An arcing detection method

Abstract

The application provides an arc detection method, which comprises the following steps: when an arc occurs, receiving an arc signal sent by an arc sensor; obtaining an arc picture according to the arc signal, and obtaining a first preprocessed image by preprocessing the arc picture; obtaining a first target region image by ROI intercepting the first preprocessed image, and obtaining a first difference result by performing difference value on two continuous images in the first target region image; performing binaryzation on the first difference result and a first target threshold to obtain a first processing result image; and judging whether the first processing result image contains a target gray value highlight or not to obtain a detection result. The application solves the problem of low accuracy of the arc in the prior art.

Description

Technical Field

[0001] This invention relates to the field of track inspection technology, and in particular to a method for detecting arcing. Background Technology

[0002] The overhead contact system, as the main power supply equipment for rail transit, is the primary power source for electric locomotives and is considered the "blood vessel" of the rail transit line. Therefore, the pantograph-contact relationship plays a crucial role in ensuring the normal operation of the entire rail transit line. During the operation of electric locomotives, the dynamic movement of the overhead contact system and pantograph can easily lead to unstable contact between them, causing the air gap to break down and resulting in discharge, forming an arc in the contact system. The generation of contact system arcs is mainly due to factors such as excessive changes in contact pressure of the overhead contact system, excessive height differences within the contact wire span or between adjacent positioning points, poor contact system smoothness, hard bends, twists, hard spots in the conductor, and excessive train speed. The generation of arcs affects the safe and stable operation of electric locomotives, increases the wear and tear on the pantograph slide and contact wire, and burns equipment such as sectional insulators, electrical connection wires, and clamps. In summary, improving the accuracy of pantograph-contact arc detection can help workers accurately locate potential safety hazards (such as excessive changes in contact pressure, excessive height differences within the contact wire span or between adjacent positioning points, and poor contact system smoothness), thereby improving the safety and stability of electric locomotive operation.

[0003] For arc detection, the existing mature solution is to use photomultiplier tubes to convert light signals within a certain wavelength range into electrical signals, thereby realizing the detection of arcing in the pantograph-catenary system. However, this method is easily affected by visible light interference, and since pantograph-catenary detection equipment usually uses supplementary lighting in tunnels, the accuracy of arc detection using only photomultiplier tubes is relatively low. Summary of the Invention

[0004] In view of the shortcomings of the prior art, the present invention provides an arc detection method, which solves the problem of low accuracy of arc detection in the prior art.

[0005] This invention provides a method for detecting arcing, characterized in that the detection method includes: receiving an arcing signal emitted by an arcing sensor when an arcing occurs; acquiring an arcing image based on the arcing signal; preprocessing the arcing image to obtain a first preprocessed image; performing ROI cropping on the first preprocessed image to obtain a first target region image; performing difference calculation on two consecutive frames in the first target region image to obtain a first difference result; binarizing the first difference result and a first target threshold to obtain a first processed result image; and determining whether the first processed result image contains a bright spot with a target grayscale value to obtain a detection result.

[0006] Optionally, determining whether the first processed result image contains a bright spot with the target gray value and obtaining the detection result includes: if the first processed result image contains a bright spot with the target gray value, then it is determined that a visible light arc has occurred.

[0007] Optionally, determining whether the first processed result image contains a bright spot with the target gray value to obtain a detection result includes: if the first processed result image does not contain a bright spot with the target gray value, performing model matching on the first processed result image to obtain a matching result image; performing ROI cropping on the matching result image to obtain a second target region image; performing difference on two consecutive frames in the second target region image to obtain a second difference result; binarizing the second difference result and the second target threshold to obtain a second processed result image; and determining whether the second processed result image contains a bright spot with the target gray value to obtain a detection result.

[0008] Optionally, determining whether the second processed image contains a bright spot with the target gray value to obtain a detection result includes: if the second processed image contains a bright spot with the target gray value, it is determined that a visible light arc has occurred; if the second processed image does not contain a bright spot with the target gray value, it is determined that a visible light arc has not occurred.

[0009] Optionally, performing model matching on the first processed result image to obtain a matching result image includes: performing model matching on the bus region in the first processed result image.

[0010] Optionally, performing model matching on the first processed image to obtain a matching result image further includes: cropping and gamma correction on the first processed image to obtain a second preprocessed image; dividing the second preprocessed image into several 8*8 small units and calculating the gradient histogram of each small unit; normalizing the histogram using a 2*2 sliding window to obtain a normalized result; extracting feature vectors from the normalized result and matching the feature vectors with a standard model to obtain the matching result image.

[0011] Optionally, the gamma correction formula is:

[0012] f(x) = x γ

[0013] Where x represents the gray value of a pixel, and γ represents the gamma coefficient.

[0014] Optionally, calculating the gradient histogram for each small unit includes:

[0015] The horizontal and vertical gradients are calculated using the kernel matrix.

[0016] Optionally, the feature vector includes magnitude and direction;

[0017]

[0018]

[0019] Where g represents the amplitude, g x G represents the horizontal gradient. y represents the vertical gradient, and v represents the direction angle.

[0020] Optionally, after determining whether the first processed image contains a bright spot with a target grayscale value and obtaining the detection result, the detection method further includes: storing the detection result in a database and displaying it visually.

[0021] Compared with the prior art, the present invention has the following beneficial effects:

[0022] When an arcing signal is received from an arcing sensor, the occurrence of arcing can be confirmed. An arcing image is then acquired, and preprocessed to obtain a first preprocessed image. A Region of Interest (ROI) is extracted from the first preprocessed image to obtain a first target region image. The difference between two consecutive frames in the first target region image is obtained to obtain a first difference result. The first difference result and a first target threshold are binarized to obtain a first processed result image. It is then determined whether the first processed result image contains a bright spot with a target grayscale value to obtain a detection result, thus achieving visible light arcing detection. Compared to existing technologies that use electromultiplier tubes exposed to external visible light, this application uses image recognition to determine whether an object containing a visible light arc is present by judging whether the first processed result image contains a bright spot with a target grayscale value. This eliminates interference from external visible light, thereby improving the accuracy of arcing detection. Attached Figure Description

[0023] Figure 1 A flowchart of an arc detection method provided in an embodiment of the present invention.

[0024] Figure 2 A flowchart of another arc detection method provided in an embodiment of the present invention;

[0025] Figure 3 This is a flowchart of another arc detection method provided in an embodiment of the present invention. Detailed Implementation

[0026] The technical solutions of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0027] Figure 1 A flowchart of an arc detection method provided in an embodiment of the present invention is shown below. Figure 1As shown, the detection method includes the following steps:

[0028] Step S100: When arcing occurs, receive the arcing signal emitted by the arcing sensor;

[0029] In step S100, when an arc occurs, the arc sensor can detect the arc signal, which includes information such as arc duration and intensity.

[0030] Step S200: Obtain an arc image based on the arc signal, and preprocess the arc image to obtain a first preprocessed image;

[0031] In step S200, arcing is divided into invisible light arcing and visible light arcing. When an arcing signal is received, it proves that arcing has occurred. The camera is activated to take a picture to obtain an arcing image, and then the arcing image is preprocessed to obtain a first preprocessed image.

[0032] Specifically, preprocessing includes image format conversion and image rotation operations, converting 3-channel images into single-channel grayscale images for easier subsequent transmission and calculation, thus improving image processing speed. Image rotation employs a radial transformation scheme. A 3x3 transformation matrix is ​​created based on the required rotation angle and multiplied with the original image to obtain the rotated image. Compared to traditional image rotation schemes, this method offers significantly faster processing speeds.

[0033] Step S300: Perform ROI cropping on the first preprocessed image to obtain the first target region image, and perform difference calculation on two consecutive frames in the first target region image to obtain the first difference result;

[0034] In step S300, the first target region image is obtained by performing ROI cropping on the first preprocessed image, and only the image of the target region is identified.

[0035] Step S400: Binarize the first difference result and the first target threshold to obtain the first processing result image;

[0036] In step S400, the first difference result and the first target threshold (thr_1) are binarized, wherein the first target threshold (thr_1) is set according to actual needs and experience.

[0037] Step S500: Determine whether the first processed image contains a bright spot with the target gray value, and obtain the detection result.

[0038] In step S500, if the first processed image contains a bright spot with a target grayscale value, it is determined that a visible light arc has occurred. Specifically, if the first processed image contains a bright spot with a target grayscale value of 255, it is determined that a visible light arc has occurred.

[0039] In this embodiment, when an arcing signal is received from an arcing sensor, the occurrence of arcing can be confirmed. An arcing image is then acquired, and preprocessed to obtain a first preprocessed image. A ROI (Region of Interest) is extracted from the first preprocessed image to obtain a first target region image. The difference between two consecutive frames in the first target region image is calculated to obtain a first difference result. The first difference result and a first target threshold are binarized to obtain a first processed result image. It is then determined whether the first processed result image contains a bright spot with a target grayscale value to obtain a detection result, thus achieving the detection of visible light arcing. Compared to existing technologies that use electromultiplier tubes exposed to external visible light, this application uses image recognition to determine whether an object containing visible light arcing exists by judging whether the first processed result image contains a bright spot with a target grayscale value. This eliminates interference from external visible light, thereby improving the accuracy of arcing detection.

[0040] It should be noted that the program for this arc detection method uses a JSON file to store the relevant configuration information, mainly including: camera configuration information, calibration parameter information, log storage information, video parameter information, UDP broadcast information, and other algorithm-related parameters. The program determines the locomotive's driving status based on inertial navigation information. When the locomotive starts, the program prioritizes activating all hardware devices: camera, LED light source, etc. Since hardware activation requires a fixed time, the system reserves 30-60 seconds for hardware device startup time. Subsequently, the system loads the JSON configuration parameters, initializes the camera, UDP broadcast receiver / sender, LED light source, and other devices, and checks whether they are operating normally. After verification, it takes pictures at a certain frequency and transmits them back. When the locomotive is detected to have no speed, the system does not immediately shut down the equipment but continues to operate normally to avoid operational failure due to a short stop of the locomotive. If the locomotive is detected to have no speed signal for an extended period, the system will automatically shut down the algorithm program and all hardware devices. This program and equipment control method extracts parameter information into a configuration file, which enhances system stability and facilitates system maintenance, management, and upgrades.

[0041] Figure 2 A flowchart of another arc detection method provided in an embodiment of the present invention is shown below. Figure 2 As shown, determining whether the first processed image contains bright spots with the target grayscale value to obtain the detection result includes the following steps:

[0042] Step S510: If the first processed result image does not contain a bright spot with the target gray value, perform model matching on the first processed result image to obtain a matching result image;

[0043] In step S510, if the first processed result image does not contain a bright spot with the target gray value (the gray value can be 255), it means that no visible light arc was detected in the image. Then, model matching is performed on the first processed result image. Specifically, model matching is performed on the bus region in the first processed result image.

[0044] Step S520: Perform ROI cropping on the matching result image to obtain the second target region image, and perform difference calculation on two consecutive frames in the second target region image to obtain the second difference result;

[0045] In step S520, after obtaining the matching result image, the ROI is cropped again to obtain the second target region image, and the difference between two consecutive frames in the second target region image is obtained to obtain the second difference result.

[0046] Step S530: Binarize the second difference result and the second target threshold to obtain the second processing result image;

[0047] In step S530, the second difference result and the second target threshold (thr_2) are binarized, wherein the first target threshold (thr_2) is set according to actual needs and experience.

[0048] Step S540: Determine whether the second processed image contains a bright spot with the target gray value, and obtain the detection result.

[0049] In step S540, if the second processed image contains a bright spot with the target grayscale value, it is determined that a visible light arc has occurred; if the second processed image does not contain a bright spot with the target grayscale value, it is determined that no visible light arc has occurred. Specifically, if the second processed image contains a bright spot with a target grayscale value of 255, it is determined that a visible light arc has occurred.

[0050] Figure 3 A flowchart of another arc detection method provided in an embodiment of the present invention is shown below. Figure 3 As shown, performing model matching on the first processed image to obtain a matched image further includes the following steps:

[0051] Step S511: Crop and gamma correction are performed on the first processed image to obtain a second preprocessed image;

[0052] In step S511, the first processed image is cropped and scaled to a fixed size, and then gamma correction is performed to adjust the image contrast and reduce the influence of lighting on the image, such as uneven lighting and local shadows, so that the overexposed or underexposed image is restored to normal and is closer to the image seen by the human eye.

[0053] The gamma correction formula is as follows:

[0054] f(x) = x γ

[0055] Where x represents the gray value of a pixel, and γ represents the gamma coefficient.

[0056] Step S512: Divide the second preprocessed image into several 8*8 small units and calculate the gradient histogram of each small unit;

[0057] In step S512, the horizontal and vertical gradients are calculated using the kernel matrix.

[0058] Step S513: Normalize the histogram using a 2*2 sliding window to obtain the normalized result;

[0059] Step S514: Extract feature vectors from the normalization result, match the feature vectors with the standard model, and obtain the matching result image.

[0060] In step S514, the feature vector includes magnitude and direction;

[0061]

[0062]

[0063] Where g represents the amplitude, g x G represents the horizontal gradient. y Θ represents the vertical gradient, and Θ represents the direction angle.

[0064] In another embodiment of the invention, such as Figure 1 As shown, after determining whether the first processed result image contains a bright spot with a target gray value and obtaining the detection result, the detection method further includes step S600: storing the detection result in a database and displaying it visually.

[0065] In this embodiment, the detection results are used to annotate the image with necessary information and write the corresponding data into the database. The corresponding data includes: location information, warning information, etc. The corresponding time sequence diagram and spatial diagram are drawn based on the pole number information stored in the database for visualization display.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for detecting arcing, characterized in that, The detection method includes: When arcing occurs, it receives the arcing signal emitted by the arcing sensor; An arc image is obtained based on the arc signal, and the arc image is preprocessed to obtain a first preprocessed image; The first target region image is obtained by performing ROI cropping on the first preprocessed image, and the first difference result is obtained by performing differential processing on two consecutive frames of the first target region image. The first difference result and the first target threshold are binarized to obtain the first processed result image; Determine whether the first processed image contains a bright spot with the target grayscale value to obtain a detection result; the detection result includes: If the first processed image contains a bright spot with the target gray value, it is determined that a visible light arc has occurred; If the first processed result image does not contain a bright spot with the target gray value, model matching is performed on the first processed result image to obtain a matching result image. The matching result image is cropped by ROI to obtain the second target region image, and the difference processing is performed on two consecutive frames in the second target region image to obtain the second difference result; The second difference result and the second target threshold are binarized to obtain the second processed result image; The detection result is obtained by determining whether the image of the second processed result contains a bright spot with the target gray value.

2. The arc detection method as described in claim 1, characterized in that, The step of determining whether the second processed image contains bright spots with the target grayscale value to obtain the detection result includes: If the image of the second processing result contains a bright spot with the target gray value, it is determined that a visible light arc has occurred; If the image of the second processing result does not contain any bright spots of the target grayscale value, it is determined that no visible light arc has occurred.

3. The arc detection method as described in claim 1, characterized in that, Model matching is performed on the first processed image to obtain a matching result image, including: Model matching is performed on the bus region in the first processed image.

4. The arc detection method as described in claim 1 or 3, characterized in that, The process of performing model matching on the first processed image to obtain a matching result image further includes: The first processed image is cropped and gamma-corrected to obtain a second preprocessed image; The second preprocessed image is divided into several 8*8 small units, and the gradient histogram of each small unit is calculated. The histogram is normalized using a 2x2 sliding window to obtain the normalized result; Feature vectors are extracted from the normalization results, and the feature vectors are matched with the standard model to obtain the matching result image.

5. The arc detection method as described in claim 4, characterized in that, The gamma correction formula is as follows: in, Represents the grayscale value of a pixel. This represents the gamma coefficient.

6. The arc detection method as described in claim 4, characterized in that, The calculation of the gradient histogram for each small unit includes: The horizontal and vertical gradients are calculated using the kernel matrix.

7. The arc detection method as described in claim 6, characterized in that, The feature vector includes magnitude and direction; in, Indicates amplitude. Represents the horizontal gradient. Represents the vertical gradient. Indicates direction and angle.

8. The arc detection method as described in claim 1, characterized in that, After determining whether the first processed image contains bright spots with the target grayscale value and obtaining the detection result, the detection method further includes: The detection results are stored in a database and then visualized.

Images