determine the matching rate ratio based on the first matching rate and the second matching rate, wherein the matching rate ratio is the ratio of the first matching rate to the second matching rate; in response to the matching rate ratio being greater than 1, determining that the target posture of the target isolating switch is an open posture; in response to the matching rate ratio being less than or equal to 1, determining that the target posture of the target isolating switch is a closed posture.

[0034] Optionally, the device also includes: a determination module, used to determine a target open posture image and a target closed posture image, wherein the model of the isolating switch in the target open posture image and the model of the isolating switch in the target closed posture image are the same as the model of the target isolating switch, and the target open posture image and the target closed posture image both include a main switch and a static contact.

[0035] According to one embodiment of the present application, a computer-readable storage medium is also provided, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the isolating switch posture determination method in the embodiment of the present application.

[0036] According to one embodiment of the present application, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the isolating switch posture determination method in the embodiment of the present application is implemented.

[0037] According to one embodiment of the present application, an electronic device is also provided, including a memory and a processor, wherein an executable program is stored in the memory, and the processor is used to run the program and is configured to run a computer program to execute the isolating switch posture determination method in the embodiment of the present application.

[0038] In the embodiment of the present application, by obtaining the target disconnector image of the target disconnector whose posture is to be determined; determining the target mark point and the first feature vector of the target disconnector image, wherein the target mark point is used to record the boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image; based on the first feature vector, the target mark point is paired with the opening mark point of the target opening posture image to obtain the first pairing point number, and based on the first feature vector, the target mark point is paired with the closing mark point of the target closing posture image to obtain the second pairing point number, wherein the first pairing point number is the number of points where the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points where the target mark point is successfully paired with the closing mark point; the target posture of the target disconnector is determined based on the first pairing point number and the second pairing point number. Thus, the purpose of automatically judging the posture of the scissor-type disconnector based on the image of the disconnector in the power grid operation area is achieved, and the technical effect of improving the accuracy of automatically judging the posture of the disconnector is achieved, thereby solving the technical problems of slow running speed and poor recognition accuracy in the related technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0040] Figure 1 is a flow chart of a method for determining the posture of an isolating switch according to an embodiment of the present application;

[0041] Figure 2 is a schematic diagram of a scissor-type isolating switch structure according to an embodiment of the present application;

[0042] Figure 3 is a schematic diagram of a posture of a scissor-type isolating switch according to an embodiment of the present application;

[0043] Figure 4 is a flow chart of a disconnector posture determination scheme according to an embodiment of the present application;

[0044] Figure 5 is a simplified flow chart of a disconnector posture determination solution according to an embodiment of the present application;

[0045] Figure 6 is a schematic diagram for determining an extreme point according to an embodiment of the present application;

[0046] Figure 7 is a schematic diagram of determining the direction of a marker point according to an embodiment of the present application;

[0047] Figure 8 is a schematic diagram of sub-region division according to an embodiment of the present application;

[0048] Fig. 9 It is a structural block diagram of a device for determining the posture of an isolating switch according to an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0050] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0051] Example 1

[0052] According to an embodiment of the present application, a method for determining the posture of an isolating switch is first provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0053] Figure 1 According to a flow chart of a method for determining the posture of an isolating switch according to an embodiment of the present application, Figure 1 As shown, the method at least includes steps S12-S18, wherein:

[0054] Step S12, obtaining a target disconnector image of a target disconnector whose posture is to be determined.

[0055] The target isolating switch is an isolating switch whose posture is to be determined. In one embodiment of the present application, the isolating switch may be a scissor-type isolating switch. The scissor-type isolating switch is used as an example for the following description. The structure is as follows: Figure 2 As shown, the scissor-type disconnector includes a rotary insulator, a base, a motor operating mechanism, a manual operating mechanism, a support insulator, a main switch and a static contact (tube busbar or soft busbar). Figure 3 is a schematic diagram of a scissor-type isolating switch posture according to an embodiment of the present application, such as Figure 3 As shown, the postures of the scissor-type disconnector include an open posture and a closed posture.

[0056] Exemplarily, to determine the posture of the disconnector, it is necessary to first obtain grayscale images of the disconnector in open and closed postures taken on-site in power grid operation areas such as high-voltage substations, as the disconnector image of the posture to be determined, that is, the target disconnector image.

[0057] Step S14, determining the target landmark point and the first feature vector of the target disconnector image, wherein the target landmark point is used to record the boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image.

[0058] Target landmarks can be understood as points on the target disconnector image used to identify or locate specific objects or locations, and are usually used for feature detection and matching in computer vision and image processing. These landmarks can be significant features in the image, such as corners, boundaries, and color blocks.

[0059] The first eigenvector can be understood as a mathematical vector that describes the content and features of an image, and is usually used to represent features such as texture, color, shape, and structure in an image. Feature vectors can be extracted through various image processing and computer vision techniques, such as edge detection, color histograms, shape descriptors, etc. Among them, the first eigenvector is a multidimensional vector used to represent features or attributes in a multidimensional space. In a multidimensional space, an object or data can be described by multiple features or attributes, so the feature vector is usually multidimensional, with each dimension corresponding to a feature or attribute. For example, for a data set containing multiple features, the feature vector of each data point can be represented as a multidimensional vector, where each dimension corresponds to a feature.

[0060] Exemplarily, after the target disconnector image is acquired, target landmark points of the target disconnector image for recording boundary information of the target disconnector image and feature vectors for describing the shape and structure of the target disconnector image are determined.

[0061] Step S16, based on the first feature vector, the target landmark point is paired with the opening landmark point of the target opening posture image to obtain a first pairing point number, and based on the first feature vector, the target landmark point is paired with the closing landmark point of the target closing posture image to obtain a second pairing point number, wherein the first pairing point number is the number of points for successful pairing of the target landmark point with the opening landmark point, and the second pairing point number is the number of points for successful pairing of the target landmark point with the closing landmark point.

[0062] The target open posture image may be a pattern image of an isolating switch in an open posture, which is used as a subsequent matching pattern and is recorded as S0. The open mark point is a mark point determined on the target open posture image, which is used to indicate that the isolating switch is in an open posture. The target closed posture image may be a pattern image of an isolating switch in a closed posture, which is used as a subsequent matching pattern and is recorded as S1. The closed mark point is a mark point determined on the target closed posture image, which is used to indicate that the isolating switch is in a closed posture.

[0063] Exemplarily, after determining the target mark point and the first feature vector of the isolating switch image to be tested, the target mark point and the opening mark point of the target opening posture image are paired based on the first feature vector to obtain a first pairing point number, which is used to indicate the matching degree between the target mark point and the opening mark point of the target opening posture image, that is, to reflect the matching degree between the isolating switch image of the posture to be determined and the opening posture. At the same time, the target mark point and the closing mark point of the target closing posture image are paired based on the first feature vector to obtain a second pairing point number, which is used to indicate the matching degree between the target mark point and the closing mark point of the target closing posture image, that is, to reflect the matching degree between the isolating switch image of the posture to be determined and the closing posture.

[0064] Step S18 determines a target posture of the target disconnect switch based on the first pairing point number and the second pairing point number.

[0065] Exemplarily, the posture of the isolating switch to be determined is determined based on the first pairing point number and the second pairing point number, that is, whether it is in an open posture or a closed posture.

[0066] Based on the above steps, by obtaining the target disconnector image of the target disconnector whose posture is to be determined; determining the target mark point and the first feature vector of the target disconnector image, wherein the target mark point is used to record the boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image; based on the first feature vector, the target mark point is paired with the opening mark point of the target opening posture image to obtain the first pairing point number, and based on the first feature vector, the target mark point is paired with the closing mark point of the target closing posture image to obtain the second pairing point number, wherein the first pairing point number is the number of points where the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points where the target mark point is successfully paired with the closing mark point; the target posture of the target disconnector is determined based on the first pairing point number and the second pairing point number. Thus, the purpose of automatically judging the posture of the scissor-type disconnector based on the image of the disconnector in the power grid operation area is achieved, and the technical effect of improving the accuracy of automatically judging the posture of the disconnector is achieved, thereby solving the technical problems of slow running speed and poor recognition accuracy in the related technology.

[0067] Optionally, in step S14, determining the target landmark point and the first feature vector of the target disconnector image includes performing the following steps:

[0068] Step S141, converting the target disconnector image into a target scale space to obtain a first disconnector image, wherein the target scale space is used to simulate a viewing angle for observing the target disconnector image;

[0069] Step S142, determining a target marking point of the first disconnector image;

[0070] Step S143, determining the direction of the target marker point based on the gradient direction of the target marker point;

[0071] Step S144, determining a first feature vector based on the target landmark point and the direction of the target landmark point.

[0072] The target scale space is a space used to measure and compare the scales or sizes of different objects, concepts or phenomena in a specific field, and is used to simulate the perspective of observing the target disconnector image, that is, to simulate the human perspective.

[0073] Exemplarily, the target disconnector image is converted into the target scale space to obtain the first disconnector image, which can be understood as the converted image of the target disconnector image in the scale space. Then, the target marker point on the image is determined based on the first disconnector image to represent the boundary information of the first disconnector image. The direction of the target marker point is determined based on the gradient direction of the target marker point (the direction in which the grayscale of the image changes fastest, usually the gradient direction is perpendicular to the edge of the image, that is, the direction of the edge). Finally, a first eigenvector is generated based on the target marker point and the direction of the target marker point to describe the shape and structure of the target disconnector image.

[0074] Optionally, step 14 further includes: establishing a target scale space based on a preset Gaussian blur scale.

[0075] Among them, the preset Gaussian blur is a commonly used image processing technique that is preset to reduce noise and details in an image, making the image smoother and more blurred. This blur effect is achieved by applying a Gaussian function, which creates a weight around each pixel of the image, and then uses these weights to recalculate the pixel value, blurring the noise and details in the image, thereby improving the quality and aesthetics of the image.

[0076] For example, the target scale space is established based on the preset Gaussian blur scale. The blur degree of the image is changed by changing the scale parameter in the Gaussian filter, and then the images with different blur degrees are combined in a certain order to form the scale space. The key points detected in the scale space are stable to the scale transformation and can be used as the essential features of the image in images of different scales.

[0077] Optionally, in step S141, converting the target disconnector image to the target scale space to obtain the first disconnector image includes performing the following steps:

[0078] Step S1411, converting the target disconnector image into a target scale space to obtain a second disconnector image;

[0079] Step S1412: performing a convolution operation on the second isolating switch image based on a Gaussian function to obtain a first isolating switch image.

[0080] Exemplarily, the isolating switch image whose posture is to be determined is converted to the target scale space based on a preset Gaussian fuzzy scale to obtain a second isolating switch image, represented as I(x, y). Then, the second isolating switch image is convolved using the Gaussian function G(x, y, σ) to obtain a first isolating switch image, represented as L(x, y, σ). The Gaussian function is:

[0081]

[0082] In the formula, (x, y) is the coordinate of the pixel point of the scissor-type disconnector image; σ is the Gaussian blur scale, also known as the spatial scale factor, which is a variable. As σ increases, the image becomes more and more blurred. For σ, it is necessary to repeatedly approximate a more appropriate value during the experiment. The value of σ depends on the actual situation and is set to 1.6 in this application.

[0083] The convolution operation is expressed as:

[0084] L(x,y,σ)=G(x,y,σ)*I(x,y) Formula (2)

[0085] Among them, * is the convolution operation of Gaussian kernel function and image.

[0086] Optionally, in step S142, determining the target landmark of the first disconnector image includes performing the following steps:

[0087] Step S1421, determining an extreme point of the first isolating switch image in the target scale space, wherein the extreme point is used to represent a point where a pixel value of the first isolating switch image is a maximum value or a minimum value;

[0088] Step S14222, determine the target landmark point based on the extreme point.

[0089] Exemplarily, the new scale space is calculated:

[0090] F(x,y,σ)=(G(x,y,kσ)-G(x,y,σ))*I(x,y)=L(x,y,kσ)-L(x,y,σ) Formula (3)

[0091] In the formula, k is a multiplier factor used to change the scale of adjacent scales, and a new scale space can be obtained. In the new scale space, the point with the maximum or minimum value of the pixel point in the first isolating switch image is selected as the extreme point of the scissor isolating switch image, expressed as X = (x, y, σ). X = (x, y, σ) represents the point (x, y) under the Gaussian blur scale σ. Since the number of pixels in the image before and after Gaussian blur remains unchanged, (x, y) is the same pixel at any scale.

[0092] Then, the target landmark point is determined based on the extreme point, and Taylor expansion is performed on the new scale space F(x, y, σ), and the following is obtained:

[0093]

[0094] The landmark point of the scissor-type disconnector image is X0 corresponding to when the derivative value of F(X) is 0, which is the target landmark point.

[0095]

[0096] Optionally, in step S143, determining the direction of the target marker point based on the gradient direction of the target marker point includes performing the following steps:

[0097] Step S1431, determining a preset neighborhood based on the target landmark point, and dividing the preset neighborhood into multiple directions;

[0098] Step S1432, determining a plurality of first gradient magnitudes and a plurality of first gradient directions of the target marker point in a plurality of directions;

[0099] Step S1433, determining a histogram based on the multiple first gradient magnitudes and the multiple first gradient directions;

[0100] Step S1434, determining the direction of the target landmark point based on the peak value of the histogram.

[0101] Exemplarily, a preset neighborhood is determined based on the target marker point, and the preset neighborhood is divided into multiple directions. The preset neighborhood can be understood as a set of pre-set pixel points adjacent to the target marker point. Usually, the neighborhood is defined as a square or circular area centered on the marker point. The size of the neighborhood can be determined according to specific application requirements. Usually, neighborhoods of sizes such as 3×3, 5×5 or 7×7 are used in image processing.

[0102] Then, the gradient in the neighborhood of the marker point (x, y) in the disconnector image L(x, y, σ) under the Gaussian fuzzy scale σ is determined:

[0103]

[0104]

[0105] Wherein, L(x, y) is the image pixel value of (x, y) in the scissor-type disconnector image under the Gaussian blur scale σ, and L(x, y) is L(x, y, σ). m(x, y) is the gradient magnitude of (x, y), that is, the first gradient magnitude, and θ(x, y) is the gradient direction of (x, y), that is, the first gradient direction. Wherein, the division of the gradient direction depends on the actual situation. In the embodiment of the present application, 45° is used as the division unit, and it is divided into 8 directions.

[0106] After determining multiple first gradient magnitudes and multiple first gradient directions in multiple directions within the target marker area, the gradient is obtained by histogram statistics, and the target marker direction is determined based on the peak value of the histogram. The gradient direction corresponding to the peak value in the histogram statistics is the main direction.

[0107] Optionally, in step S144, determining the first feature vector based on the target landmark point and the target landmark point direction includes performing the following steps:

[0108] Step S1441, dividing the preset neighborhood into a plurality of sub-regions, wherein the plurality of sub-regions include regions of various shapes;

[0109] Step S1442, determining a target coordinate system based on the direction of the target landmark point;

[0110] Step S1443, determining multiple second gradient magnitudes and multiple second gradient directions of the target marker in the multiple sub-regions in the target coordinate system;

[0111] Step S1444, determining a second eigenvector based on the plurality of second gradient magnitudes, the plurality of second gradient directions, the grayscale value means of the pixels of the plurality of sub-regions, and the grayscale value variances of the pixels of the plurality of sub-regions;

[0112] Step S1445, normalize the second eigenvector to obtain the first eigenvector.

[0113] Exemplarily, the preset area is divided into a plurality of sub-areas, wherein the plurality of sub-areas are areas including a plurality of shapes. For example, taking each mark point in the scissor-type isolating switch image as the center, For the pixels in the neighborhood with the pixel as the radius, the 16×16 pixels around the marker point are first divided into 4×4 sub-regions according to the size of 4×4, that is, divided into 6 rectangular regions. Then the remaining pixels are divided into 4 sub-regions, a total of 10 sub-regions.

[0114] Then rotate the coordinate axis to be in the same direction as the main direction, determine the target coordinate system based on the direction of the target landmark, and perform a standardized transformation on the main direction. The corresponding position change relationship is:

[0115]

[0116] In the target coordinate system, for the pixels in all sub-areas, the second gradient magnitude and the second gradient direction are obtained based on the above method, and weighted accumulation is performed to generate a gradient histogram. Then, the second eigenvector is determined based on multiple second gradient magnitudes, multiple second gradient directions, the grayscale value mean of the pixels in multiple sub-areas, and the grayscale value variance of the pixels in multiple sub-areas. Among them, the second eigenvector is used to represent the eigenvalue of the target disconnector image in the target coordinate. For example, each sub-area has 8 directional information, and the grayscale value mean and grayscale value variance of each pixel are two information, then each sub-area has a total of 10 information. Then a single marker point forms a 100-dimensional vector describing the feature. Combining the directional information of all sub-areas to generate a 100-dimensional vector describing the feature is helpful to describe the texture, shape, structure and other features of the image.

[0117] Finally, the second eigenvector is normalized to obtain the first eigenvector. Let the description of the second eigenvector be = (a1, a2, a3, ... a 100 ), then the normalized vector describing the feature is L = (l1, l2, l3, ...l 100 ), the formula is:

[0118]

[0119] Optionally, in step S16, performing landmark point pairing on the target landmark point with an opening landmark point of the target opening posture image based on the first feature vector to obtain a first pairing point number, and performing landmark point pairing on the target landmark point with a closing landmark point of the target closing posture image based on the first feature vector to obtain a second pairing point number includes performing the following steps:

[0120] Step S161, determining a third eigenvector and a fourth eigenvector, wherein the third eigenvector is used to describe the shape and structure of the target open posture image, and the fourth eigenvector is used to describe the shape and structure of the target closed posture image;

[0121] Step S162, screening the target landmark points based on the third eigenvector and the first eigenvector to obtain a first pairing point;

[0122] Step S163: Screen the target landmark points based on the fourth eigenvector and the first eigenvector to obtain second pairing points.

[0123] Exemplarily, pairing the target landmark points with the landmark points of the target posture image based on the first feature vector includes determining a third feature vector and a fourth feature vector, assuming that the vector describing the feature of the scissor-type isolating switch posture style image is X i =(x i1 , x i2 , ..., x i100 ). Wherein, the third eigenvector is used to describe the shape and structure of the target open posture image, and the fourth eigenvector is used to describe the shape and structure of the target closed posture image. Based on the third eigenvector and the first eigenvector in the isolating switch image of the posture to be determined, the feature points in the isolating switch image of the posture to be determined are screened to obtain a first pairing point, wherein the first pairing point is used to indicate a point where the target mark point is successfully paired with the open mark point. Based on the fourth eigenvector and the first eigenvector in the isolating switch image of the posture to be determined, the feature points in the isolating switch image of the posture to be determined are screened to obtain a second pairing point, wherein the second pairing point is used to indicate a point where the target mark point is successfully paired with the closed mark point.

[0124] Optionally, in step S162, screening the target landmark point based on the third feature vector and the first feature vector to obtain the first pairing point includes performing the following steps:

[0125] Step S1621, determining a first Euclidean distance between the first eigenvector and the third eigenvector;

[0126] Step S1622, filter out the landmark points whose distance in the first Euclidean distance is less than a preset threshold, and obtain the first pairing point.

[0127] For example, a vector describing a feature of the scissor-type disconnector image to be paired is represented as Y j =(y j1 ,y j2 , ..., y j100 ), determine the first Euclidean distance between the first eigenvector and the third eigenvector, wherein the first Euclidean distance is used to represent the similarity or difference between the first eigenvector and the third eigenvector, and the calculation formula is as follows:

[0128]

[0129] The preset threshold is a preset standard for measuring the first Euclidean distance, which is used to determine whether the first Euclidean distance meets the expected requirements or conditions. i , Y j ) is less than T (i.e., a preset threshold, usually set to 0.8), and the landmark points that meet the rule are screened out to obtain the first pairing point.

[0130] Optionally, in step S163, screening the target landmark point based on the fourth eigenvector and the first eigenvector to obtain the second pairing point includes performing the following steps:

[0131] Step S1631, determining a second Euclidean distance between the first eigenvector and the fourth eigenvector;

[0132] Step S1632: filter out the landmark points whose second Euclidean distance is less than a preset threshold to obtain the second pairing points.

[0133] For example, a vector describing a feature of the scissor-type disconnector image to be paired is represented as Y j =(y j1 ,y j2 , ..., y j100 ), determine the second Euclidean distance between the first eigenvector and the fourth eigenvector, wherein the first Euclidean distance is used to represent the similarity or difference between the first eigenvector and the third eigenvector, and the calculation is the same as formula (10).

[0134] The preset threshold is a preset standard for measuring the second Euclidean distance, which is used to determine whether the second Euclidean distance meets the expected requirements or conditions. i , Y j ) is less than T (i.e., a preset threshold, usually set to 0.8), and the landmark points that meet the rule are screened out to obtain the second pairing points.

[0135] Optionally, step S16 further includes executing the following steps:

[0136] Step S164, in response to the presence of a first special marker point satisfying a first special screening condition in the target marker points, determining a first correlation degree between first special feature vectors corresponding to the first special marker point, wherein the first special screening condition includes points paired with a plurality of open marker points and points paired with special points in the open marker points, and the special points in the open marker points are points paired with a plurality of target marker points;

[0137] Step S165, updating the first pairing point based on the first special marker point; or,

[0138] Step S166, in response to the presence of a second special marker point satisfying a second special screening condition in the target marker points, determining a second correlation degree between second special feature vectors corresponding to the second special marker point, wherein the second special screening condition includes points paired with a plurality of closed marker points and points paired with special points in the closed marker points, and the special points in the closed marker points are points paired with a plurality of target marker points;

[0139] Step S167: Update the second pairing point based on the second special marker point.

[0140] Exemplarily, the first special marker point is a special marker point obtained by screening the target marker point based on the first special screening condition, and the first special screening condition includes a point paired with multiple opening marker points, and a point paired with a special point in the opening marker point, wherein the special point in the opening marker point is a point paired with multiple target marker points. That is, assuming that a marker point in the opening posture style image and the scissor-type isolating switch image to be paired appear in the style image and is paired with multiple marker points in the scissor-type isolating switch image to be paired, or multiple marker points in the opening posture style image are paired with a marker point in the scissor-type isolating switch image to be paired, the first degree of association between the first special feature vectors corresponding to the first special marker point is further determined. The first degree of association is used to indicate the degree of correlation or closeness of connection between the first special feature vectors corresponding to the first special marker point. Suppose there is a vector X=(x1, x2, ..., x n ), Y=(y1,y2,...,y n), then the calculation formula of the correlation degree r of vectors X and Y is as follows:

[0141]

[0142]

[0143] Where ||X|| represents the norm of vector X.

[0144] Update the first pairing point based on the first special marker point: if the first correlation degree r between the vectors describing the features of these marker points is less than 0, then directly abandon this group of marker points. Finally, take the group of marker points with the largest correlation degree as the pairing points.

[0145] Exemplarily, the second special mark point is a special mark point obtained by screening the target mark point based on the second special screening condition, and the second special screening condition includes a point paired with multiple closed mark points, and a point paired with a special point in the closed mark point, wherein the special point in the closed mark point is a point paired with multiple target mark points. That is, assuming that a mark point in the closed posture style image and the scissor-type isolating switch image to be paired are paired with multiple mark points in the scissor-type isolating switch image to be paired, or multiple mark points in the closed posture style image are paired with a mark point in the scissor-type isolating switch image to be paired, the second degree of association between the second special feature vectors corresponding to the second special mark point is further determined. The calculation method of the first degree of association is the same as formula (11) and formula (12).

[0146] Optionally, in step S18, determining a target posture of the target disconnector based on the first pairing rate and the second pairing rate includes performing the following steps:

[0147] Step S181, determining a first matching rate based on the first matching points, wherein the first matching rate is used to represent the probability that the target disconnector is in an open position;

[0148] Step S182, determining a second matching rate based on the second matching points, wherein the second matching rate is used to represent the probability that the target disconnector is in a closed position;

[0149] Step S183: determining a target posture of the target disconnector based on the first matching rate and the second matching rate.

[0150] Exemplarily, the first pairing point is the number of points for which the target mark point is successfully paired with the opening mark point. The first pairing rate is determined based on the first pairing point. Assume that the number of mark points (target mark points) of the scissor-type isolating switch image to be paired is num, the number of mark points of the opening posture style image S0 is num_SO, and the number of mark points of the scissor-type isolating switch image to be paired and the opening posture style image S0 that are successfully paired is num_O. Then, the pairing rate of the scissor-type isolating switch image to be paired and the opening posture style image S0 is the first pairing rate, expressed as match_O, which is used to indicate the probability that the target isolating switch is in the opening posture. The calculation formula of match_O is as follows:

[0151]

[0152] Exemplarily, the second pairing point number is the number of points that the target mark point and the closed mark point are successfully paired. The second pairing rate is determined based on the second pairing point number. Assuming that the number of mark points of the closed posture style image S1 is num_SC, and the number of mark point pairs of the scissor-type disconnector image to be paired and the closed posture style image S1 is num_C, then the pairing rate of the scissor-type disconnector image to be paired and the closed posture style image S1 is the second pairing rate, expressed as match_C, which is used to indicate the probability that the target disconnector is in a closed posture. The calculation formula of match_C is as follows:

[0153]

[0154] Then, the target posture of the target disconnector is determined based on the first matching rate and the second matching rate, that is, the scissor-type disconnector to be matched is in an open posture or a closed posture.

[0155] Optionally, in step S183, determining a target posture of the target disconnector based on the first pairing rate and the second pairing rate includes performing the following steps:

[0156] Step S1831, determining a matching rate ratio based on the first matching rate and the second matching rate, wherein the matching rate ratio is a ratio of the first matching rate to the second matching rate;

[0157] Step S1832, in response to the pairing rate ratio being greater than 1, determining that the target posture of the target isolation switch is an open posture;

[0158] Step S1833: in response to the matching rate ratio being less than or equal to 1, determining that the target posture of the target isolation switch is a closed posture.

[0159] Exemplarily, a matching rate ratio match of the images is determined based on the first matching rate and the second matching rate, wherein the matching rate ratio is a ratio of the first matching rate to the second matching rate, and is expressed as:

[0160]

[0161] When the matching rate ratio is greater than 1, that is, match>1, the scissor-type isolating switch posture of the image to be paired is determined to be an open posture. When the matching rate ratio is less than or equal to 1, that is, match≤1, the scissor-type isolating switch posture of the image to be paired is determined to be a closed posture. The determination of the scissor-type isolating switch posture is now completed.

[0162] Figure 4 is a flow chart of a disconnector posture determination solution according to an embodiment of the present application, such as Figure 4 As shown in the figure, the disconnector posture determination scheme first constructs the data set, and then extracts the landmark points and the vectors describing the features by establishing the scale space, screening the landmark points, determining the directions of the landmark points, and generating the vectors describing the features. Finally, the landmark points are paired to determine the posture of the scissor-type disconnector of the paired image.

[0163] Figure 5 is a simplified flow chart of a disconnector posture determination scheme according to an embodiment of the present application, such as Figure 5 As shown, the scissor-type isolating switch is obtained to obtain an open posture style image, and its landmark points and vectors describing features are extracted; at the same time, the scissor-type isolating switch is obtained to obtain a closed posture style image, and its landmark points and vectors describing features are extracted. The landmark points and vectors describing features of the posture template image are stored, and after the scissor-type isolating switch image to be paired is obtained and its landmark points and vectors describing features are extracted, the landmark points are paired to determine the scissor-type isolating switch posture of the image to be paired.

[0164] In one embodiment of the present application, a data set is constructed by selecting 1,000 photos of scissor-type disconnectors taken at a substation, converting them into grayscale scissor-type disconnector images with an image resolution of approximately 4,000*3,000 pixels, and selecting 2 images, one of which is an open posture style image S0, and the other is a closed posture style image S1, and the remaining 998 are scissor-type disconnector images to be paired.

[0165] Then, the landmark points of the two posture style images are extracted, and the obtained landmark points and vectors describing the features are stored. The process of extracting landmark points and vectors describing the features is as follows:

[0166] a) Establish scale space. Convert the scissor-type disconnector image into an image in scale space. During the implementation, the Gaussian blur scale σ0 is set to 1.6. The new scale space is obtained by the formula. k is set to The principle of determining the extreme point is that if the value of a certain pixel reaches the maximum or minimum value, then the pixel is selected as an extreme point. The maximum and minimum values ​​here refer to the maximum values ​​of the 26 pixels in the upper and lower three-layer scale space. These 26 pixels are composed of 9 pixels in the upper and lower layers and 8 pixels in the middle layer. Figure 6 is a schematic diagram of determining an extreme point according to an embodiment of the present application, such as Figure 6 The figure shows the process of determining extreme points based on hierarchical scale.

[0167] b) Select the landmark point. Substitute the obtained extreme point into the derivative of F(X). If the result is 0, it is the landmark point.

[0168] c) Determine the direction of the landmark point. Figure 7 is a schematic diagram of determining the direction of a marker point according to an embodiment of the present application, such as Figure 7 As shown in the figure, the gradient magnitude and gradient direction of the pixel points in the neighborhood of the marker point are calculated. The "0°-360°" is divided into 8 intervals, each interval is 45 degrees, and the gradient directions of all pixel points are attributed to these 8 intervals. The peak value is defined as the main direction of the marker point by using histogram statistics.

[0169] d) Generate a vector describing the features. Figure 8 is a schematic diagram of sub-area division according to an embodiment of the present application, such as Figure 8 As shown, its neighborhood is divided into 10 sub-regions. Among them, the pixels contained in the 4 sub-regions other than the 6 rectangular sub-regions are the pixels in the neighborhood of the landmark point (only the pixels completely contained in the neighborhood are considered). Then the coordinate axis is transformed to the same direction as the main direction, the gradient of the pixel point is calculated, the histogram is statistically calculated, and the grayscale mean and variance are calculated. The final vector describing the feature is obtained, normalized, and the final vector describing the feature is obtained.

[0170] Then, the following operations are performed on the remaining 998 scissor-type disconnector images to be paired (taking the first image as an example):

[0171] e) Extracting landmark points and vectors describing features, the same as the above-mentioned landmark point extraction process.

[0172] f) Mark point pairing. The open posture style image S0 has 2104 marker points, the closed posture style image S1 has 1982 marker points, and the first scissor-type disconnector image to be paired has 1796 marker points. By calculating the Euclidean distance and the correlation coefficient of the special marker points, it is found that the image to be paired has 1420 matching points with the open posture style image S0 and 612 matching points with the closed posture style image S1.

[0173] g) Determine the posture of the image to be paired. Calculation shows that match_O = 67.50%, match_C = 30.88%, match = 2.19 is greater than 1, and it can be determined that the posture of the scissor-type isolation switch of the image to be paired is the open posture.

[0174] In summary, the present application obtains the target disconnector image of the target disconnector whose posture is to be determined; determines the target mark point and the first feature vector of the target disconnector image, wherein the target mark point is used to record the boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image; based on the first feature vector, the target mark point is paired with the opening mark point of the target opening posture image to obtain the first pairing point number, and based on the first feature vector, the target mark point is paired with the closing mark point of the target closing posture image to obtain the second pairing point number, wherein the first pairing point number is the number of points where the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points where the target mark point is successfully paired with the closing mark point; and determines the target posture of the target disconnector based on the first pairing point number and the second pairing point number. Thus, the purpose of automatically judging the posture of the scissor-type disconnector based on the image of the disconnector in the power grid operation area is achieved, and the technical effect of improving the accuracy of automatically judging the posture of the disconnector is achieved, thereby solving the technical problems of slow running speed and poor recognition accuracy in the related technology.

[0175] Example 2

[0176] According to an embodiment of the present application, a disconnector posture determination device for implementing the method in Embodiment 1 is also provided. Fig. 9 is a structural block diagram of a device for determining the posture of an isolating switch according to an embodiment of the present application, such as Fig. 9As shown, the isolating switch posture determination device 90 includes: an acquisition module 92, which is used to acquire a target isolating switch image of a target isolating switch whose posture is to be determined; a first determination module 94, which is used to determine a target mark point and a first feature vector of the target isolating switch image, wherein the target mark point is used to record the boundary information of the target isolating switch image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target isolating switch image; a pairing module 96, which is used to perform mark point pairing of the target mark point with the opening mark point of the target opening posture image based on the first feature vector to obtain a first pairing point number, and perform mark point pairing of the target mark point with the closing mark point of the target closing posture image based on the first feature vector to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points at which the target mark point is successfully paired with the closing mark point; a second determination module 98, which is used to determine the target posture of the target isolating switch based on the first pairing point number and the second pairing point number.

[0177] Optionally, the first determination module 94 is also used to convert the target isolating switch image into a target scale space to obtain a first isolating switch image, wherein the target scale space is used to simulate the viewing angle of observing the target isolating switch image; determine the target marker point of the first isolating switch image; determine the direction of the target marker point based on the gradient direction of the target marker point; and determine the first eigenvector based on the target marker point and the target marker point direction.

[0178] Optionally, the device further includes: a scale establishment module, configured to establish a target scale space based on a preset Gaussian blur scale.

[0179] Optionally, the first determination module 94 is further configured to convert the target isolating switch image into a target scale space to obtain a second isolating switch image; and perform a convolution operation on the second isolating switch image based on a Gaussian function to obtain the first isolating switch image.

[0180] Optionally, the first determination module 94 is further used to determine extreme points of the first isolating switch image in the target scale space, wherein the extreme points are used to represent points where the pixel values ​​of the first isolating switch image are maximum or minimum; and determine the target landmark points based on the extreme points.

[0181] Optionally, the first determination module 94 is also used to determine a preset neighborhood based on the target marker point, and divide the preset neighborhood into multiple directions; determine multiple first gradient sizes and multiple first gradient directions of the target marker point in multiple directions; determine a histogram based on the multiple first gradient sizes and multiple first gradient directions; and determine the direction of the target marker point based on the peak value of the histogram.

[0182] Optionally, the first determination module 94 is also used to divide a preset neighborhood into multiple sub-regions, wherein the multiple sub-regions include regions of various shapes; determine a target coordinate system based on the direction of a target marker point; in the target coordinate system, determine multiple second gradient magnitudes and multiple second gradient directions of the target marker point in the multiple sub-regions; determine a second eigenvector based on multiple second gradient magnitudes, multiple second gradient directions, grayscale value means of pixels in multiple sub-regions, and grayscale value variances of pixels in multiple sub-regions; and normalize the second eigenvector to obtain a first eigenvector.

[0183] Optionally, the pairing module 96 is also used to determine a third eigenvector and a fourth eigenvector, wherein the third eigenvector is used to describe the shape and structure of the target open posture image, and the fourth eigenvector is used to describe the shape and structure of the target closed posture image; based on the third eigenvector and the first eigenvector, the target landmark points are screened to obtain a first pairing point; based on the fourth eigenvector and the first eigenvector, the target landmark points are screened to obtain a second pairing point.

[0184] Optionally, the pairing module 96 is further configured to determine a first Euclidean distance between the first feature vector and the third feature vector; and filter out landmark points whose distance in the first Euclidean distance is less than a preset threshold to obtain first pairing points.

[0185] Optionally, the pairing module 96 is further configured to determine a second Euclidean distance between the first eigenvector and the fourth eigenvector; and select landmark points whose distance in the second Euclidean distance is less than a preset threshold to obtain second pairing points.

[0186] Optionally, the pairing module 96 is also used to determine a first degree of association between first special feature vectors corresponding to the first special marker point in response to the presence of a first special marker point that meets a first special filtering condition in the target marker points, wherein the first special filtering condition includes points paired with multiple open marker points and points paired with special points in the open marker points, and the special points in the open marker points are points paired with multiple target marker points; update the first pairing point based on the first special marker point; or, in response to the presence of a second special marker point that meets a second special filtering condition in the target marker points, determine a second degree of association between second special feature vectors corresponding to the second special marker point, wherein the second special filtering condition includes points paired with multiple closed marker points and points paired with special points in the closed marker points, and the special points in the closed marker points are points paired with multiple target marker points; update the second pairing point based on the second special marker point.

[0187] Optionally, the second determination module 98 is also used to determine a first matching rate based on the first matching points, wherein the first matching rate is used to indicate the probability that the target isolating switch is in an open posture; determine a second matching rate based on the second matching points, wherein the second matching rate is used to indicate the probability that the target isolating switch is in a closed posture; and determine the target posture of the target isolating switch based on the first matching rate and the second matching rate.

[0188] Optionally, the second determination module 98 is also used to determine the matching rate ratio based on the first matching rate and the second matching rate, wherein the matching rate ratio is the ratio of the first matching rate to the second matching rate; in response to the matching rate ratio being greater than 1, determining that the target posture of the target isolating switch is an open posture; in response to the matching rate ratio being less than or equal to 1, determining that the target posture of the target isolating switch is a closed posture.

[0189] Optionally, the device also includes: a determination module, used to determine a target open posture image and a target closed posture image, wherein the model of the isolating switch in the target open posture image and the model of the isolating switch in the target closed posture image are the same as the model of the target isolating switch, and the target open posture image and the target closed posture image both include a main switch and a static contact.

[0190] It should be noted that each module in the isolating switch posture determination device in the embodiment of the present application corresponds one by one to each implementation step of the isolating switch posture determination method in Example 1. Since a detailed description has been given in Example 1, some details not reflected in this embodiment can be referred to Example 1 and will not be repeated here.

[0191] Example 3

[0192] According to an embodiment of the present application, a non-volatile storage medium is also provided, which includes a stored computer program, wherein the device where the non-volatile storage medium is located executes the isolating switch posture determination method in Example 1 by running the computer program.

[0193] Optionally, the device where the non-volatile storage medium is located implements the following steps by running the computer program:

[0194] Step S12, obtaining a target disconnector image of a target disconnector whose posture is to be determined;

[0195] Step S14, determining a target marker point and a first feature vector of the target disconnector image, wherein the target marker point is used to record boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image;

[0196] Step S16, based on the first feature vector, the target mark point is paired with the opening mark point of the target opening posture image to obtain a first pairing point number, and based on the first feature vector, the target mark point is paired with the closing mark point of the target closing posture image to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points at which the target mark point is successfully paired with the closing mark point;

[0197] Step S18, determining a target posture of the target disconnector based on the first pairing point number and the second pairing point number.

[0198] According to an embodiment of the present application, a computer program product is also provided, including a computer program, and when the computer program is executed by a processor, the method for determining the posture of the isolating switch in Embodiment 1 of the present application is implemented.

[0199] Optionally, when the computer program is running, the following steps are performed:

[0200] Step S12, obtaining a target disconnector image of a target disconnector whose posture is to be determined;

[0201] Step S14, determining a target marker point and a first feature vector of the target disconnector image, wherein the target marker point is used to record boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image;

[0202] Step S16, based on the first feature vector, the target mark point is paired with the opening mark point of the target opening posture image to obtain a first pairing point number, and based on the first feature vector, the target mark point is paired with the closing mark point of the target closing posture image to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points at which the target mark point is successfully paired with the closing mark point;

[0203] Step S18, determining a target posture of the target disconnector based on the first pairing point number and the second pairing point number.

[0204] According to an embodiment of the present application, an electronic device is also provided, which includes: a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to execute the isolating switch posture determination method in Example 1 through the computer program.

[0205] Optionally, the processor is configured to implement the following steps by executing a computer program:

[0206] Step S12, obtaining a target disconnector image of a target disconnector whose posture is to be determined;

[0207] Step S14, determining a target marker point and a first feature vector of the target disconnector image, wherein the target marker point is used to record boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image;

[0208] Step S16, based on the first feature vector, the target mark point is paired with the opening mark point of the target opening posture image to obtain a first pairing point number, and based on the first feature vector, the target mark point is paired with the closing mark point of the target closing posture image to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target mark point is successfully paired with the opening mark point, and the second pairing point number is the number of points at which the target mark point is successfully paired with the closing mark point;

[0209] Step S18, determining a target posture of the target disconnector based on the first pairing point number and the second pairing point number.

[0210] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0211] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0212] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0213] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0214] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0215] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the relevant technology or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0216] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for determining the posture of an isolating switch, characterized in that: The method comprises: Acquire a target disconnector image of a target disconnector whose posture is to be determined; Determine a target marker point and a first feature vector of the target disconnector image, wherein the target marker point is used to record boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image; Based on the first feature vector, the target landmark point is paired with the opening landmark point of the target opening gesture image to obtain a first pairing point number, and based on the first feature vector, the target landmark point is paired with the closing landmark point of the target closing gesture image to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target landmark point is successfully paired with the opening landmark point, and the second pairing point number is the number of points at which the target landmark point is successfully paired with the closing landmark point; A first pairing rate is determined based on the first pairing points, wherein the first pairing rate is used to indicate the probability that the target isolating switch is in an open position; a second pairing rate is determined based on the second pairing points, wherein the second pairing rate is used to indicate the probability that the target isolating switch is in a closed position; and a target position of the target isolating switch is determined based on the first pairing rate and the second pairing rate.

2. The method according to claim 1, characterized in that Determining the target landmark point and the first feature vector of the target disconnector image comprises: Converting the target disconnector image to a target scale space to obtain a first disconnector image, wherein the target scale space is used to simulate a viewing angle for observing the target disconnector image; Determining the target landmark point of the first disconnector image; Determine the direction of the target marker point based on the gradient direction of the target marker point; The first feature vector is determined based on the target landmark point and the target landmark point direction.

3. The method according to claim 2, characterized in that The method further comprises: The target scale space is established based on a preset Gaussian blur scale.

4. The method according to claim 2, characterized in that: The converting the target isolating switch image into a target scale space to obtain a first isolating switch image comprises: Converting the target isolating switch image to the target scale space to obtain a second isolating switch image; A convolution operation is performed on the second isolating switch image based on a Gaussian function to obtain the first isolating switch image.

5. The method according to claim 2, characterized in that: The determining the target landmark point of the first isolating switch image comprises: Determine an extreme point of the first isolating switch image in the target scale space, wherein the extreme point is used to represent a point where a pixel value of the first isolating switch image is a maximum value or a minimum value; The target landmark point is determined based on the extreme point.

6. The method according to claim 2, characterized in that Determining the direction of the target marker point based on the gradient direction of the target marker point comprises: Determine a preset neighborhood based on the target landmark point, and divide the preset neighborhood into multiple directions; Determine a plurality of first gradient magnitudes and a plurality of first gradient directions of the target marker point in the plurality of directions; determining a histogram based on the plurality of first gradient magnitudes and the plurality of first gradient directions; The direction of the target landmark point is determined based on the peak value of the histogram.

7. The method according to claim 6, characterized in that The determining the first feature vector based on the target landmark point and the direction of the target landmark point comprises: Dividing the preset neighborhood into a plurality of sub-regions, wherein the plurality of sub-regions include regions of various shapes; Determine a target coordinate system based on the direction of the target landmark point; In the target coordinate system, determining a plurality of second gradient magnitudes and a plurality of second gradient directions of the target marker point in the plurality of sub-regions; Determine a second eigenvector based on the multiple second gradient magnitudes, the multiple second gradient directions, the grayscale value means of the pixels of the multiple sub-regions, and the grayscale value variances of the pixels of the multiple sub-regions; The second eigenvector is normalized to obtain the first eigenvector.

8. The method according to claim 1, characterized in that The performing marker point pairing on the target marker point with the opening marker point of the target opening posture image based on the first feature vector to obtain a first pairing point number, and performing marker point pairing on the target marker point with the closing marker point of the target closing posture image based on the first feature vector to obtain a second pairing point number comprises: Determine a third eigenvector and a fourth eigenvector, wherein the third eigenvector is used to describe the shape and structure of the target open posture image, and the fourth eigenvector is used to describe the shape and structure of the target closed posture image; Screening the target landmark point based on the third eigenvector and the first eigenvector to obtain a first pairing point; The target landmark point is screened based on the fourth eigenvector and the first eigenvector to obtain a second pairing point.

9. The method according to claim 8, characterized in that The screening of the target landmark point based on the third feature vector and the first feature vector to obtain a first pairing point comprises: determining a first Euclidean distance between the first eigenvector and the third eigenvector; The landmark points whose distance in the first Euclidean distance is smaller than a preset threshold are screened out to obtain the first pairing points.

10. The method according to claim 9, characterized in that The screening of the target landmark points based on the fourth eigenvector and the first eigenvector to obtain the second pairing point comprises: determining a second Euclidean distance between the first eigenvector and the fourth eigenvector; The landmark points whose distance in the second Euclidean distance is smaller than the preset threshold are screened out to obtain the second pairing points.

11. The method according to claim 9 or 10, characterized in that: The method further comprises: In response to the presence of a first special marker point that meets a first special screening condition among the target marker points, determining a first correlation degree between first special feature vectors corresponding to the first special marker point, wherein the first special screening condition includes points paired with a plurality of the start marker points and points paired with a special point among the start marker points, and the special point among the start marker points is a point paired with a plurality of target marker points; Update the first pairing point based on the first special marker point; or, In response to the presence of a second special marker point that meets a second special screening condition in the target marker points, determining a second degree of association between second special feature vectors corresponding to the second special marker point, wherein the second special screening condition includes points paired with a plurality of the closed marker points and points paired with special points in the closed marker points, and the special points in the closed marker points are points paired with a plurality of the target marker points; The second pairing point is updated based on the second special landmark point.

12. The method according to claim 1, characterized in that Determining the target posture of the target isolation switch based on the first pairing rate and the second pairing rate includes: Determining a matching rate ratio based on the first matching rate and the second matching rate, wherein the matching rate ratio is a ratio of the first matching rate to the second matching rate; In response to the pairing rate ratio being greater than 1, determining that the target posture of the target isolation switch is an open posture; In response to the matching rate ratio being less than or equal to 1, determining that the target posture of the target isolation switch is a closed posture.

13. The method according to claim 1, characterized in that The method further comprises: Determine the target open posture image and the target closed posture image, wherein the model of the disconnector in the target open posture image and the model of the disconnector in the target closed posture image are the same as the model of the target disconnector, and both the target open posture image and the target closed posture image include a main switch and a static contact.

14. A device for determining the posture of an isolating switch, characterized in that: The device comprises: An acquisition module, used for acquiring a target disconnector image of a target disconnector whose posture is to be determined; A first determination module is used to determine a target marker point and a first feature vector of the target disconnector image, wherein the target marker point is used to record boundary information of the target disconnector image, and the first feature vector is a multidimensional vector, and the first feature vector is used to describe the shape and structure of the target disconnector image; a pairing module, configured to perform pairing of the target marker point with an opening marker point of a target opening gesture image based on the first feature vector to obtain a first pairing point number, and perform pairing of the target marker point with a closing marker point of a target closing gesture image based on the first feature vector to obtain a second pairing point number, wherein the first pairing point number is the number of points at which the target marker point is successfully paired with the opening marker point, and the second pairing point number is the number of points at which the target marker point is successfully paired with the closing marker point; A second determination module is used to determine a first matching rate based on the first matching points, wherein the first matching rate is used to indicate the probability that the target isolating switch is in an open posture; determine a second matching rate based on the second matching points, wherein the second matching rate is used to indicate the probability that the target isolating switch is in a closed posture; and determine a target posture of the target isolating switch based on the first matching rate and the second matching rate.

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program is configured to execute the disconnector posture determination method described in any one of claims 1 to 13 when running on a computer or a processor.

16. A computer program product, characterized in that The invention comprises a computer program, which, when executed by a processor, implements the disconnector posture determination method described in any one of claims 1 to 13.

17. An electronic device comprising a memory and a processor, characterized in that: The memory stores a computer program, and the processor is configured to run the computer program to execute the disconnector posture determination method described in any one of claims 1 to 13 above.

Images