Knife switch on-off state identification method and device
By performing binary processing and overlapping analysis of feature images on the original image of the knife switch, the problem of low sensitivity to identify abnormal states in the prior art is solved, and efficient and accurate state recognition of various types of knife switches is achieved.
Patent Information
- Application Number
- CN202310846926.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-11
- Publication Date
- 2025-01-14
AI Technical Summary
The prior art has low sensitivity and poor generalization when identifying abnormal states of circuit breakers and isolation switches, especially for non-dual-arm open type knife switches.
By binarizing the original image of the knife switch, the knife switch area is divided, and the unique feature images of different judgment states are used for overlap analysis, and the actual state of the knife switch is determined based on the threshold judgment, including the split, merging and intermediate states. .
It improves the sensitivity and generalization of the identification of abnormal states of separation and combination, and is suitable for various types of knife switches, enhancing the accuracy and efficiency of knife switch status recognition.
Smart Images
Figure CN119313912A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of video monitoring, and in particular relates to a method and a device for identifying the opening and closing states of a knife switch. Background Art
[0002] "One-key sequential control" is an operation mode for substation switching operation, which can realize the prefabrication of operation project software, modular construction of operation tasks, automatic identification of equipment status, intelligent verification of anti-error interlocking, one-key start of operation steps, and automatic sequential execution of operation process. "Double confirmation" means that when the equipment is operated remotely, at least two non-same principle or non-same source indications should change accordingly, and all these indications should change accordingly at the same time to confirm that the equipment has been operated in place.
[0003] At present, the main problem of one-button sequential control is the double confirmation of the positions of circuit breakers and disconnectors. In particular, there are many types of knife switches. The existing video double confirmation is mainly for the opening and closing states, not for the abnormal opening and closing states, and the sensitivity of identifying the abnormal opening and closing states is not enough; it is only for a certain type of knife switch (generally double-arm opening type knife switch), not applicable to all types of knife switches; some knife switch state recognition algorithms based on deep learning methods mainly train models for specific knife switches, and the application deployment experience and effect are not good, mainly because the generalization performance of new and unprecedented knife switch types is poor. Summary of the invention
[0004] The object of the present invention is to provide a method and device for identifying the opening and closing states of knife switches, so as to solve the problem of low sensitivity and poor generalization in identifying the abnormal opening and closing states.
[0005] In order to solve the above technical problems, the present invention provides a method for identifying the opening and closing states of a knife switch, comprising the following steps:
[0006] 1) Binarizing the original image of the knife switch at a certain moment to segment the knife switch in the image and obtain the knife switch area in the binary image;
[0007] 2) According to the amount of overlap between the knife switch area in the binary image obtained in step 1) and the knife switch feature area in the unique feature images of the knife switch in different judgment states, determine the actual state of the knife switch at a certain moment; wherein, the unique feature images of the knife switch in different judgment states are obtained by first binarizing the original images of the knife switch in different actual states and then performing logical operations; the logical operation refers to subtracting the overlapping part of the binary image of the knife switch in a certain actual state from the part of the binary image of the knife switch in other actual states to obtain the unique feature image of the knife switch in a certain judgment state.
[0008] The beneficial effects are as follows: in order to solve the problem of low sensitivity and poor generalization in identifying the abnormal state of separation and combination, the present invention performs binarization processing on the original image of the knife switch at a certain moment, and judges the opening and closing state of the current knife switch according to the knife switch area in the binarized image and the knife switch area in the unique feature image of different judgment states of the knife switch established, so as to find out the separation and abnormal state. The present invention not only identifies the separation and combination state, but also identifies the abnormal separation and combination state, has high sensitivity and generalization, makes the knife switch identification efficient, and improves the accuracy of knife switch state identification.
[0009] Furthermore, different actual states of the knife switch include: an open state, a closed state and an intermediate state.
[0010] Furthermore, the characteristic images of different judgment states of the knife switch include: a unique characteristic image of the open state, a unique characteristic image of the closed state and the intermediate state, a unique characteristic image of the closed state, and a unique characteristic image of the open state and the intermediate state; wherein the calculation expression of the unique characteristic image of the open state is as follows:
[0011] mask1={open_and-(middle_or∪closed_or)};
[0012] The unique feature images of the combined state and the intermediate state are calculated as follows:
[0013] mask2={(middle_or∪closed_or)–open_or};
[0014] The unique feature image of the combined state is calculated as follows:
[0015] mask3={closed_and-(middle_or∪open_or)};
[0016] The unique feature images of the open state and the intermediate state are calculated as follows:
[0017] mask4={(middle_or∪open_or)–closed_or};
[0018] Among them, mask1, mask2, mask3 and mask4 are the unique feature images of the open state, the feature images of the closed state and the intermediate state, the unique feature images of the closed state, and the feature images of the open state and the intermediate state respectively; open is the binarized image of the original image of the actual open state of the knife gate; middle is the binarized image of the original image of the actual intermediate state of the knife gate; closed is the binarized image of the original image of the actual closed state of the knife gate; and and or respectively represent the intersection and union operations on the binarized images of the original images of the same actual states of the knife gate.
[0019] The beneficial effects are: finding unique feature images of different states of the knife switch, the feature images are helpful to judge the state of the knife switch at any time, not only to identify the separation and combination state, but also to identify the abnormal separation and combination state, with high sensitivity and generalization.
[0020] Furthermore, the means for determining the actual state of the knife switch at a certain moment is as follows: if the area where the knife switch area in the binary image of the knife switch at a certain moment overlaps with the knife switch feature area in the unique feature image of the knife switch in the open state is greater than the knife switch feature area in the unique feature image of the knife switch in the open state multiplied by a set first threshold, and the area where the knife switch area in the binary image of the knife switch at a certain moment overlaps with the knife switch feature area in the unique feature image of the knife switch in the closed state and the intermediate state is less than the knife switch feature area in the unique feature image of the knife switch in the closed state and the intermediate state multiplied by a set second threshold, then the actual state of the knife switch is determined to be the open state; wherein the value range of the first threshold is 0.9-0.99, and the value range of the second threshold is 0.01-0.1.
[0021] The beneficial effects are as follows: the above judgment method can effectively identify whether the actual state of the knife switch is open, which is conducive to obtaining the actual opening and closing state of the knife switch, and is conducive to finding the abnormal opening and closing state of the knife switch, making the knife switch identification efficient, and improving the accuracy of the knife switch state identification.
[0022] Further, if the area where the knife gate area in the binary image of the knife gate at a certain moment overlaps with the knife gate feature area in the unique feature image of the knife gate in the closed state is greater than the knife gate feature area in the unique feature image of the knife gate in the closed state multiplied by the set third threshold, and the area where the knife gate area in the binary image of the knife gate at a certain moment overlaps with the knife gate feature area in the unique feature image of the knife gate in the open state and the intermediate state is less than the knife gate feature area in the unique feature image of the knife gate in the open state and the intermediate state multiplied by the set fourth threshold, then the actual state of the knife gate is determined to be the closed state; the value range of the third threshold is 0.9-0.99, and the value range of the fourth threshold is 0.01-0.1.
[0023] The beneficial effects are: the above judgment method can effectively identify whether the actual state of the knife switch is in the closed state, which is conducive to obtaining the actual opening and closing state of the knife switch, and is conducive to finding the abnormal opening and closing state of the knife switch, making the knife switch identification efficient and improving the accuracy of the knife switch state identification.
[0024] Further, if the judgment conditions for judging whether the knife switch is in the open state and the closed state are not met, it is determined that the actual state of the knife switch is an intermediate state.
[0025] The beneficial effects are: the above judgment method can effectively identify whether the actual state of the knife switch is the intermediate state, which is conducive to obtaining the actual opening and closing state of the knife switch, and is conducive to finding the abnormal opening and closing state of the knife switch, making the knife switch identification efficient and improving the accuracy of the knife switch state identification.
[0026] To solve the above technical problems, the present invention also provides a device for identifying the opening and closing states of a knife switch, comprising a memory and a processor, wherein the processor is used to execute computer program instructions stored in the memory to implement the above-mentioned method for identifying the opening and closing states of a knife switch.
[0027] The beneficial effect is that the device ensures the effective and reliable execution of a method for identifying the opening and closing states of a knife switch. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] Figure 1-1 It is a schematic diagram of a video frame-by-frame combined state image folder of the present invention;
[0029] Figure 1-2 It is a schematic diagram of an image folder of an intermediate state of video frame segmentation of the present invention;
[0030] Figure 1-3 It is a schematic diagram of an image folder in a video frame-opening state of the present invention;
[0031] Figure 2 This is a knife switch segmentation effect diagram of the present invention;
[0032] Figure 3-1 is a visualization diagram of the open state characteristics of the present invention;
[0033] Figure 3-2 It is a characteristic visualization diagram of the combined state and intermediate state of the present invention;
[0034] Figure 3-3 It is a visualization diagram of the combined state characteristics of the present invention;
[0035] Figure 3-4 It is a characteristic visualization diagram of the open state and the intermediate state of the present invention;
[0036] Figure 4-1 It is a combined state recognition effect diagram of the present invention;
[0037] Figure 4-2 It is the intermediate state recognition effect diagram a of the present invention;
[0038] Figure 4-3 It is the intermediate state recognition effect diagram b of the present invention;
[0039] Figure 4-4 It is a state recognition effect diagram of the present invention;
[0040] Figure 5 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION
[0041] The basic concept of the present invention is: the present invention first performs video frame processing based on the inspection point knife switch action video, and places the original images of different actual states in the corresponding folders; by first performing binary processing on the original image of the opening and closing state, and then performing logical operations, the unique feature images of the knife switch in different judgment states are obtained, and the obtained unique feature images of the judgment state are stored and modeled; finally, based on the above model, the subsequent knife switch is identified for abnormal opening and closing faults. Based on this concept, a knife switch opening and closing state identification method and a knife switch opening and closing state identification device of the present invention can be realized.
[0042] The present invention will be described in detail below with reference to the accompanying drawings and method embodiments.
[0043] Method Example:
[0044] A method for identifying the switch opening and closing states of the present invention is shown in the flowchart as follows: Figure 5 As shown, the specific implementation steps include:
[0045] Step 1: First, mark the opening and closing status according to the knife switch action video of the inspection point. Through video framing and file movement, the original image of the knife switch is stored in the folders of the three actual states of open state, closed state, and intermediate state, such as Figure 1-1 to Figure 1-3 shown.
[0046] Step 2: By performing a series of processing on the original image of the actual state, a unique feature image of the switch judgment state is obtained and stored and modeled. The processing process specifically includes:
[0047] 1) Perform image segmentation (binarization) on the original image of the knife switch after framing, including but not limited to image segmentation methods based on global threshold, traditional image segmentation methods such as Otsu's optimal global threshold, and image segmentation methods based on deep learning. The segmentation effect is as follows: Figure 2 shown.
[0048] 2) Perform set and logical operations on the segmented image data set to obtain unique feature images of different judgment states of the switch, as shown in the visualization diagram Figure 3-1 to Figure 3-4 The logical operations and descriptions of feature tags are shown in Table 1, and the basic set operations are shown in Table 2.
[0049] Table 1
[0050] Unique features of opening and closing states describe mask1={open_and-(middle_or∪closed_or)} It is open, not closed, and not in between. mask2={(middle_or∪closed_or)–open_or} It is closed, or in the middle, not open. mask3={closed_and-(middle_or∪open_or)} It is closed, not open, and not in between. mask4={(middle_or∪open_or)–closed_or} Is it open, or in the middle, not closed
[0051] Table 2
[0052]
[0053] Among them, mask1, mask2, mask3 and mask4 are the unique feature images of the open state, the unique feature images of the closed state and the intermediate state, the unique feature images of the closed state and the unique feature images of the open state and the intermediate state respectively; mask1 = {open_and-(middle_or∪closed_or)} means that the binary image of the actual state of the knife switch is the open state minus the binary image of the actual state of other states (the union of all the binary images of the intermediate state and the union of all the binary images of the closed state) with the same value is used to obtain the unique feature image of the knife switch judgment state as the open state. The principles of the remaining mask2, mask3, and mask4 are similar to mask1.
[0054] 3) Write mask1, mask2, mask3, and mask4 obtained in step 2) into a json file to complete feature storage modeling.
[0055] Step 3: Based on the above template, the subsequent knife switch action video is used to identify abnormal opening and closing faults. First, the json model file is loaded, and then the current knife switch action video is framed, and then the framed image is binarized. Finally, the actual state of the knife switch is judged by combining the opening and closing state judgment logic rules. The actual state judgment logic rules are as follows: Figures 4-1 to 4-4 shown.
[0056] Table 3
[0057] Open and close state judgment logic Status Classification if s1>y1*k1 or s2<y2*k2 Open state if s3>y3*k3 or s4<y4*k4 Close state other Intermediate state
[0058] Among them, k1 is the area of the knife switch feature region in the unique feature image of the open state; k2 is the area of the knife switch feature region in the unique feature image of the closed state and the intermediate state; k3 is the area of the knife switch feature region in the unique feature image of the closed state; k4 is the area of the knife switch feature region in the unique feature image of the open state and the intermediate state; y1, y3 are the first threshold value of the positive criterion and the third threshold value of the positive criterion, respectively, generally taken as (0.9-0.99); y2, y4 are the second threshold value of the negative criterion and the fourth threshold value of the negative criterion, respectively, generally taken as (0.01-0.1); s1 is the difference between the knife switch region in the binary image of the knife switch at a certain moment and the knife switch opening s1 is the area of the knife gate area in the binary image at a certain moment of the knife gate overlapping with the knife gate feature area in the unique feature image of the knife gate closed state and the intermediate state; s2 is the area of the knife gate area in the binary image at a certain moment of the knife gate overlapping with the knife gate feature area in the unique feature image of the knife gate closed state; s3 is the area of the knife gate area in the binary image at a certain moment of the knife gate overlapping with the knife gate feature area in the unique feature image of the knife gate closed state; s4 is the area of the knife gate area in the binary image at a certain moment of the knife gate overlapping with the knife gate feature area in the unique feature image of the knife gate open state and the intermediate state. The above-mentioned area values are equal to the sum of the number of pixels in their respective corresponding areas.
[0059] Among them, the criterion logic for the actual state of the disconnecting switch being in the open state is s1 > y1 * k1 or s2 < y2 * k2. Specifically, it means that if the area (s1) of the overlapping region between the disconnecting switch region in the binary image of the disconnecting switch at a certain moment and the disconnecting switch feature region in the unique feature image of the open state of the disconnecting switch is greater than the area (k1) of the disconnecting switch feature region in the unique feature image of the open state of the disconnecting switch multiplied by the set threshold y1, and the area (s2) of the overlapping region between the disconnecting switch region in the binary image of the disconnecting switch at a certain moment and the disconnecting switch feature regions in the unique feature images of the closed state and the intermediate state of the disconnecting switch is less than the area (k2) of the disconnecting switch feature regions in the unique feature images of the closed state and the intermediate state of the disconnecting switch multiplied by the set threshold, then it is determined that the actual state of the disconnecting switch is in the open state.
[0060] The present invention is applicable to all types of disconnecting switches, including but not limited to double-arm open type, rotary type, folding type, V-type disconnecting switches, and has strong usability; the present invention not only identifies the open and closed states, but also identifies abnormal open and closed states, and has high sensitivity and generalization.
[0061] Device embodiment:
[0062] A device for identifying the open and closed states of a disconnecting switch according to the present invention includes a memory and a processor. The processor is used to execute computer program instructions stored in the memory to implement a method for identifying the open and closed states of a disconnecting switch introduced in the method embodiment of the present invention. Among them, the processor can be a processing device such as a programmable logic device FPGA. The memory can be various memories that store information by means of electric energy, such as RAM, ROM, etc., and can also be other types of memories.
Claims
1. A method for identifying the opening and closing states of a knife switch, characterized in that: The steps include: 1) Binarizing the original image of the knife switch at a certain moment to segment the knife switch in the image and obtain the knife switch area in the binary image; 2) According to the amount of overlap between the knife switch area in the binary image obtained in step 1) and the knife switch feature area in the unique feature images of the knife switch in different judgment states, determine the actual state of the knife switch at a certain moment; wherein, the unique feature images of the knife switch in different judgment states are obtained by first binarizing the original images of the knife switch in different actual states and then performing logical operations; the logical operation refers to subtracting the overlapping part of the binary image of the knife switch in a certain actual state from the part of the binary image of the knife switch in other actual states to obtain the unique feature image of the knife switch in a certain judgment state.
2. The method for identifying the switch opening and closing state according to claim 1 is characterized in that: The different actual states of the knife switch include: open state, closed state and intermediate state.
3. The method for identifying the switch opening and closing state according to claim 1 or 2, characterized in that: The characteristic images of different judgment states of the knife switch include: unique characteristic images of the open state, unique characteristic images of the closed state and the intermediate state, unique characteristic images of the closed state, and unique characteristic images of the open state and the intermediate state; wherein the calculation expression of the unique characteristic image of the open state is as follows: mask1={open_and-(middle_or∪closed_or)}; The unique feature images of the combined state and the intermediate state are calculated as follows: mask2={(middle_or∪closed_or)–open_or}; The unique feature image of the combined state is calculated as follows: mask3={closed_and-(middle_or∪open_or)}; The unique feature images of the open state and the intermediate state are calculated as follows: mask4={(middle_or∪open_or)–closed_or}; Among them, mask1, mask2, mask3 and mask4 are the unique feature images of the open state, the feature images of the closed state and the intermediate state, the unique feature images of the closed state, and the feature images of the open state and the intermediate state respectively; open is the binarized image of the original image of the actual open state of the knife gate; middle is the binarized image of the original image of the actual intermediate state of the knife gate; closed is the binarized image of the original image of the actual closed state of the knife gate; and and or respectively represent the intersection and union operations on the binarized images of the original images of the same actual states of the knife gate.
4. The method for identifying the switch opening and closing state according to claim 3 is characterized in that: The means for determining the actual state of the knife switch at a certain moment is as follows: if the area where the knife switch area in the binary image of the knife switch at a certain moment overlaps with the knife switch feature area in the unique feature image of the knife switch in the open state is greater than the knife switch feature area in the unique feature image of the knife switch in the open state multiplied by a set first threshold, and the area where the knife switch area in the binary image of the knife switch at a certain moment overlaps with the knife switch feature area in the unique feature image of the knife switch in the closed state and the intermediate state is less than the knife switch feature area in the unique feature image of the knife switch in the closed state and the intermediate state multiplied by a set second threshold, then the actual state of the knife switch is determined to be the open state; wherein the value range of the first threshold is 0.9-0.99, and the value range of the second threshold is 0.01-0.
1.
5. The method for identifying the switch opening and closing state according to claim 4 is characterized in that: If the area where the knife switch area in the binary image of the knife switch at a certain moment overlaps with the knife switch feature area in the unique feature image of the knife switch in the closed state is greater than the knife switch feature area in the unique feature image of the knife switch in the closed state multiplied by the set third threshold, and the area where the knife switch area in the binary image of the knife switch at a certain moment overlaps with the knife switch feature area in the unique feature image of the knife switch in the open state and the intermediate state is less than the knife switch feature area in the unique feature image of the knife switch in the open state and the intermediate state multiplied by the set fourth threshold, then the actual state of the knife switch is determined to be the closed state; the value range of the third threshold is 0.9-0.99, and the value range of the fourth threshold is 0.01-0.
1.
6. The method for identifying the switch opening and closing state according to claim 5 is characterized in that: If the judgment conditions for judging whether the knife switch is in the open state and the closed state are not met, the actual state of the knife switch is judged to be the intermediate state.
7. A device for identifying the opening and closing state of a knife switch, characterized in that: It comprises a memory and a processor, wherein the processor is used to execute computer program instructions stored in the memory to implement the knife switch opening and closing state identification method as described in any one of claims 1 to 6.