A direct current arc detection method, a detection system and a storage medium

Abstract

The application discloses a direct-current arc detection method and system and a storage medium, and the method comprises the following steps: fixing two cameras to form a binocular camera, and acquiring parameters of the binocular camera; a master control unit configures the parameters of the binocular camera, synchronizes the time of acquiring image frames of the binocular camera, and performs shooting; the master control unit processes the shot images, forms an event camera, and outputs event frames; whether the event frames are higher than a behavior trigger condition is judged, and the event frames are transmitted to an upper computer; the upper computer performs image preprocessing and feature point extraction on the arc occurrence event frames, determines the distance between the arc occurrence position and the binocular camera, and then performs three-dimensional space re-projection to obtain the spatial position of the arc feature point relative to the camera, and the arc length is obtained by conversion; the upper computer judges the occurrence of the arc, gives a warning, and performs an arc extinguishing operation. The application effectively highlights the arc event, reduces the data volume and equipment cost of the core detection algorithm, improves the processing speed and effect, expands the detection range, and realizes arc generation warning.

Description

Technical Field

[0001] This application belongs to the field of power electronics technology, specifically relating to a DC arc detection method, detection system, and storage medium. Background Technology

[0002] An electric arc is a powerful discharge phenomenon that occurs in air or other insulating media, typically accompanied by high temperature, intense light, and a strong current. In most modern electrical appliances, it is an accidental and potentially harmful discharge. In DC systems, once an arc forms, its voltage remains non-zero for an extended period, making it more likely to persist compared to arcs in AC systems where the voltage may cross zero, leading to potential risks such as high-temperature fires. Therefore, effective arc detection is crucial to minimizing its impact.

[0003] Arc detection methods often rely on changes in electrical characteristics for judgment. However, this approach can lead to false alarms when power changes are large. In complex electrical scenarios, it is difficult to pinpoint the exact location of the arc, and it cannot provide information about the specific on-site conditions at the time of arc generation or provide early warnings of the derivative hazards caused by the arc. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an optical-based DC arc detection method, detection system, and storage medium capable of precisely locating the arc generation position.

[0005] The present invention adopts the following technical solution.

[0006] The first aspect of this invention discloses a method for detecting a DC electric arc, comprising the following steps:

[0007] Two cameras are fixed at a fixed point at a distance from the location to be monitored for electric arc, forming a binocular camera system. The intrinsic and extrinsic parameters of the binocular camera are acquired. The parameters of the binocular camera are configured through the main control unit, and the acquisition time of the image frames by the binocular camera is synchronized and the images are captured.

[0008] The main control unit processes the captured images to obtain a temporally continuous regular sequence. By subtracting two consecutive frames in the sequence, an event-like camera is formed and an event frame is output. It is determined whether the event frame exceeds the behavior triggering condition. If so, the event frame is saved as an arc occurrence event frame and transmitted to the host computer.

[0009] The host computer performs image preprocessing on the arc occurrence event frame and extracts feature points to determine the distance between the arc occurrence location and the binocular camera. Based on the determined distance, it performs three-dimensional spatial reprojection to obtain the spatial position of the arc feature points relative to the camera and calculates the arc length.

[0010] The host computer combines the behavior triggering conditions of the arc event occurrence frame with the arc length to determine the occurrence of an arc, issue an early warning and execute the corresponding arc extinguishing operation, and save the image of the arc occurrence.

[0011] Preferably, the fixed point is located on the perpendicular bisector of the line connecting the two cameras. The position of the lens and camera is adjusted so that the position to be monitored for the electric arc is completely within the shooting range, and the position is centered and fills the shooting range.

[0012] Preferably, the acquisition of the intrinsic and extrinsic parameters of the binocular camera is achieved by obtaining the intrinsic and extrinsic parameters of each binocular camera through a calibration device and a calibration algorithm.

[0013] Preferably, the output event frame is generated by the main control unit buffering multiple frames of images acquired by the two cameras in the order of shooting time, so that the captured images form a temporally continuous regular sequence. By subtracting two consecutive frames in the regular sequence, the anomalies of the two frames are extracted to form an event-like camera, and the event frame is output.

[0014] Preferably, the behavior triggering condition is as follows: by performing histogram statistics on the event frames, a triggering threshold is set. When the event frame is higher than the triggering threshold, the event frame is saved as an arc occurrence event frame and transmitted to the host computer.

[0015] Preferably, the trigger threshold can be either a brightness threshold or a pixel count threshold, and the behavior trigger condition is when the brightness or pixel count of the event frame is higher than the threshold.

[0016] Preferably, the host computer performs image preprocessing on the arc occurrence event frames by using the intrinsic parameters to preprocess the arc occurrence event frames acquired by the two cameras to obtain a binocular vision image.

[0017] The second aspect of the present invention discloses a DC arc detection system that performs the DC arc detection method described in the first aspect, comprising: a camera, a main control unit, and a host computer;

[0018] Two cameras are fixed at a set distance from the location to be monitored by a mounting plate, forming a binocular camera. The signal terminals of the two cameras are connected to the main control unit via data cables, and the main control unit is connected to the host computer via data cables.

[0019] A third aspect of the present invention discloses a terminal, including a processor and a storage medium; the storage medium is used to store instructions;

[0020] The processor is configured to operate according to the instructions to perform the steps of the method according to the first aspect.

[0021] A fourth aspect of the present invention discloses a computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the steps of the method described in the first aspect.

[0022] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0023] This invention solves the problem of expensive or low-resolution cameras used in general visual arc detection by simulating an ordinary RGB camera as an event-like camera. The event-like camera effectively highlights arc events, reduces the data volume and equipment cost of the core detection algorithm, improves processing speed and detection effect, and realizes the monitoring and acquisition of non-electrical physical features such as arc length through binocular vision, expanding the detection range and realizing early warning of arc generation. It not only improves detection accuracy, but also has the feature of flexible component configuration. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the components of the present invention;

[0025] Figure 2 This is a flowchart of the method of the present invention;

[0026] Figure 3 This is a schematic diagram of the algorithm of this invention;

[0027] In the diagram: 1. Fixture; 2. First camera; 3. Second camera; 4. Main control unit; 5. Host computer. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of this invention. The embodiments described in this application are merely some embodiments of this invention, and not all embodiments. Based on the spirit of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of this invention.

[0029] Embodiment 1 of the present invention provides an optical-based DC arc detection method, comprising the following steps:

[0030] Step 1: Fix two cameras at a fixed point at a set distance from the location to be monitored for electric arc, forming a binocular camera, and acquire the intrinsic and extrinsic parameters of each binocular camera.

[0031] Step 1.1: Fix a pair of cameras at a suitable distance from the location to be monitored to form a binocular camera, and make the fixing point lie on the perpendicular bisector of the line connecting the two cameras. Adjust the lens and camera positions so that the location to be monitored can be clearly and completely captured by the camera.

[0032] In a preferred but non-limiting embodiment, in order to form a binocular camera, two originally independent cameras need to be fixed at a fixed distance and in the same direction to form a binocular camera. By changing the lens, the parameters such as the lens focal length and field of view of the binocular camera and the distance relative to the monitored area are adjusted so that the monitored area can be completely within the shooting range, and the position is centered and fills the shooting range.

[0033] Step 1.2: Obtain the intrinsic and extrinsic parameters of each binocular camera using a calibration device and calibration algorithm;

[0034] The intrinsic parameters include: focal length, principal point, scaling factor, distortion parameters, etc.; the extrinsic parameters include: the distance between the binocular cameras, camera attitude, etc.

[0035] In a preferred but non-limiting embodiment, the calibration device is not limited to checkerboard, dot calibration plate and misaligned origin calibration plate, but also includes other plate-like objects with easily identifiable repeating high-precision geometric array patterns on a certain substrate.

[0036] More preferably, the calibration algorithm can be Zhang Zhengyou's calibration method, Tsai and Huang's calibration method, checkerboard calibration method, or any one of the relevant calibration algorithm libraries such as OpenCV, Matlab, and Halcon.

[0037] Step 2: Configure the relevant parameters of the camera through the main control unit, synchronize the acquisition of image frame times by the two cameras and take pictures.

[0038] Specifically, the main control unit configures camera parameters such as frame rate, exposure, and sensitivity, and adjusts the position and optical lens of the binocular cameras so that the two cameras can simultaneously, accurately, completely, and without overexposure or severe distortion capture arc images. Then, the binocular cameras synchronize the acquisition of image frame times so that the images captured by the binocular cameras are at the same moment.

[0039] In a preferred but non-limiting embodiment, adjusting the position of the binocular camera in step 2 is a further adjustment to step 1.1 until the best shooting effect is achieved. Specifically, it involves fine-tuning the focus and position so that the binocular camera can capture a clear and complete image of the detected area.

[0040] The distortion is related to the lens parameters used. The distortion can be observed visually. When there is not much difference between what the camera captures and what the human eye sees, the distortion is small. If the human eye sees something significantly different, the distortion is large.

[0041] Step 3: The main control unit processes the images acquired by each camera to obtain a temporally continuous regular sequence. By subtracting two consecutive frames in the sequence, an event-like camera is formed and an "event frame" is output. It is determined whether the "event frame" is higher than the behavior trigger condition. If so, the "event frame" is saved as an arc occurrence event frame and transmitted to the host computer.

[0042] The main control unit buffers multiple frames of images acquired by each camera in chronological order of shooting time, so that the captured images form a regular sequence that is continuous in time.

[0043] Next, by subtracting two consecutive frames in the regular sequence, the outliers of the two frames are extracted to form an event-like camera and output an "event frame".

[0044] It is worth noting that the aforementioned event camera, as defined in this invention, is a type of ordinary RGB camera that uses an algorithm to simulate the working characteristics of an event camera.

[0045] The behavior triggering condition is set by performing histogram statistics on the "event frame" and setting a trigger threshold. When the "event frame" is higher than the trigger threshold, the "event frame" is saved as an arc occurrence event frame and transmitted to the host computer.

[0046] The trigger threshold can be a brightness threshold or a pixel count threshold. When the brightness of the event frame is higher than a threshold, or when the number of pixels is higher than a threshold, it is used as a trigger condition.

[0047] Generally, the brightness range of an image captured by a camera should be between 0 and 255. This range can be set a priori, such as manually generating an electric arc within the image brightness range and extracting its brightness; or by using binarization algorithms such as the maximum inter-class method to calculate whether the separation threshold is significantly greater than the previous frame.

[0048] Step 4: The host computer performs image preprocessing on the arc occurrence event frames acquired by the binocular camera to obtain a binocular visual image and extract feature points. It determines the distance between the arc occurrence location and the binocular camera. Based on the determined distance, it performs three-dimensional spatial reprojection to obtain the spatial position of the arc feature points relative to the camera and calculates the arc length.

[0049] Step 4.1: The host computer uses the intrinsic parameters obtained in Step 1 to preprocess the arc event frames acquired by each camera to obtain an ideal binocular vision image. For the processed binocular vision image, feature matching algorithms such as ORB and SIFT are used to match features and extract the positional differences of feature points between the two frames.

[0050] In a preferred but non-limiting embodiment, the image preprocessing includes, but is not limited to, any one or more of the following: distortion correction, epipolar correction and filtering, sharpening, binarization, and morphological processing.

[0051] Step 4.2: Using the extrinsic parameters obtained in Step 1, namely the calibrated binocular camera spacing and attitude parameters, and the positional differences of the feature points obtained in Step 4.1, depth calculation is performed to determine the distance from the arc occurrence location to the binocular camera. Then, using the determined distance, three-dimensional spatial reprojection is performed to obtain the spatial position of the arc feature points relative to the camera, and then the arc length is calculated.

[0052] In a preferred but non-limiting embodiment, during the conversion, the electric arc can be considered to consist of a series of inflection points and straight lines connecting the inflection points one by one. In the previous step, a three-dimensional spatial reprojection was performed, which can obtain the three-dimensional coordinate position data of each inflection point and corner point relative to the camera position in the actual space. Then, calculations can be performed as needed, such as calculating the arc length of the main and branch arcs, the distance between the two ends of the discharge, etc.

[0053] Step 5: The host computer combines the arc brightness obtained from the arc event frame in Step 3 and the arc length obtained in Step 4 to determine the occurrence of an arc, issue an early warning and execute the corresponding arc extinguishing operation, and save the image of the arc generation.

[0054] like Figure 2 As shown, Embodiment 2 of the present invention provides an optical-based DC arc detection method. Based on Embodiment 1, a preferred embodiment of the DC arc detection method is described in detail using a specific arc detection in an engineering experiment as an example.

[0055] Specifically, the optical-based DC arc detection method includes the following steps:

[0056] In terms of hardware composition:

[0057] Regarding the camera: Lens options include two MVL-MF1224M-5MPE lenses; Camera body options include two MV-CS004-11GC lenses; The main control unit options include a ZYNQ7100-based main control board with two Gigabit Ethernet ports, one PCIe controller, and one synchronization signal generator.

[0058] The host computer can be a personal desktop PC. The configuration used in this case is as follows: CPU: i7-14700K, graphics card: RTX4080, RAM: 64GB, ROM: 512G (SSD) + 4TB (HDD).

[0059] There are no special requirements for the mounting plate.

[0060] In terms of algorithms, such as Figure 3 As shown:

[0061] By fixing two cameras with synchronous triggering function at a suitable distance on the same flat bracket or plane, ensuring that the optical axes of the two cameras are parallel and shooting in the same direction, and adjusting the lens parameters and the position of the bracket, the monitored area can be clearly imaged on both cameras without bias.

[0062] Next, a camera calibration board with a periodically occurring, high-precision, high-contrast, and simple black-and-white image, such as a dotted calibration board, is selected and placed in front of the camera. The orientation of the calibration board is changed by translation or rotation, and two cameras are used to clearly capture complete images of the calibration board in different orientations. Then, Zhang Zhengyou's calibration method is used to calibrate the camera on these images, obtaining the camera's intrinsic and extrinsic parameters and distortion model matrix.

[0063] The focal length fl of the two cameras is obtained through internal parameters.

[0064] The extrinsic parameter is used to obtain a relatively accurate baseline distance B:

[0065] The calibration image pairs are taken by different cameras at the same position and in the same orientation of the calibration board. Each pair has corresponding extrinsic parameters (Rl, Tl) and (Rr, Tr). The subscripts l and r represent the left and right cameras, respectively, and are only used to distinguish between different cameras. In actual deployment, the left-right relationship may not be used. R is the rotation matrix, which describes the rotation state, and T is the translation matrix, which describes the relative position of the calibration board and the camera.

[0066] The rotation matrices Rc and Tc of camera r relative to camera l are calculated and given by the following formulas:

[0067]

[0068] Since the optical axes of actual cameras are not perfectly parallel, it is necessary to calculate the epipolar correction matrix Rp to obtain an image with ideally parallel optical axes. The formula is as follows:

[0069]

[0070] Where R rect r l and r r It is derived from the following formula:

[0071] R rect =[e1e2e3] T

[0072]

[0073] e1e2e3 are given by the following equations:

[0074]

[0075] Using Tl and Tr, the baseline distance B between them can be obtained using the following formula:

[0076] B = |R c T l -T r |

[0077] Next, the main control unit sends a synchronized shooting trigger signal to control both cameras to take pictures in real time and transmit the images back to the main control unit. At this time, the main control unit extracts event frames, and the steps include:

[0078] The original images are temporarily stored and each forms a 1-second FIFO image sequence. Then, the images are grayscaled, which also forms a 1-second grayscale FIFO image sequence. The spatial pixels of two adjacent frames are subtracted to extract the difference between the two adjacent frames, which is the generation of the "like" event frame.

[0079] Next, grayscale histogram statistics are performed on the frame, and the grayscale image is binarized using the maximum inter-class method to obtain the separation threshold. If the images captured by both cameras meet the condition that the separation threshold is significantly greater than that of the previous frame, it is considered that an electric arc has occurred. At this point, the next step of electric arc parameter extraction is performed.

[0080] First, distortion correction is performed on the respective grayscale arc event frames. The ideal image is calculated using the distortion model matrix of each camera mentioned above. Then, epipolar correction is performed on the respective grayscale arc event frames using the epipolar correction matrix calculated from the calibration parameters. Next, Gaussian sharpening and maximum inter-class binarization are performed to extract the arc, followed by a closing operation to ensure arc continuity. Then, a skeleton search algorithm and the Harris corner algorithm are used to extract the topological structure of each branch point and feature inflection point, and feature point matching is performed to form feature point pairs tl(ul,vl) and tr(ur,vr). Finally, the next step, depth calculation, is performed.

[0081] The depth calculation includes, firstly, performing disparity calculation on feature point pairs: calculating the coordinate distance Di of feature point pair i in pixel coordinates for each pair. Since epipolar correction has already been performed, this can be degenerated into:

[0082] D i =ur i -ul i (The difference in pixel coordinates u between camera r and camera l)

[0083] Based on the focal length fl obtained above, the depth S is calculated using the following formula. i :

[0084] S i = fl * B / D i

[0085] Next, calculate the spatial position of each feature point i, i.e., the 3D reprojection. First, obtain the center pixel position (cx, cy) of camera l. The spatial position of each feature point i relative to camera l can be obtained using the following formula:

[0086] X i =Si*(ul i -cx) / f

[0087] Y i =Si*(vl) i -cy) / f

[0088] Z i =Si

[0089] Since an electric arc can be considered to consist of a series of inflection points and straight lines connecting these inflection points, and the spatial position information of each support point and corner point is calculated in the previous step, the spatial characterization of the electric arc and the calculation of required information can be performed. For example, the distance Ed between the two ends of the discharge, A and B (with coordinate vectors EA = (xa, ya, za) and EB = (x2, y2, z2) respectively), can be calculated using the following formula:

[0090] Ed=|EA-EB|

[0091] Since the data frame of the electric arc and its corresponding time have already been obtained, this time can not only be used to send an electric arc alarm, but also to capture and retain the original image, and for other purposes such as: extracting and further analyzing optical information such as arc brightness and arc halo; and monitoring the on-site situation after the electric arc is generated. It can also connect to other physical acquisition devices to retain desired data; for example, a microphone can be added to retain the sound information when the electric arc is generated.

[0092] Embodiment 3 of the present invention provides an optical-based DC current detection system, including: a camera, a main control unit, and a host computer.

[0093] Two cameras are fixed at a known distance using a mounting plate, and optical lenses with identical parameters are installed according to the actual scene requirements, making them symmetrical binocular cameras. The signal terminals of the two cameras are respectively connected to the same main control unit and the host computer.

[0094] Specifically, such as Figure 1 As shown, the first camera 2 and the second camera 3 are fixed at a known distance by the fixing member 1. The first camera 2 and the second camera 3 are each connected to the main control unit 4 by a data cable. The main control unit 4 is connected to the host computer 5 by a data cable.

[0095] In a preferred but non-limiting embodiment of the present invention, the fixing plate 1 is not strictly limited and can be either rod-shaped or plate-shaped, as long as a certain method can be used to prevent the two cameras from undergoing relative displacement.

[0096] More preferably, the camera can be any one of a regular RGB camera, an infrared camera, a multispectral camera, or an event camera; at the same time, the two cameras can also be composed of asymmetrical cameras, that is, the lenses and camera parameters are not exactly the same.

[0097] The main control unit 4 and the host computer 5 mainly perform the following functions: camera parameter configuration, frame synchronization, frame stacking, distortion correction, event preprocessing, arc extraction, feature matching, depth calculation, 3D reprojection, and physical feature calculation.

[0098] In a preferred but non-limiting embodiment of the present invention, the main control unit 4 and the host computer 5 are not strictly distinguished, and the main control unit 4 and the host computer 5 can be combined into one device.

[0099] Embodiment 4 of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is loaded onto the processor, it implements an optical-based DC arc detection method according to Embodiment 1.

[0100] Embodiment 5 of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements an optical-based DC arc detection method according to Embodiment 1.

[0101] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0102] This invention addresses the problem of expensive or low-resolution cameras used in conventional visual arc detection by simulating a regular RGB camera as an event-like camera. The event-like camera effectively highlights arc events, reduces the data volume and equipment cost of the core detection algorithm, improves processing speed and detection results, and enables the monitoring and acquisition of non-electrical physical features such as arc length through binocular vision. This expands the detection range, enables early warning of arc occurrence, improves detection accuracy, and offers flexible component configuration.

[0103] This disclosure can be a system, method, and / or computer program product. A computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for causing a processor to implement various aspects of this disclosure.

[0104] Computer-readable storage media can be tangible devices capable of holding and storing instructions for use by an instruction execution device. Computer-readable storage media can be, for example—but not limited to—electrical storage devices, magnetic storage devices, optical storage devices, electromagnetic storage devices, semiconductor storage devices, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory sticks, floppy disks, mechanical encoding devices, such as punch cards or recessed protrusions storing instructions thereon, and any suitable combination of the foregoing. The computer-readable storage media used herein are not to be construed as transient signals themselves, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., light pulses through fiber optic cables), or electrical signals transmitted through wires.

[0105] The computer-readable program instructions described herein can be downloaded from computer-readable storage media to various computing / processing devices, or downloaded via a network, such as the Internet, local area network, wide area network, and / or wireless network, to an external computer or external storage device. The network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards them to the computer-readable storage media in the respective computing / processing device.

[0106] Computer program instructions used to perform the operations of this disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-dependent instructions, microcode, firmware instructions, status setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Smalltalk, C++, etc., and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The computer-readable program instructions may execute entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or may be connected to an external computer (e.g., via the Internet using an Internet service provider). In some embodiments, electronic circuitry, such as programmable logic circuitry, field-programmable gate arrays (FPGAs), or programmable logic arrays (PLAs), is personalized by utilizing the status information of the computer-readable program instructions to implement various aspects of this disclosure.

[0107] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for detecting DC arc, characterized in that, Includes the following steps: Two cameras are fixed at a fixed point at a distance from the location to be monitored for electric arc, forming a binocular camera system. The intrinsic and extrinsic parameters of the binocular camera are acquired. The parameters of the binocular camera are configured through the main control unit, and the acquisition time of the image frames by the binocular camera is synchronized and the images are captured. The main control unit processes the captured images. It performs multi-frame buffering on the images acquired by the two cameras in the order of shooting time, so that the captured images form a temporally continuous regular sequence. By subtracting two consecutive frames in the regular sequence, the anomaly points of the two frames are extracted to form an event-like camera and output an event frame. It determines whether the event frame exceeds the behavior triggering condition. If so, the event frame is saved as an arc occurrence event frame and transmitted to the host computer. The behavior triggering condition is as follows: by performing histogram statistics on event frames, a trigger threshold is set. When the event frame exceeds the trigger threshold, the event frame is saved as an arc occurrence event frame and transmitted to the host computer. The host computer performs image preprocessing on the arc occurrence event frame and extracts feature points to determine the distance between the arc occurrence location and the binocular camera. Based on the determined distance, it performs three-dimensional spatial reprojection to obtain the spatial position of the arc feature points relative to the camera and calculates the arc length. The host computer combines the behavior triggering conditions of the arc occurrence event frame with the arc length to determine the occurrence of an arc, issue an early warning and execute the corresponding arc extinguishing operation, and save the image of the arc occurrence.

2. The DC arc detection method according to claim 1, characterized in that, The fixed point is located on the perpendicular bisector of the line connecting the two cameras. The position of the lens and camera is adjusted so that the position to be monitored for the electric arc is within the shooting range, and the position is centered and fills the shooting range.

3. The DC arc detection method according to claim 1, characterized in that, The intrinsic and extrinsic parameters of the stereo camera are obtained by using a calibration device and a calibration algorithm to obtain the intrinsic and extrinsic parameters of each stereo camera.

4. The DC arc detection method according to claim 1, characterized in that, The trigger threshold can be either a brightness threshold or a pixel count threshold. When the brightness or pixel count of the event frame is higher than the threshold, it serves as a behavior trigger condition.

5. The DC arc detection method according to claim 1, characterized in that, The host computer performs image preprocessing on the arc occurrence event frames by using the intrinsic parameters to preprocess the arc occurrence event frames acquired by the two cameras to obtain a binocular vision image.

6. A DC arc detection system, comprising the DC arc detection method according to any one of claims 1-5, characterized in that, include: Camera, main control unit, and host computer; Two cameras are fixed at a set distance from the location to be monitored by a mounting plate, forming a binocular camera. The signal terminals of the two cameras are connected to the main control unit via data cables, and the main control unit is connected to the host computer via data cables.

7. A terminal, comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is configured to operate according to the instructions to perform the steps of the method according to any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the steps of the method according to any one of claims 1-5.

Images