Automatic measurement method and system for dangerous source distance in substation based on binocular ranging

Abstract

The present invention provides a method and system for automatically measuring the distance of dangerous sources in a substation based on binocular ranging, wherein the method comprises: step 1: obtaining a substation image in the substation taken by a binocular camera; step 2: training a dangerous source detection neural network model; step 3: determining the dangerous source in the substation based on the dangerous source detection neural network model and according to the substation image; step 4: determining the relative distance between the dangerous source in the substation and the shooting position according to the shooting characteristics of the dangerous source in the substation. The method and system for automatically measuring the distance of dangerous sources in a substation based on binocular ranging of the present invention introduces the substation image in the substation taken by a binocular camera, and at the same time, trains the dangerous source detection neural network model. According to the substation image and the dangerous source detection neural network model, the dangerous source in the substation is determined. According to the shooting characteristics of the dangerous source in the substation, the relative distance between the dangerous source in the substation and the shooting position is determined, thereby improving the perception efficiency of the dangerous source distance and enhancing safety.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer substation safety, and in particular to a method and system for automatically measuring the distance of a hazardous source in a transformer substation based on binocular ranging. Background Art

[0002] As an important part of the power system, substations have some potential sources of danger, such as high voltage danger, electromagnetic radiation, chemical substances and physical dangers. High voltage danger refers to the fact that substations process high voltage power signals, so there is a risk of electric shock and arc flash. High voltage lines and equipment may cause electric shock hazards to the human body, while arc flash may cause fire and explosion; electromagnetic radiation refers to the electromagnetic radiation generated by high voltage equipment and transmission lines in substations. Long-term exposure to high-intensity electromagnetic fields may have an impact on human health, including ionizing radiation and non-ionizing radiation; chemical substances refer to some chemical substances used in substations (such as insulating oil, coolant and cleaning agent) that may be harmful to human health. Improper handling, leakage or leakage may lead to chemical contact and poisoning risks; physical danger refers to the physical dangers of equipment and structures in substations (for example, the noise and vibration generated by the operation of large equipment may affect the hearing and health of workers, sharp edges in equipment and structures, high temperature surfaces and risks of falling or collision).

[0003] The invention patent with application number: CN202110381160.5 discloses an electronic fence area generation system based on hazard source monitoring and personnel positioning, wherein the system includes: a visual perception module, a data analysis module, a hidden danger assessment module and a fence dynamic generation module, which solves the problem that the existing fixed electronic fence cannot adjust the fence range in real time and accurately according to environmental conditions, has poor safety and high cost. The above invention conducts a comprehensive analysis of the state and trend of the hazard source, can effectively monitor a variety of major hazard sources, and reduce the incidence of danger; the virtual electronic fence generated by the above invention is organically and dynamically combined with the positioning technology, and automatic graded alarm is realized according to each heat point, and there is no need to manually arrange the virtual electronic fence, which is convenient and quick to set up, greatly reducing the cost and workload of security personnel. At the same time, it can adjust the effective area in real time according to the degree of danger, significantly improving the safety protection efficiency of dangerous areas.

[0004] However, when the above-mentioned prior art generates a virtual electronic fence, it is necessary to dynamically obtain the fence boundary information of the dangerous area before generating the electronic fence, which has a certain lag and a low efficiency in sensing danger, further increasing the danger.

[0005] In view of this, there is an urgent need for an automatic measurement method and system for the distance of hazardous sources in a substation based on binocular ranging to at least solve the above-mentioned deficiencies. Summary of the invention

[0006] One of the purposes of the present invention is to provide a method and system for automatically measuring the distance of dangerous sources in a substation based on binocular ranging, introduce a substation image taken by a binocular camera, and at the same time, train a neural network model for dangerous source detection. According to the substation image and the neural network model for dangerous source detection, the dangerous source in the station is determined. According to the shooting characteristics of the dangerous source in the station, the relative distance between the dangerous source in the station and the shooting position of the binocular camera is determined, thereby improving the perception efficiency of the dangerous source distance and enhancing safety.

[0007] The embodiment of the present invention provides a method for automatically measuring the distance of a hazardous source in a substation based on binocular ranging, comprising:

[0008] Step 1: Obtain the substation image inside the substation taken by the binocular camera;

[0009] Step 2: Training the hazard source detection neural network model;

[0010] Step 3: Based on the hazard source detection neural network model, determine the hazard sources in the substation according to the substation image;

[0011] Step 4: According to the shooting characteristics of the dangerous sources in the station, determine the relative distance between the dangerous sources in the station and the shooting position of the binocular camera.

[0012] Preferably, step 1: obtaining a substation image in the substation taken by a binocular camera, comprises:

[0013] Obtain the distribution of in-station equipment within the substation;

[0014] Determine the suspected locations of dangerous sources based on the distribution of equipment within the station;

[0015] Determine the drivable routes within the station;

[0016] According to the drivable routes and suspected dangerous source locations within the station, set the moving path of the mobile intelligent distance measuring device;

[0017] Control the mobile intelligent distance measuring device to move based on the moving path, and at the same time, establish a data transmission link between the mobile intelligent distance measuring device and the binocular camera preset on the mobile intelligent distance measuring device;

[0018] Based on the data transmission link, the substation image is obtained.

[0019] Preferably, the moving path of the mobile intelligent distance measuring device is set according to the drivable routes in the station and the suspected locations of the danger sources, including:

[0020] Get the shooting range of the binocular camera;

[0021] Performing path planning based on drivable routes within the station to obtain multiple first planned routes;

[0022] Based on the first planned route and the shooting range, determine the second planned route with a comprehensive inspection level for the suspicious points of the hazard source;

[0023] Determine whether there are repeated passes of suspicious points of dangerous sources on the second planned route;

[0024] The moving path is determined based on the existence judgment result of repeated passing of suspicious points of danger sources on the second planned route.

[0025] Preferably, determining the moving path based on the existence determination result of the repeated passing of the suspicious point of the danger source on the second planned route includes:

[0026] If the existence determination result is non-existence, the corresponding second planned route will be used as the third planned route;

[0027] Obtain the first execution duration of the third planned route;

[0028] If the existence determination result is existence, the corresponding second planned route is used as the fourth planned route;

[0029] Obtain the second execution time of the fourth planned route;

[0030] According to the repeated passing situation, determine the number of times the suspected dangerous source point is passed;

[0031] Obtain inspection priority for suspected hazardous source locations;

[0032] According to the number of passes and the inspection priority, the reasonableness of the number of passes is determined, and based on the reasonableness, the second execution time is adjusted downward to obtain a third execution time;

[0033] The first execution duration is compared with the third execution duration, and the moving paths in the third planned route and the fourth planned route are determined according to the comparison result of the duration.

[0034] Preferably, step 2: training a hazard source detection neural network model comprises:

[0035] Obtain historical operation records of substations;

[0036] Determine potential sources of danger based on historical operating records;

[0037] Get the substation upload database;

[0038] Obtain relevant images of potential danger sources based on the substation upload database;

[0039] Based on the yolov8 target model and the associated images, the neural network model for hazard source detection is trained.

[0040] Preferably, potential sources of danger are determined based on historical operation records, including:

[0041] Obtain the device operation status recorded in the historical operation record;

[0042] According to the device operation status, determine the first hazard source in the historical operation records of abnormal device operation status;

[0043] Obtain personnel accident events recorded in historical operation records;

[0044] According to the personnel accident event, determine the second hazard source that caused the personnel accident event;

[0045] Treat the first and second hazard sources together as potential hazard sources;

[0046] Wherein, according to the operation status of the device, determining the first hazard source in the historical operation record of abnormal operation status of the device includes:

[0047] Extract the dialogue flow in the historical operation record of abnormal operation status of the device, and determine the first hazard source based on the dialogue flow;

[0048] Among them, according to the personnel accident event, the second hazard source that caused the personnel accident event is determined, including:

[0049] Obtain the accident source attribution model, and determine the second hazard source based on the personnel accident events and the accident source attribution model.

[0050] Preferably, step 4: determining the relative distance between the hazard source in the station and the shooting position of the binocular camera according to the shooting characteristics of the hazard source in the station, comprises:

[0051] Calculate the parallax of feature points based on the shooting features;

[0052] Determine the horizontal offset according to the parallax of the feature points;

[0053] Measure the binocular physical distance between binocular cameras;

[0054] According to the binocular physical distance, determine the distance measurement error calculation coefficient;

[0055] Get the focal length parameters of the binocular camera;

[0056] According to the binocular physical distance, focal length parameters, horizontal offset and ranging error calculation coefficient, the relative distance is determined. The calculation formula of the relative distance is:

[0057]

[0058] Where D is the relative distance, dis i is the i-th horizontal offset, Base is the binocular physical distance, fli is the i-th focal length parameter, I is the total number of horizontal offsets, exp is the ranging error calculation coefficient function, and exp(r) is the ranging error calculation coefficient.

[0059] The method for automatically measuring the distance of a hazardous source in a substation based on binocular ranging provided by an embodiment of the present invention further includes:

[0060] Obtain the number of binocular cameras, and determine the target triggering camera based on the number of cameras;

[0061] Wherein, based on the number of cameras, determining the target triggering camera includes:

[0062] When the number of cameras is equal to 1, the corresponding binocular camera is used as the target trigger camera;

[0063] When the number of cameras is greater than 1, at least one camera to be triggered in the binocular cameras is determined based on the trigger mechanism, and an evaluation result of the camera to be triggered is obtained based on a preset ranging accuracy evaluation system;

[0064] According to the evaluation results, the target is determined to trigger the camera.

[0065] Preferably, determining at least one camera to be triggered in the binocular camera, and obtaining an evaluation result of the camera to be triggered based on a preset distance measurement accuracy evaluation system, includes:

[0066] Determine the required focal length range based on the shooting characteristics;

[0067] Determine the application focal length range based on the binocular physical distance;

[0068] Determine the camera to be triggered according to the required focal length range and the application focal length range;

[0069] Obtain the pre-shot image of the camera to be triggered;

[0070] Based on the preset distance measurement accuracy evaluation system, the evaluation result of the camera to be triggered is obtained according to the pre-shot image.

[0071] The embodiment of the present invention provides an automatic measurement system for the distance of dangerous sources in a substation based on binocular ranging, comprising:

[0072] The substation image acquisition subsystem is used to acquire the substation image in the substation taken by the binocular camera;

[0073] Model training subsystem, used to train the hazard source detection neural network model;

[0074] The subsystem for determining hazardous sources within the station is used to determine hazardous sources within the station based on the hazard source detection neural network model and the substation image;

[0075] The relative distance determination subsystem is used to determine the relative distance between the hazardous source in the station and the shooting position of the binocular camera according to the shooting characteristics of the hazardous source in the station.

[0076] The beneficial effects of the present invention are:

[0077] The present invention introduces substation images taken by binocular cameras, and at the same time, trains a neural network model for hazard source detection. According to the substation images and the neural network model for hazard source detection, the hazard source in the substation is determined. According to the shooting characteristics of the hazard source in the substation, the relative distance between the hazard source in the substation and the shooting position of the binocular camera is determined, thereby improving the perception efficiency of the hazard source distance and enhancing safety.

[0078] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures specifically pointed out in the present application documents.

[0079] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0081] Figure 1 Schematic diagram of an automatic measurement method of the distance to a hazardous source in a substation based on binocular ranging in an embodiment of the present invention;

[0082] Figure 2 It is a schematic diagram of an automatic measurement system for the distance to hazardous sources in a substation based on binocular ranging in an embodiment of the present invention. DETAILED DESCRIPTION

[0083] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0084] The embodiment of the present invention provides a method for automatically measuring the distance of dangerous sources in a substation based on binocular ranging, such as Figure 1 As shown, including:

[0085] Step 1: Obtain a substation image in the substation taken by a binocular camera; wherein the substation image is: an image in the substation taken by a binocular camera in the substation where hazard source detection is required;

[0086] Step 2: training a hazard source detection neural network model; wherein the hazard source detection neural network model is: a target detection model for hazard source detection;

[0087] Step 3: Based on the danger source detection neural network model, determine the danger source in the substation according to the substation image; wherein the danger source in the substation is: the danger source in the substation image detected by the danger source detection neural network model;

[0088] Step 4: According to the shooting characteristics of the dangerous source in the station, determine the relative distance between the dangerous source in the station and the shooting position of the binocular camera. The shooting characteristics are: the image characteristics of the image taken by the left camera of the binocular camera and the image characteristics of the image taken by the right camera of the binocular camera. The image characteristics are the characteristic representation of the substation image, including color, texture and shape, etc. The shooting position is: the position of the binocular camera; the relative distance is: the distance between the dangerous source in the station and the shooting position.

[0089] The working principle and beneficial effects of the above technical solution are:

[0090] This application introduces substation images taken by binocular cameras, and at the same time, trains a neural network model for hazard source detection. According to the substation images and the neural network model for hazard source detection, the hazard source in the station is determined. According to the shooting characteristics of the hazard source in the station, the relative distance between the hazard source in the station and the shooting position of the binocular camera is determined, which improves the perception efficiency of the hazard source distance and improves safety.

[0091] In one embodiment, step 1: obtaining a substation image in the substation taken by a binocular camera, includes:

[0092] Obtaining the distribution of in-station equipment in the substation; wherein the in-station equipment distribution includes: the types of equipment in the substation and the distribution locations of the equipment;

[0093] Determine the suspected danger source points according to the distribution of equipment in the station; the suspected danger source points are: the substation equipment locations where the danger source may exist, determine whether there is a danger possibility according to the type of equipment, and if there is a danger possibility, the corresponding equipment distribution location is regarded as the suspected danger source point;

[0094] Determine the drivable routes within the station; wherein, the drivable routes within the station are: the passage routes that can be driven within the substation;

[0095] According to the drivable routes and suspected dangerous source locations within the station, set the moving path of the mobile intelligent distance measuring device;

[0096] Control the mobile intelligent distance measuring device to move based on the moving path, and at the same time, establish a data transmission link between the mobile intelligent distance measuring device and the binocular camera preset on the mobile intelligent distance measuring device; wherein the mobile intelligent distance measuring device is: a distance measuring device equipped with a binocular camera; the data transmission link is: a communication link between a central processor and the binocular camera;

[0097] Based on the data transmission link, the substation image is obtained.

[0098] The working principle and beneficial effects of the above technical solution are:

[0099] When a binocular camera is set on a mobile intelligent ranging device to perform real-time detection of hazardous sources in a substation, it is necessary to plan the driving route of the mobile intelligent ranging device. Therefore, this application obtains the distribution of in-station equipment in the substation and determines the suspected locations of hazardous sources. Obtain the drivable routes in the station, set the moving path of the mobile intelligent ranging device according to the drivable routes in the station and the suspected locations of hazardous sources, and control the mobile intelligent ranging device to move based on the moving path. During movement, establish a data transmission link between the central processing unit and the mobile intelligent ranging device, obtain the substation image based on the data transmission link, and introduce a mobile intelligent ranging device to obtain the substation image, thereby improving the convenience of the substation image.

[0100] In one embodiment, according to the drivable routes in the station and the suspected locations of dangerous sources, the moving path of the mobile intelligent distance measuring device is set, including:

[0101] Get the shooting range of the binocular camera; where the shooting range is: the shooting range of the binocular camera;

[0102] Performing path planning based on the drivable routes within the station to obtain multiple first planned routes; wherein the first planned route is: a route passing through a suspected point of a dangerous source planned based on the path planning technology;

[0103] Based on the first planned route and the shooting range, a second planned route is determined in which the inspection comprehensiveness of the suspicious dangerous source points is comprehensive; wherein the inspection comprehensiveness is the result obtained by dividing the number of suspicious dangerous source points that can be photographed by the number of all suspicious dangerous source points; the inspection comprehensiveness is comprehensive when each suspicious dangerous source point is photographed;

[0104] Determine whether there is a repeated passing of the suspected danger source point on the second planned route; wherein the repeated passing situation is: one or more suspected danger source points on the second planned route are photographed more than once;

[0105] Based on the existence determination result of the repeated passing of the suspicious points of the hazard source on the second planned route, the moving path is determined. The existence determination result is: existence or non-existence; the moving path is: the driving route of the mobile intelligent ranging device in the substation.

[0106] The working principle and beneficial effects of the above technical solution are:

[0107] This application obtains the shooting range of the binocular camera and introduces path planning technology to plan the first planning route. According to the first planning route and the shooting range, the first planning route with comprehensive inspection is determined and used as the second planning route. The repeated passing of suspicious points of the hazard source is obtained, and the moving path is determined according to the different existence judgment results, which improves the rationality of the moving path formulation.

[0108] In one embodiment, based on the existence determination result of the repeated passing of the suspicious point of the danger source on the second planned route, determining the moving path includes:

[0109] If the existence determination result is non-existence, the corresponding second planned route will be used as the third planned route;

[0110] Obtaining a first execution time of the third planned route; wherein the first execution time is: the time required for the mobile intelligent distance measuring device to complete one driving of the third planned route;

[0111] If the existence determination result is existence, the corresponding second planned route is used as the fourth planned route;

[0112] Obtaining a second execution time of the fourth planned route; wherein the second execution time is: the time required for the mobile intelligent ranging device to complete one driving of the fourth planned route;

[0113] According to the repeated passing situation, the number of times the suspicious point of the danger source is passed is determined; wherein the number of passing is: the number of times the suspicious point of the danger source on the fourth planned route is photographed, for example: twice;

[0114] Obtain the inspection priority of the suspected hazardous source points; the inspection priority is determined according to the frequency of hazardous sources being discovered at the suspected hazardous source points in history. The higher the frequency of hazardous sources being discovered, the higher the corresponding inspection priority.

[0115] According to the number of passes and the inspection priority, the reasonableness of the number of passes is determined, and based on the reasonableness, the second execution time is adjusted downward to obtain the third execution time; wherein, when determining the reasonableness of the number of passes according to the number of passes and the inspection priority, the theoretical number of passes is determined according to the inspection priority, and the comparison difference and value of all suspicious points of dangerous sources are determined by comparing the theoretical number of passes and the corresponding number of passes, and the reasonableness is determined according to the comparison number difference and value, and the larger the comparison difference and value, the lower the reasonableness; the downward adjustment rule for adjusting the second execution time downward based on the reasonableness is manually preset;

[0116] The first execution duration is compared with the third execution duration, and the moving paths in the third planned route and the fourth planned route are determined according to the comparison result of the durations. The comparison result is: the first execution duration is greater than or equal to the third execution duration and the first execution duration is less than the third execution duration; when determining the moving paths in the third planned route and the fourth planned route: when the first execution duration is greater than or equal to the third execution duration, the fourth planned route is used as the moving path, and when the first execution duration is less than the third execution duration, the third planned route is used as the moving path.

[0117] The working principle and beneficial effects of the above technical solution are:

[0118] The present application obtains a first execution duration of a third planned route that does not have any suspected hazardous source points that are passed through repeatedly, and at the same time, obtains a second execution duration of a fourth planned route that has any suspected hazardous source points that are passed through repeatedly, determines the number of times the suspected hazardous source points are passed through based on the repeated passing situation, determines the reasonableness of the number of times of passing through based on the number of passing through and the inspection priority of the suspected hazardous source points, lowers the second execution duration based on the reasonableness, obtains the third execution duration, compares the first execution duration with the third execution duration, uses the fourth planned route as the moving path when the first execution duration is greater than or equal to the third execution duration, and uses the third planned route as the moving path when the first execution duration is less than the third execution duration, thereby improving the accuracy of moving path formulation.

[0119] In one embodiment, step 2: training a hazard source detection neural network model includes:

[0120] Obtain the historical operation record of the substation; wherein the historical operation record is: the historical operation status record of the substation;

[0121] According to the historical operation records, potential danger sources are determined; wherein the potential danger sources are: danger sources of the substation obtained based on the historical operation records;

[0122] Obtain the substation upload database; wherein the substation upload database includes the substation data uploaded by the Industrial Internet exchange platform;

[0123] According to the substation upload database, the associated image of the potential hazard source is obtained; wherein the associated image is: an image containing the potential hazard source;

[0124] Based on the yolov8 target model, the hazard source detection neural network model is trained according to the associated images. Among them, the yolov8 target model is a version of the YOLO series target detection algorithm. When training the hazard source detection neural network model according to the associated images, the target detection task is converted into a single neural network regression problem to achieve it.

[0125] The working principle and beneficial effects of the above technical solution are:

[0126] The data used for training the target model is different, and the subsequent detection objects of the model are also different. Therefore, this application introduces the historical operation records of the substation to determine the potential sources of danger in the historical operation records. The substation upload database is introduced to obtain the associated images of the potential sources of danger. The associated images are used as training data to train the yolov8 target model to obtain the neural network model for hazard source detection, which improves the suitability and detection accuracy of the neural network model for hazard source detection.

[0127] In one embodiment, determining potential danger sources based on historical operation records includes:

[0128] Obtain the device operation status recorded in the historical operation record; wherein the device operation status is: the operation status of the substation equipment in the historical operation record;

[0129] According to the device operation state, determining the first hazard source in the historical operation record of abnormal device operation state; wherein the first hazard source is: the substation equipment with abnormal device operation state in the historical operation record;

[0130] Obtain the personnel accident items recorded in the historical operation records; wherein the personnel accident items are: personnel accident event record items;

[0131] According to the personnel accident event, determine the second hazard source that caused the personnel accident event; wherein the second hazard source is: the target that caused the personnel accident recorded in the personnel accident event record item;

[0132] Treat the first and second hazard sources together as potential hazard sources;

[0133] Wherein, according to the operation status of the device, determining the first hazard source in the historical operation record of abnormal operation status of the device includes:

[0134] Extract the dialogue flow in the historical operation record of the abnormal operation state of the device, and determine the first hazard source according to the dialogue flow; wherein the dialogue flow is: the dialogue record of the detection personnel who detected the abnormal operation state of the device; when determining the first hazard source, parse the semantics of the dialogue record to obtain the first hazard source mentioned;

[0135] Among them, according to the personnel accident event, the second hazard source that caused the personnel accident event is determined, including:

[0136] Obtain the accident source attribution model, and determine the second hazard source based on the personnel accident event and the accident source attribution model. The accident source attribution model is an artificial intelligence model used to attribute the accident source of the accident event.

[0137] The working principle and beneficial effects of the above technical solution are:

[0138] This application introduces historical operation records, obtains the operating status of the device, and determines the first hazard source in the historical operation records of the abnormal operating status of the device based on the operating status of the device. When determining, extract the conversation records of the detection personnel who detected the abnormal operating status of the device, and determine the first hazard source based on the conversation flow. Obtain the personnel accident events recorded in the historical operation records, and at the same time, obtain the accident source attribution model. The accident source attribution model is an artificial intelligence model for attributing the accident source of the accident event. According to the personnel accident events and the accident source attribution model, determine the second hazard source, and treat the first hazard source and the second hazard source together as potential hazard sources, thereby improving the comprehensiveness of the acquisition of potential hazard sources.

[0139] In one embodiment, step 4: determining the relative distance between the danger source in the station and the shooting position of the binocular camera according to the shooting characteristics of the danger source in the station includes:

[0140] Calculate feature point parallax based on the shooting features; wherein the feature point parallax is: the position difference between the image features of the left camera image and the image features of the right camera image at the same image position;

[0141] Determine the horizontal offset according to the parallax of the feature points;

[0142] Measure the binocular physical distance between binocular cameras; wherein the binocular physical distance is: the baseline length between the binocular cameras;

[0143] According to the binocular physical distance, the ranging error calculation coefficient is determined; wherein the ranging error calculation coefficient is: a parameter for correcting or modifying the relative distance;

[0144] Get the focal length parameters of the binocular camera; where focal length parameters: the shooting focal length of the binocular camera;

[0145] According to the binocular physical distance, focal length parameters, horizontal offset and ranging error calculation coefficient, the relative distance is determined. The calculation formula of the relative distance is:

[0146]

[0147] Where D is the relative distance, dis i is the i-th horizontal offset, Base is the binocular physical distance, f li is the i-th focal length parameter, I is the total number of horizontal offsets, exp is the ranging error calculation coefficient function, and exp(r) is the ranging error calculation coefficient.

[0148] The working principle and beneficial effects of the above technical solution are:

[0149] This application introduces shooting features to calculate feature point parallax, and determines the horizontal offset based on the feature point parallax. Determine the ranging error calculation coefficient based on the measured binocular physical distance between the binocular cameras. Determine the relative distance based on the binocular physical distance, focal length parameters, horizontal offset and ranging error calculation coefficient, and the relative distance calculation is more accurate.

[0150] The embodiment of the present invention provides a method for automatically measuring the distance of a hazardous source in a substation based on binocular ranging, and further includes:

[0151] The number of binocular cameras is obtained, and based on the difference in the number of cameras, the target triggering camera is determined; wherein the number of cameras is: the number of binocular cameras; the target triggering camera is: the camera in the binocular camera that triggers the execution of the shooting task;

[0152] Wherein, based on the number of cameras, determining the target triggering camera includes:

[0153] When the number of cameras is equal to 1, the corresponding binocular camera is used as the target trigger camera;

[0154] When the number of cameras is greater than 1, at least one camera to be triggered among the binocular cameras is determined based on the trigger mechanism, and an evaluation result of the camera to be triggered is obtained based on a preset distance measurement accuracy evaluation system; wherein the trigger mechanism is: a mechanism for triggering the camera to be triggered; the preset distance measurement accuracy evaluation system is manually preset; and the evaluation result is: an evaluation result of the image quality taken by the camera to be triggered;

[0155] According to the evaluation results, the target is determined to trigger the camera.

[0156] The working principle and beneficial effects of the above technical solution are:

[0157] When there are multiple binocular cameras, it is necessary to consider determining the binocular camera that is most suitable for shooting. Therefore, this application obtains the number of binocular cameras. When the number of cameras is equal to 1, the target trigger camera is directly determined. When the number of cameras is greater than 1, a trigger mechanism is introduced to determine at least one camera to be triggered. A ranging accuracy evaluation system is introduced to obtain the evaluation results of the cameras to be triggered, and the target trigger camera with the best evaluation results is determined, which improves the suitability of the target trigger camera selection.

[0158] In one embodiment, determining at least one camera to be triggered in the binocular camera and obtaining an evaluation result of the camera to be triggered based on a preset distance measurement accuracy evaluation system includes:

[0159] According to the shooting characteristics, the required focal length range is determined; wherein the required focal length range is: the focal length range required for shooting the substation image;

[0160] Determine the application focal length range according to the binocular physical distance; wherein the application focal length range is: the focal length range that the binocular camera can capture at the binocular physical distance;

[0161] Determine the camera to be triggered according to the required focal length range and the application focal length range;

[0162] Acquire a pre-shot image of the camera to be triggered; wherein the pre-shot image is: an image of the danger source in the station taken by the camera to be triggered;

[0163] Based on the preset distance measurement accuracy evaluation system, the evaluation result of the trigger camera to be evaluated is obtained according to the pre-shot image. Among them, the preset distance measurement accuracy evaluation system is: the distance measurement accuracy evaluation system is preset manually; the evaluation result is: the shooting quality evaluation of the pre-shot image.

[0164] The working principle and beneficial effects of the above technical solution are:

[0165] This application determines the required focal length range based on the shooting characteristics, determines the applied focal length range based on the binocular physical distance, and determines the camera to be triggered based on the required focal length range and the applied focal length range. The pre-shot image of the camera to be triggered is obtained, and a distance measurement accuracy evaluation system is introduced. Based on the pre-shot image, the evaluation result of the camera to be triggered is obtained, which improves the suitability of the evaluation of the camera to be triggered and makes the selection of the camera to be triggered more accurate.

[0166] The embodiment of the present invention provides an automatic measurement system for the distance of dangerous sources in a substation based on binocular ranging, such as Figure 2 As shown, including:

[0167] The substation image acquisition subsystem 1 is used to acquire the substation image in the substation taken by the binocular camera;

[0168] Model training subsystem 2, used to train the hazard source detection neural network model;

[0169] The subsystem 3 for determining the danger source in the station is used to determine the danger source in the station according to the substation image based on the danger source detection neural network model;

[0170] The relative distance determination subsystem 4 is used to determine the relative distance between the danger source in the station and the shooting position of the binocular camera according to the shooting characteristics of the danger source in the station.

[0171] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. The automatic measurement method of the distance of dangerous sources in the substation based on binocular ranging is characterized by: include: Step 1: Obtain the substation image inside the substation taken by the binocular camera; Step 2: Training the hazard source detection neural network model; Step 3: Based on the hazard source detection neural network model, determine the hazard sources in the substation according to the substation image; Step 4: According to the shooting characteristics of the danger source in the station, determine the relative distance between the danger source in the station and the shooting position of the binocular camera; Step 1: Obtain the substation image inside the substation taken by the binocular camera, including: Obtain the distribution of in-station equipment within the substation; Determine the suspected locations of dangerous sources based on the distribution of equipment within the station; Determine the drivable routes within the station; According to the drivable routes and suspected dangerous source locations within the station, set the moving path of the mobile intelligent distance measuring device; Control the mobile intelligent distance measuring device to move based on the moving path, and at the same time, establish a data transmission link between the mobile intelligent distance measuring device and the binocular camera preset on the mobile intelligent distance measuring device; Based on the data transmission link, the substation image is obtained; According to the drivable routes within the station and the suspected locations of dangerous sources, the moving path of the mobile intelligent distance measuring device is set, including: Get the shooting range of the binocular camera; Performing path planning based on drivable routes within the station to obtain multiple first planned routes; Based on the first planned route and the shooting range, determine the second planned route with a comprehensive inspection level for the suspicious points of the hazard source; Determine whether there are repeated passes of suspicious points of dangerous sources on the second planned route; Determine the moving path based on the existence judgment result of the repeated passing of the suspicious points of the hazard source on the second planned route; Based on the existence determination result of the repeated passing of the suspicious points of the hazard source on the second planned route, the moving path is determined, including: If the existence determination result is non-existence, the corresponding second planned route will be used as the third planned route; Obtain the first execution duration of the third planned route; If the existence determination result is existence, the corresponding second planned route is used as the fourth planned route; Obtain the second execution time of the fourth planned route; Determine the number of times the suspected hazardous source point has been passed based on repeated passing situations; Obtain inspection priority for suspected hazardous source locations; According to the number of passes and the inspection priority, the reasonableness of the number of passes is determined, and based on the reasonableness, the second execution time is adjusted downward to obtain a third execution time; Comparing the first execution duration with the third execution duration, and determining the moving paths in the third planned route and the fourth planned route according to the comparison result of the duration; Step 4: According to the shooting characteristics of the danger source in the station, determine the relative distance between the danger source in the station and the shooting position of the binocular camera, including: Calculate the parallax of feature points based on the shooting features; Determine the horizontal offset according to the parallax of the feature points; Measure the binocular physical distance between binocular cameras; According to the binocular physical distance, determine the distance measurement error calculation coefficient; Get the focal length parameters of the binocular camera; According to the binocular physical distance, focal length parameters, horizontal offset and ranging error calculation coefficient, the relative distance is determined. The calculation formula of the relative distance is: Where D is the relative distance, dis i is the i-th horizontal offset, Base is the binocular physical distance, f li is the i-th focal length parameter, I is the total number of horizontal offsets, exp is the ranging error calculation coefficient function, and exp (r) is the ranging error calculation coefficient.

2. The method for automatically measuring the distance of dangerous sources in a substation based on binocular ranging according to claim 1, characterized in that: Step 2: Training the hazard source detection neural network model, including: Obtain historical operation records of substations; Determine potential sources of danger based on historical operation records; Get the substation upload database; Obtain relevant images of potential danger sources based on the substation upload database; Based on the yolov8 target model and the associated images, the neural network model for hazard source detection is trained.

3. The method for automatically measuring the distance of dangerous sources in a substation based on binocular ranging as claimed in claim 2, characterized in that: Based on historical operation records, identify potential sources of danger, including: Obtain the device operation status recorded in the historical operation record; According to the device operation status, determine the first hazard source in the historical operation records of abnormal device operation status; Obtain personnel accident events recorded in historical operation records; According to the personnel accident event, determine the second hazard source that caused the personnel accident event; Treat the first and second hazard sources together as potential hazard sources; Wherein, according to the operation status of the device, determining the first hazard source in the historical operation record of abnormal operation status of the device includes: Extract the dialogue flow in the historical operation record of abnormal operation status of the device, and determine the first hazard source based on the dialogue flow; Among them, according to the personnel accident event, the second hazard source that caused the personnel accident event is determined, including: Obtain the accident source attribution model, and determine the second hazard source based on the personnel accident events and the accident source attribution model.

4. The method for automatically measuring the distance of dangerous sources in a substation based on binocular ranging according to claim 1, characterized in that: Also includes: Obtain the number of binocular cameras, and determine the target triggering camera based on the number of cameras; Wherein, based on the number of cameras, determining the target triggering camera includes: When the number of cameras is equal to 1, the corresponding binocular camera is used as the target trigger camera; When the number of cameras is greater than 1, at least one camera to be triggered in the binocular cameras is determined based on the trigger mechanism, and an evaluation result of the camera to be triggered is obtained based on a preset ranging accuracy evaluation system; According to the evaluation results, the target is determined to trigger the camera.

5. The method for automatically measuring the distance of dangerous sources in a substation based on binocular ranging as claimed in claim 4, characterized in that: Determine at least one camera to be triggered in the binocular camera, and obtain the evaluation result of the camera to be triggered based on a preset distance measurement accuracy evaluation system, including: Determine the required focal length range based on the shooting characteristics; Determine the application focal length range based on the binocular physical distance; Determine the camera to be triggered according to the required focal length range and the application focal length range; Obtain the pre-shot image of the camera to be triggered; Based on the preset distance measurement accuracy evaluation system, the evaluation result of the camera to be triggered is obtained according to the pre-shot image.

6. The automatic measurement system of the distance of dangerous sources in the substation based on binocular ranging is characterized by: include: The substation image acquisition subsystem is used to acquire the substation image in the substation taken by the binocular camera; Model training subsystem, used to train the hazard source detection neural network model; The subsystem for determining hazardous sources within the station is used to determine hazardous sources within the station based on the hazard source detection neural network model and the substation image; The relative distance determination subsystem is used to determine the relative distance between the danger source in the station and the shooting position of the binocular camera according to the shooting characteristics of the danger source in the station; The substation image acquisition subsystem performs the following operations: Obtain the distribution of in-station equipment within the substation; Determine the suspected locations of dangerous sources based on the distribution of equipment within the station; Determine the drivable routes within the station; According to the drivable routes and suspected dangerous source locations within the station, set the moving path of the mobile intelligent distance measuring device; Control the mobile intelligent distance measuring device to move based on the moving path, and at the same time, establish a data transmission link between the mobile intelligent distance measuring device and the binocular camera preset on the mobile intelligent distance measuring device; Based on the data transmission link, the substation image is obtained; According to the drivable routes within the station and the suspected locations of dangerous sources, the moving path of the mobile intelligent distance measuring device is set, including: Get the shooting range of the binocular camera; Performing path planning based on drivable routes within the station to obtain multiple first planned routes; Based on the first planned route and the shooting range, determine the second planned route with a comprehensive inspection level for the suspicious points of the hazard source; Determine whether there are repeated passes of suspicious points of dangerous sources on the second planned route; Determine the moving path based on the existence judgment result of the repeated passing of the suspicious points of the hazard source on the second planned route; Based on the existence determination result of the repeated passing of the suspicious points of the hazard source on the second planned route, the moving path is determined, including: If the existence determination result is non-existence, the corresponding second planned route will be used as the third planned route; Obtain the first execution duration of the third planned route; If the existence determination result is existence, the corresponding second planned route is used as the fourth planned route; Obtain the second execution time of the fourth planned route; Determine the number of times the suspected hazardous source point has been passed based on repeated passing situations; Obtain inspection priority for suspected hazardous source locations; According to the number of passes and the inspection priority, the reasonableness of the number of passes is determined, and based on the reasonableness, the second execution time is adjusted downward to obtain a third execution time; Comparing the first execution duration with the third execution duration, and determining the moving paths in the third planned route and the fourth planned route according to the comparison result of the duration; The relative distance determination subsystem performs the following operations: Calculate the parallax of feature points based on the shooting features; Determine the horizontal offset according to the parallax of the feature points; Measure the binocular physical distance between binocular cameras; According to the binocular physical distance, determine the distance measurement error calculation coefficient; Get the focal length parameters of the binocular camera; According to the binocular physical distance, focal length parameters, horizontal offset and ranging error calculation coefficient, the relative distance is determined. The calculation formula of the relative distance is: Where D is the relative distance, dis i is the i-th horizontal offset, Base is the binocular physical distance, f li is the i-th focal length parameter, I is the total number of horizontal offsets, exp is the ranging error calculation coefficient function, and exp(r) is the ranging error calculation coefficient.

Images