Open stope inspection method, system and device based on unmanned aerial vehicle

Abstract

The invention provides an open-air stope inspection method, system and device based on an unmanned aerial vehicle, and the method comprises the steps: obtaining an inspection image of a to-be-inspected area in an open-air stope, and enabling a data collection unit carried on the unmanned aerial vehicle to carry out the collection of the inspection image when the unmanned aerial vehicle flies on the open-air stope according to a preset inspection route; the name of the preset inspection route comprises the name of a target object which needs to be collected when the unmanned aerial vehicle flies along the preset inspection route; determining an identification algorithm corresponding to the at least one target object from a preset algorithm set according to the name of the preset inspection route; and determining the fault or information of the target object according to the recognition algorithm corresponding to the target object. On the basis that the unmanned aerial vehicle shoots the images of the open-air stopes, specific recognition algorithms are set for different target objects, the corresponding recognition algorithms are selected according to the target objects to monitor the target objects, faults of the target objects are found in time, and therefore the monitoring precision and efficiency of the open-air stopes are improved.

Description

Technical Field

[0001] This application relates to the technical field of unmanned aerial vehicles, and particularly to a method, system and device for inspecting open-pit mines based on unmanned aerial vehicles. Background Art

[0002] An open-pit mine (open-pit stope) refers to a workplace where stripping and mining are directly carried out on the surface. Among them, during the mining process of an open-pit mine, mining is carried out layer by layer from top to bottom. Minerals and rocks in the mining area need to be divided into horizontal layers with a certain thickness or piled up to form slopes. During continuous mining, it is necessary to monitor the slopes to avoid slope cracks and landslides. In addition, to ensure the safe production of the open-pit stope, it is also necessary to monitor the power supply lines, drainage pipelines, personnel, vehicles, etc. in the open-pit stope, promptly discover existing faults, and eliminate potential safety hazards.

[0003] However, in the prior art, for the monitoring of open-pit mines, most still rely on manual inspections of open-pit mines by workers, resulting in low inspection efficiency and difficulty in promptly discovering faults. Therefore, in order to improve efficiency and reduce the workload of workers, some open-pit mines use unmanned aerial vehicles to take pictures of the open-pit mines, and then workers view the pictures and rely on experience to discover faults and potential safety hazards in the open-pit mines. Although this method can take pictures of the open-pit mines in real time, relying on experience to discover faults and potential safety hazards from the pictures has errors and also affects the inspection efficiency.

[0004] In summary, the current inspection methods for open-pit mines are inefficient and difficult to promptly and accurately discover faults and potential safety hazards. Summary of the Invention

[0005] This application provides a method, system and device for inspecting open-pit mines based on unmanned aerial vehicles to solve the technical problems mentioned in the background art section.

[0006] In a first aspect, this application provides a method for inspecting an open-pit mine based on an unmanned aerial vehicle, including: Obtaining an inspection image of an area to be inspected in the open-pit mine, where the inspection image is collected by a data collection unit carried on the unmanned aerial vehicle when the unmanned aerial vehicle flies over the open-pit mine along a preset inspection route. The name of the preset inspection route includes the name of the target object that needs to be collected when the unmanned aerial vehicle flies along the preset inspection route. The target object is at least one of the following: slope, power supply line, drainage pipeline, electric switch, fire, personnel, and vehicle; Determine the recognition algorithm corresponding to at least one target object from the preset algorithm set according to the name of the preset inspection route, where the preset algorithm set includes at least one of the following algorithms: an image processing algorithm based on semantic segmentation corresponding to the slope, an image processing algorithm for identifying broken strands of the power supply line corresponding to the power supply line, an image processing algorithm for identifying water leakage corresponding to the drainage pipeline, a temperature monitoring algorithm corresponding to the switch, a target detection algorithm for identifying flame and smoke corresponding to the fire, and a target detection algorithm corresponding to the personnel and vehicles; Determine the fault or information of the target object according to the recognition algorithm corresponding to the target object.

[0007] Optionally, when the target object is the slope, the determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Determine the category of each pixel in the inspection image according to the image processing algorithm based on semantic segmentation; Determine whether there are slope cracks and landslides on the slope according to the category of each pixel in the inspection image.

[0008] Optionally, when the target object is the power supply line, the determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Perform image cropping on the inspection image according to the image processing algorithm for identifying broken strands of the line to obtain multiple sub-inspection images; Determine the category of each pixel in each sub-inspection image according to the image processing algorithm for identifying broken strands of the line, and determine the sub-inspection images where the power supply line exists according to the category of each pixel; Judge whether there is a wire breakage fault in the power supply line according to the sub-inspection images where the power supply line exists; For the power supply line with the wire breakage fault, determine the position of the wire breakage fault in the power supply line.

[0009] Optionally, the judging whether there is a wire breakage fault in the power supply line according to the sub-inspection images where the power supply line exists includes: Obtain the temperature value in the sub-inspection image where the power supply line exists; Determine at least one sub-inspection image where the power supply line has a wire breakage fault according to the temperature value in the sub-inspection image and the preset temperature value, and record it as the first set of sub-inspection images; Determine at least one sub-inspection image where the power supply line has a wire breakage fault according to the image processing algorithm for identifying broken strands of the line, and record it as the second set of sub-inspection images; Based on the first sub-inspection image set and the second sub-inspection image set, determine whether there is a wire strand break fault in the power supply line.

[0010] Optionally, when the target object is a drainage pipeline, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Determine whether there is a water leakage fault in the drainage pipeline according to the inspection image and the target detection model corresponding to the water leakage fault, where the target detection model corresponding to the water leakage fault is obtained by training with images corresponding to the drainage pipeline containing the water leakage fault; When there is a drainage pipeline with a water leakage fault in the inspection image, determine the position of the water leakage fault in the open-pit mine through the target detection model corresponding to the water leakage fault.

[0011] Optionally, when the target object is a switch, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Obtain the temperature value in the inspection image including the switch; Determine whether there is a fault in the switch according to the temperature value and the switch temperature warning value corresponding to the switch.

[0012] Optionally, when the target object is a fire, or personnel and vehicles, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Determine whether the target object exists in the inspection image according to the inspection image and the target detection model corresponding to the target object, where the target detection model corresponding to the target object is obtained by training with images including the target object; When the inspection image includes the target object, determine the target detection model corresponding to the target object and determine the position of the target object in the open-pit mine.

[0013] Optionally, after determining the fault or information of the target object according to the recognition algorithm corresponding to the target object, it further includes: Output a warning message through the drone.

[0014] In a second aspect, the present application provides an open-pit mine inspection system based on a drone, including: a drone body, a processing unit, a data acquisition unit, and a communication unit, where the data acquisition unit and the communication unit are both connected to the processing unit, and the data acquisition unit and the communication unit are both carried on the drone body; The data acquisition unit is configured to take pictures of the area to be inspected in the open-pit mine under the control of the processing unit to obtain inspection images; The communication unit is used for communication among the UAV body, the data acquisition unit and the processing unit; The processing unit is used to execute the method according to any item of the first aspect.

[0015] In a third aspect, the present application provides an open-pit stope inspection device based on a UAV, including: An acquisition module, configured to acquire inspection images of an area to be inspected in the open-pit stope. The inspection images are acquired by a data acquisition unit carried on the UAV when the UAV flies over the open-pit stope according to a preset inspection route. The name of the preset inspection route includes the name of the target object to be acquired when the UAV flies along the preset inspection route. The target object is at least one of the following: slope, power supply line, drainage pipeline, electric switch, fire, personnel and vehicle; A determination module, configured to determine, according to the name of the preset inspection route, an identification algorithm corresponding to at least one target object from a preset algorithm set. The preset algorithm set includes at least one of the following algorithms: an image processing algorithm based on semantic segmentation corresponding to the slope, an image processing algorithm for identifying broken strands of a line corresponding to the power supply line, an image processing algorithm for identifying water leakage corresponding to the drainage pipeline, a temperature monitoring algorithm corresponding to the electric switch, a target detection algorithm for identifying flames and smoke corresponding to the fire, and a target detection algorithm corresponding to the personnel and vehicle; An identification module, configured to determine faults or information of the target object according to the identification algorithm corresponding to the target object Optionally, when the target object is the slope, when the identification module determines the fault of the target object according to the identification algorithm corresponding to the target object, it specifically is used for: Determine the category of each pixel in the inspection image according to the image processing algorithm based on semantic segmentation; Determine whether there are slope ground fissures and landslides on the slope according to the category of each pixel in the inspection image.

[0016] Optionally, when the target object is a power supply line, when the identification module determines the fault of the target object according to the identification algorithm corresponding to the target object, it specifically is used for: Perform image cropping on the inspection image according to the image processing algorithm for identifying broken strands of a line to obtain multiple sub-inspection images; Determine the category of each pixel in each sub-inspection image according to the image processing algorithm for identifying broken strands of a line, and determine the sub-inspection images where the power supply line exists according to the category of each pixel; Judge whether there is a wire breakage fault in the power supply line according to the sub-inspection images where the power supply line exists; For a power supply line with the wire strand breakage fault, determine the location of the wire strand breakage fault in the power supply line.

[0017] Optionally, when the recognition module determines whether there is a wire strand breakage fault in the power supply line according to the sub-inspection image of the power supply line, it is specifically used for: Obtain the temperature value in the sub-inspection image of the power supply line; According to the temperature value in the sub-inspection image and the preset temperature value, determine at least one sub-inspection image of the power supply line with a wire strand breakage fault, denoted as the first sub-inspection image set; According to the image processing algorithm for wire strand breakage recognition, determine at least one sub-inspection image of the power supply line with a wire strand breakage fault, denoted as the second sub-inspection image set; According to the first sub-inspection image set and the second sub-inspection image set, determine whether there is a wire strand breakage fault in the power supply line.

[0018] Optionally, when the target object is a drainage pipeline, when the recognition module determines the fault of the target object according to the recognition algorithm corresponding to the target object, it is specifically used for: According to the inspection image and the target detection model corresponding to the water leakage fault, determine whether there is a water leakage fault in the drainage pipeline, where the target detection model corresponding to the water leakage fault is obtained by training with the image corresponding to the drainage pipeline containing the water leakage fault; When there is a drainage pipeline with a water leakage fault in the inspection image, determine the location of the water leakage fault in the open-pit mine through the target detection model corresponding to the water leakage fault.

[0019] Optionally, when the target object is a switch, when the recognition module determines the fault of the target object according to the recognition algorithm corresponding to the target object, it is specifically used for: Obtain the temperature value in the inspection image including the switch; According to the temperature value and the switch temperature warning value corresponding to the switch, determine whether there is a fault in the switch.

[0020] Optionally, when the target object is a fire, or personnel and vehicles, when the recognition module determines the fault of the target object according to the recognition algorithm corresponding to the target object, it is specifically used for: According to the inspection image and the target detection model corresponding to the target object, determine whether the target object exists in the inspection image, where the target detection model corresponding to the target object is obtained by training with the image containing the target object; When the inspection image contains the target object, determine the target detection model corresponding to the target object and determine the location of the target object in the open-pit mine.

[0021] Optionally, it further includes: an early warning module; After the recognition module determines the fault or information of the target object according to the recognition algorithm corresponding to the target object, the early warning module is configured to: Output an early warning message through the UAV.

[0022] In a fourth aspect, the present application provides an electronic device, including: a processor and a memory; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory, so that the processor executes the method described in any item of the first aspect.

[0023] In a fifth aspect, an embodiment of the present application provides a readable storage medium, including a program or instructions. When the program or instructions are run on a computer, the method described in any item of the first aspect above is executed.

[0024] In a sixth aspect, an embodiment of the present application provides a computer program product, including a computer program. When the computer program is executed by a processor, the method described in any item of the first aspect is implemented.

[0025] The method, system and device for inspecting an open-pit stope based on a UAV provided by the present application obtain inspection images of an area to be inspected in the open-pit stope. The inspection images are collected by a data collection unit carried on the UAV when the UAV flies over the open-pit stope according to a preset inspection route. The name of the preset inspection route includes the name of the target object to be collected when the UAV flies along the preset inspection route; according to the name of the preset inspection route, an identification algorithm corresponding to at least one target object is determined from a preset algorithm set; according to the identification algorithm corresponding to the target object, the fault or information of the target object is determined. On the basis of the UAV taking images of the open-pit stope, specific identification algorithms are set for different target objects, and corresponding identification algorithms are selected according to the target objects to monitor the target objects, and the faults of the target objects are found in time, thereby improving the monitoring accuracy and efficiency of the open-pit stope. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0027] Figure 1 It is a flowchart of a method for inspecting an open-pit stope based on a UAV provided by an embodiment of the present application; Figure 2 Structural diagram of an open-pit stope inspection system based on an unmanned aerial vehicle provided by an embodiment of the present application; Figure 3 Schematic structural diagram of an open-pit stope inspection device based on an unmanned aerial vehicle provided by an embodiment of the present application; Figure 4 Schematic structural diagram of an electronic device provided by an embodiment of the present application. Detailed implementation manners

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below. Apparently, the described embodiments are some but not all of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts also belong to the scope of protection of the present application.

[0029] In the prior art, for the monitoring of open-pit stopes, most of them still rely on manual inspection of open-pit stopes by workers, resulting in low inspection efficiency and difficulty in timely detecting faults. Therefore, in order to improve efficiency and reduce the workload of workers, some open-pit stopes use unmanned aerial vehicles to take pictures of the open-pit stope, and then workers view the pictures and rely on experience to find faults and potential safety hazards in the open-pit stope. Although this method can take pictures of the open-pit stope in real time, relying on experience to find faults and potential safety hazards from the pictures has errors and also affects the inspection efficiency.

[0030] Therefore, to solve the technical problems existing in the prior art, the present application proposes an open-pit stope inspection method, system, and device based on an unmanned aerial vehicle. In the system, the recognition algorithm corresponding to each target object is set according to its fault attribute or its own information attribute, and the system controls the inspection route of the unmanned aerial vehicle, so that the data acquisition unit carried on the unmanned aerial vehicle takes pictures of multiple target objects on the inspection route and sends the inspection images to the processing unit. In this way, the processing unit can retrieve the corresponding recognition algorithm according to the target object to be inspected by the user, and perform fault or information recognition on the target object to be inspected, which can quickly and accurately identify abnormal situations on site, reduce manual intervention, improve inspection efficiency, reduce safety risks and costs, and meet the requirements of inspection scenarios for different open-pit stopes.

[0031] Figure 1 Flowchart of an open-pit stope inspection method based on an unmanned aerial vehicle provided by an embodiment of the present application. Among them, the execution subject of this embodiment can be the open-pit stope inspection system mentioned below, specifically, the processing unit. As Figure 1 shown, the method of this embodiment includes: S101. Obtain the inspection image of the area to be inspected in the open-pit mine.

[0032] Among them, the inspection image is obtained by the data acquisition unit carried on the unmanned aerial vehicle (UAV) when the UAV flies over the open-pit mine according to the preset inspection route. The name of the preset inspection route includes the name of the target object to be collected when the UAV flies along the preset inspection route. The target object is at least one of the following: slope, power supply line, drainage pipeline, electric switch, fire, personnel, and vehicle.

[0033] In this step, the data acquisition unit mainly includes a camera and a thermal imager. The camera can be a monocular camera or a five-eye camera. Here, the five-eye camera is taken as an example for illustration.

[0034] Mount the five-eye camera and the thermal imager on the UAV. The UAV cruises over the open-pit mine according to the inspection route issued by the user. Avoiding light, take pictures of the open-pit mine through the five-eye camera, and the thermal imager synchronously obtains the thermal image of the open-pit mine image taken by the five-eye camera. Among them, the temperature distribution on the surface of the object is displayed on the thermal image. The image of the open-pit mine taken by the five-eye camera and the thermal image synchronously obtained by the thermal imager are collectively referred to as the inspection image.

[0035] Among them, during the mining process of the open-pit mine, a slope will be formed. When the open-pit mine is blasted, vibration is generated, resulting in cracks and landslides on the slope. If slope cracks and landslides occur, it will affect the mining of the open-pit mine. Therefore, it is necessary to monitor the slope and timely detect slope cracks and landslides. Therefore, through UAV cruising, obtain the image corresponding to the slope.

[0036] Moreover, the mining area of the open-pit mine is directly exposed, and water accumulation is likely to occur in the open-pit mine. It is necessary to drain water through the drainage pipeline. Therefore, it is necessary to ensure that there is no leakage problem with the drainage pipeline. Therefore, through UAV cruising, take pictures of the drainage pipeline.

[0037] Moreover, when the open-pit mine is mined, it is necessary to use the power supply line for power supply. If there are faults such as leakage in the power supply line, it will not only affect the mining progress, but may also cause safety accidents such as electric shock. Therefore, it is necessary to timely eliminate the leakage problem. Therefore, it is necessary to cruise through the UAV and take pictures of the power supply line to obtain the image of the power supply line.

[0038] Moreover, overheating of the electric switch will also affect the power supply of the open-pit mine and even cause a fire. Therefore, through UAV cruising, take pictures of the electric switch to obtain the image of the electric switch.

[0039] Since gunpowder is used for blasting during the mining of the open-pit mine and there are many electrical equipment, it is easy to cause a fire. Therefore, it is necessary to monitor the open-pit mine for fire and timely detect the fire. Therefore, the UAV can cruise in the area where a fire is likely to occur and take pictures of the open-pit mine.

[0040] For personnel and vehicles, it is necessary to proofread the personnel information and vehicle information to ensure the accuracy of personnel and vehicles. In addition, for the safety of personnel and vehicles, when personnel and vehicles are found in dangerous areas, the corresponding personnel and vehicles should be notified in time to keep them away from the dangerous areas.

[0041] Therefore, when the drone conducts patrols, multiple preset patrol routes that can cover all the targets in the open-pit mine as much as possible are set. Among them, each preset patrol route is named after the name of the target that needs to be collected when the drone flies along the preset patrol route. For example, when the name of the preset patrol route is "slope power supply line", it means that when the drone flies along this preset patrol route, the patrol images captured include the areas corresponding to the slope and the power supply line. Among them, the name of the patrol image captured through the preset patrol route can be the same as the name of the preset patrol route, which is convenient for finding the patrol image corresponding to each preset patrol route.

[0042] Therefore, when the drone flies along each preset patrol route, the patrol images captured by the five-eye camera and the thermal imager need to include at least one of the slope, power supply line, drainage pipeline, switch, fire, personnel and vehicle. It should be noted that for fire, it does not mean taking pictures of the scene corresponding to the fire, but identifying whether there is a fire or fire hazard in the areas prone to fire in the open-pit mine.

[0043] S102. Determine the recognition algorithm corresponding to at least one target from the preset algorithm set according to the name of the preset patrol route.

[0044] The preset algorithm set includes at least one of the following algorithms: an image processing algorithm based on semantic segmentation corresponding to the slope, an image processing algorithm for identifying broken strands of the power supply line corresponding to the power supply line, an image processing algorithm for identifying water leakage corresponding to the drainage pipeline, a temperature monitoring algorithm corresponding to the switch, a target detection algorithm for identifying flames and smoke corresponding to the fire, and a target detection algorithm corresponding to personnel and vehicles.

[0045] In this step, in the processing unit, according to the characteristics of each target and the characteristics of the corresponding faults, the corresponding recognition algorithms are specially set for each target. In this way, after the processing unit obtains the patrol images captured when the drone flies along the preset patrol route, according to the name of the preset patrol route, it determines the targets included in the patrol images, and then determines the recognition algorithms corresponding to the targets. Thus, when the recognition algorithm corresponding to each target identifies the target, it can quickly and accurately identify the faults or information of the target, improving the accuracy and performance of the system.

[0046] S103. Determine the faults or information of the target object according to the recognition algorithm corresponding to the target object.

[0047] In this step, optionally, when the target object is a slope, the specific implementation manner of S104 is as follows: S201. Determine the category of each pixel in the inspection image according to the image processing algorithm based on semantic segmentation.

[0048] Specifically, slope ground fissures usually appear as irregular and slender cracks, whose width and shape vary due to different environmental conditions and geological features. The landslide body usually covers a wide area and has an irregular boundary, and the open-pit mine environment is complex, and the distance is far when the drone takes pictures, etc. These factors will cause the slope cracks and landslides on the inspection image to be unclear. Therefore, for slopes, an identification algorithm that can identify the boundary and shape is required.

[0049] The image processing algorithm based on semantic segmentation classifies each pixel in the image to achieve pixel-level fine processing. Therefore, when identifying slope cracks and landslide bodies, the boundary parts of slope cracks and landslide bodies can be accurately extracted. Therefore, in this embodiment, the image processing algorithm based on semantic segmentation is used to identify slope cracks and landslide bodies.

[0050] For example, in this embodiment, image semantic segmentation based on a fully convolutional network can be used. When identifying slope cracks and landslide bodies, the inspection image is input into the image semantic segmentation based on the fully convolutional network. This algorithm processes the inspection image and directly outputs the category corresponding to each pixel in the inspection image. Here, the category refers to the target object. For example, the pixel corresponds to a landslide, or a vehicle, or a road, etc., or an object that is not monitored in this application in the open-pit mine, such as a road, a mineral deposit.

[0051] It should be noted that the image semantic segmentation-based method for identifying slope cracks and landslide bodies needs to be trained through model training, that is, the slope images of open-pit mines with slope cracks and / or landslide bodies and the slope images of open-pit mines without slope cracks and / or landslide bodies are used as training data to train the image processing algorithm based on semantic segmentation. The training process can refer to the prior art and will not be elaborated here.

[0052] S202. Determine whether there are slope ground fissures and landslide bodies on the slope according to the category of each pixel in the inspection image.

[0053] Specifically, in the image processing algorithm based on semantic segmentation, for pixels of the same category, their corresponding features are the same. According to the contours formed by the pixels with the same features in the inspection image, the areas corresponding to slope cracks and landslide bodies can be determined, and the slope cracks and landslide bodies are extracted from the inspection image, so as to determine whether there are slope cracks and landslide bodies on the slope.

[0054] Optionally, when the target is a power supply line, the specific implementation of S104 is as follows: S301. According to the image processing algorithm for line strand break identification, crop the inspection image to obtain multiple sub-inspection images.

[0055] Specifically, the power supply line has the characteristics of being slender. Therefore, the detection of wire strand breaks requires accurate identification of slender targets, and the break points are usually small, requiring the algorithm to have high-precision positioning capabilities. To solve this problem, an identification algorithm based on image cropping, semantic segmentation, classification, and object detection is adopted.

[0056] First, use the image cropping model in the identification algorithm to crop the inspection image to obtain multiple sub-inspection images. Among them, the sizes of the multiple sub-inspection images can be the same or different. Among them, the image cropping technology can adopt a threshold-based cropping algorithm to divide the pixels in the image into different regions according to the gray value, so as to crop out the region corresponding to the power supply line and reduce the influence of other regions except the power supply line in the inspection image on the power supply line, improving the efficiency and accuracy of wire strand break identification.

[0057] S302. According to the image processing algorithm for line strand break identification, determine the category of each pixel in each sub-inspection image, and determine the sub-inspection images with a power supply line according to the category of each pixel.

[0058] Specifically, determine the category of each pixel in the sub-inspection image based on the semantic segmentation model in the identification algorithm. Then, according to the category of each marked pixel and the category of the pixels corresponding to the power supply line, determine the sub-inspection images with a power supply line. Among them, the semantic segmentation model adopted can also be image semantic segmentation based on a fully convolutional network.

[0059] S303. According to the sub-inspection images with a power supply line, determine whether there is a wire strand break fault in the power supply line.

[0060] Specifically, for the sub-inspection images without a power supply line, there is no need to identify wire strand break faults. For the sub-inspection images with a power supply line, the identification method can be as follows: S3031. Obtain the temperature value in the sub-inspection image with a power supply line.

[0061] Specifically, when a fault occurs in the power supply line, it will be reflected in the temperature. Generally, for wire strand break faults, the temperature will rise. Therefore, for the thermal image in the inspection image, obtain the temperature value of the thermal image.

[0062] S3032. Determine at least one sub-inspection image indicating a wire strand break fault in the power supply line based on the temperature value in the sub-inspection image and the preset temperature value, and denote it as the first sub-inspection image set.

[0063] Specifically, the preset temperature value is the temperature value on the thermal image when the thermal imager takes pictures when the power supply line has no fault. Therefore, by comparing the temperature value in the sub-inspection image with the preset temperature value, when there is a temperature value greater than the preset temperature value in the sub-inspection image, it indicates that there is a wire strand break fault in the power supply line in the sub-inspection image.

[0064] Therefore, obtain at least one sub-inspection image with a temperature value greater than the preset temperature value, and denote it as the first sub-inspection image set, where the first sub-inspection image set can be an empty set.

[0065] S3033. Determine at least one sub-inspection image indicating a wire strand break fault in the power supply line according to the image processing algorithm for wire strand break identification, and denote it as the second sub-inspection image set.

[0066] Specifically, the classification model in the recognition algorithms based on semantic segmentation, image cropping, classification, and object detection has the advantage of analyzing local features, so that the wire strand break of the power supply line can be identified. Therefore, according to the classification model, determine at least one sub-inspection image with a wire strand break fault from the sub-inspection images with the power supply line, and denote it as the second sub-inspection image set, where the first sub-inspection image set can be an empty set.

[0067] Among them, the classification model can be obtained by training with the power supply lines with wire strand break faults and the power supply lines without wire strand break faults. Therefore, the power supply lines can be classified into two types: power supply lines with wire strand break faults and power supply lines without wire strand break faults through the classification model.

[0068] S3034. Judge whether there is a wire strand break fault in the power supply line according to the first sub-inspection image set and the second sub-inspection image set.

[0069] Specifically, when both the first sub-inspection image set and the second sub-inspection image set are empty sets, confirm that there is no wire strand break fault in the power supply line; when one of the first sub-inspection image set and the second sub-inspection image set is an empty set, or both the first sub-inspection image set and the second sub-inspection image set are not empty sets and there is no intersection, in order to avoid failing to detect the wire strand break fault in the power supply line in time, the drone can be controlled to fly on the preset inspection flight path corresponding to the power supply line every preset time period, take the inspection images of the power supply line, and confirm whether there is a wire strand break fault; when there is an intersection between the first sub-inspection image set and the second sub-inspection image set, determine that there is a wire strand break fault in the power supply line.

[0070] Judge whether there is a wire breakage fault in the power supply line in two ways to improve the accuracy of judgment and reduce the error rate.

[0071] S304. For the power supply line with wire breakage fault, determine the position of the wire breakage fault in the power supply line.

[0072] Specifically, in the recognition algorithms based on semantic segmentation, image cropping, classification, and object detection, the object detection model is used to identify specific objects. For example, the wire breakage in the power supply line. Therefore, for the power supply line determined to have a wire breakage fault through the classification model, then use the object detection model to determine the specific position of the wire breakage in the power supply line and set a mark to quickly detect the wire breakage in the power supply line and repair it in time. Among them, the object detection model can be, for example, the YOLOv8 algorithm.

[0073] Optionally, when the object is a drainage pipeline, the specific implementation method of S104 is as follows: S401. According to the inspection image and the object detection model corresponding to the water leakage fault, determine whether there is a water leakage fault in the drainage pipeline.

[0074] Among them, the object detection model corresponding to the water leakage fault is obtained by training with the image corresponding to the drainage pipeline containing the water leakage fault.

[0075] Specifically, for the detection of water leakage in the drainage pipeline, it mainly focuses on identifying local areas. Since the water leakage area usually shows a jet of water, these areas have the characteristics of jetting in the inspection image and have obvious visual differences from the surrounding environment. Therefore, the water leakage phenomenon has the significant characteristics of concentrating in a specific area and being in a jet state. And the object detection algorithm is used to specifically detect specific objects. Therefore, when judging whether there is a water leakage phenomenon in the drainage pipeline, use the object detection algorithm, such as the YOLOv8 algorithm.

[0076] Among them, for the YOLOv8 algorithm used to judge whether there is a water leakage phenomenon in the drainage pipeline, model training needs to be carried out in advance. When training the model, the training data are the drainage pipelines without water leakage and the drainage pipelines with water leakage. After training, the YOLOv8 algorithm can accurately detect the water leakage fault in the drainage pipeline.

[0077] Through the trained YOLOv8 algorithm, analyze the inspection image to identify the water leakage fault.

[0078] S402. When there is a drainage pipeline with a water leakage fault in the inspection image, determine the position of the water leakage fault in the open-pit mine through the object detection model corresponding to the water leakage fault.

[0079] Specifically, when it is determined that there is a leakage fault in the drainage pipeline, it is necessary to promptly determine the location of the leakage point. Therefore, the YOLOv8 algorithm is used to quickly and accurately locate the leakage point.

[0080] Optionally, when the target is a switch, the specific implementation of S104 is as follows: S501. Obtain the temperature value in the inspection image containing the switch. Specifically, when a switch fails, for example, when the switch has poor contact, it generally manifests in temperature, usually showing an increase in temperature. Therefore, the detection of the switch involves monitoring temperature anomalies. Therefore, through a temperature monitoring algorithm, such as infrared thermal imaging technology, temperature data of the switch area is obtained. Generally, a thermal image of the open-pit mine is taken by a thermal imager, and the temperature at each point in the thermal image is identified to obtain the temperature value in the thermal image.

[0081] S502. Determine whether the switch has a fault according to the temperature value and the switch temperature warning value corresponding to the switch.

[0082] Specifically, the switch temperature warning value is the temperature value corresponding to when the switch has no fault. Therefore, when there is a temperature value in the thermal image that exceeds the switch temperature warning value, a warning is issued to monitor the switch.

[0083] Optionally, when the target is a fire, or personnel and vehicles, the specific implementation of S104 is as follows: S601. Determine whether there is a target in the inspection image according to the inspection image and the target detection model corresponding to the target.

[0084] Among them, the target detection model corresponding to the target is obtained by training with images containing the target.

[0085] Specifically, both fire detection and personnel and vehicle identification involve real-time identification and positioning of targets.

[0086] For fire detection, the main targets are to identify flames and smoke in the image. These features usually have distinct colors and dynamic changes. Therefore, the target detection model, such as the YOLOv8 algorithm, is trained with images containing fires to obtain the YOLOv8 algorithm that can identify fires. Then, the inspection image is processed by the YOLOv8 algorithm that can identify fires to determine whether there is a fire phenomenon in the inspection image.

[0087] For personnel and vehicles, the YOLOv8 algorithm is trained with images containing personnel and vehicles, so that the trained YOLOv8 algorithm processes the inspection image to identify the personnel and vehicles in the inspection image.

[0088] S602. When the inspection image contains the target object, determine the target detection model corresponding to the target object and determine the position of the target object in the open-pit mine.

[0089] Specifically, if a fire phenomenon is recognized in the inspection image, the position of the fire in the open-pit mine is determined through the YOLOv8 algorithm, and the fire is extinguished in time.

[0090] For personnel and vehicles, the positions of the personnel and vehicles are determined through the corresponding YOLOv8 algorithm. For the personnel or vehicles located in the dangerous area, the relevant personnel and the drivers of the vehicles are notified in time.

[0091] Optionally, after S104, the method further includes: S105. Output a warning message through the drone.

[0092] In this step, for the target object with a fault, for example, there is a broken wire in the power supply line, the drainage pipeline leaks water, or the personnel or vehicle is in the dangerous area, a warning message is output through the drone. For example, a loudspeaker is set on the drone to shout at the staff in the open-pit mine remotely to remind the staff.

[0093] In this embodiment, by acquiring the inspection image of the area to be inspected in the open-pit mine, the inspection image is acquired by the data acquisition unit carried on the drone when the drone flies over the open-pit mine according to the preset inspection route, and the name of the preset inspection route includes the name of the target object that needs to be acquired when the drone flies along the preset inspection route; according to the name of the preset inspection route, the recognition algorithm corresponding to at least one target object is determined from the preset algorithm set; according to the recognition algorithm corresponding to the target object, the fault or information of the target object is determined. On the basis of the drone taking pictures of the open-pit mine, specific recognition algorithms are set for different target objects, and the corresponding recognition algorithms are selected according to the target objects to monitor the target objects, and the faults of the target objects are found in time, thereby improving the monitoring accuracy and efficiency of the open-pit mine.

[0094] Figure 2 This is the structural diagram of the open-pit mine inspection system based on the drone provided by an embodiment of the present application. As Figure 2 shown, the open-pit mine inspection system based on the drone includes: a drone body 210, a processing unit 220, a data acquisition unit 230, and a communication unit 240. Among them, the data acquisition unit 230 and the communication unit 240 are both connected to the processing unit 220, and the data acquisition unit 230 and the communication unit 240 are both carried on the drone body 210.

[0095] The data acquisition unit 230 is configured to take pictures of the area to be inspected in the open-pit mine under the control of the processing unit 220 to obtain inspection images; The communication unit 240 is used for communication between the UAV body 210, the data acquisition unit 230 and the processing unit 220; The processing unit 220 is used to execute the solution of any one of the above method embodiments.

[0096] In this embodiment, the communication unit 240 can be, for example, a base station to realize communication between the UAV body 210, the data acquisition unit 230 and the processing unit 220. For example, the UAV body 210 sends real-time position information to the processing unit 220, and the processing unit 220 controls the flight of the UAV body 210, such as the inspection route, or the flight posture. Or, the data acquisition unit 230 sends the collected data to the processing unit 220.

[0097] The data acquisition unit 230 can be, for example, a thermal imager, a monocular camera, or a five-eye camera, which is carried on the UAV body 210 and is used to capture images of the open-pit mine and send the inspection images to the processing unit 220.

[0098] The processing unit 220 is used to send control commands to the UAV body 210 and receive the inspection images transmitted by the data acquisition unit 230. Among them, as Figure 2 shown, the processing unit 220 includes: a data processing sub-module 221, which is used to process the inspection images transmitted by the data acquisition unit 230 and identify the faults or information of the target object. Specifically, the specific working process of the processing unit 220 can refer to the above method embodiments.

[0099] Optionally, as Figure 2 shown, the processing unit 220 further includes: a service application sub-module 222, where the service application sub-module 222 is used to implement the planning of the inspection route of the UAV body 210, the designation of the inspection plan, the push of faults, the query of the airport status, and the query of the status of the UAV body.

[0100] Optionally, as Figure 2 shown, the processing unit 220 further includes: a data storage sub-module 223, and the data storage sub-module 223 is used for local server deployment, storing the inspection images sent by the data acquisition unit 230, and storing the processing results of the inspection images by the data processing sub-module 221, that is, the identified faults or information.

[0101] In this embodiment, the unmanned aerial vehicle (UAV) inspection system for open-pit mines includes a UAV body 210, a processing unit 220, a data acquisition unit 230, and a communication unit 240. Among them, the data acquisition unit 230 and the communication unit 240 are both connected to the processing unit 220, and the data acquisition unit 230 and the communication unit 240 are both carried on the UAV body 210. By setting specific recognition algorithms for different targets in the processing unit 220 and selecting the corresponding recognition algorithm according to the target to monitor the target, the faults of the target can be detected in time, thereby improving the monitoring accuracy and efficiency of the open-pit mine.

[0102] Figure 3 FIG. is a schematic structural diagram of a UAV-based inspection device for an open-pit mine provided by an embodiment of the present application. As Figure 3 shown, the UAV-based inspection device for an open-pit mine includes: an acquisition module 310, a determination module 320, and an identification module 330. Optionally, the UAV-based inspection device for an open-pit mine further includes: an early warning module 340.

[0103] Among them, the acquisition module 310 is used to acquire inspection images of the area to be inspected in the open-pit mine. The inspection images are acquired by the data acquisition unit carried on the UAV when the UAV flies over the open-pit mine according to a preset inspection route. The name of the preset inspection route includes the name of the target to be acquired when the UAV flies along the preset inspection route. The target is at least one of the following: slope, power supply line, drainage pipeline, electric switch, fire, personnel, and vehicle; The determination module 320 is used to determine the recognition algorithm corresponding to at least one target from a preset algorithm set according to the name of the preset inspection route. Among them, the target is the name of the target included in the name of the preset inspection route. The preset algorithm set includes at least one of the following algorithms: an image processing algorithm based on semantic segmentation corresponding to the slope, an image processing algorithm for identifying broken strands of the power supply line, an image processing algorithm for identifying water leakage corresponding to the drainage pipeline, a temperature monitoring algorithm corresponding to the electric switch, a target detection algorithm for identifying flames and smoke corresponding to the fire, and a target detection algorithm corresponding to personnel and vehicles; The identification module 330 is used to determine the fault or information of the target according to the recognition algorithm corresponding to the target. Optionally, when the target is the slope, when the identification module 330 determines the fault of the target according to the recognition algorithm corresponding to the target, it is specifically used for: Determine the category of each pixel in the inspection image according to the image processing algorithm based on semantic segmentation; Determine whether there are slope ground fissures and landslide bodies on the slope according to the category of each pixel in the inspection image.

[0104] Optionally, when the target object is a power supply line, when the recognition module 330 determines the fault of the target object according to the recognition algorithm corresponding to the target object, it specifically is used for: Perform image cropping on the inspection image according to the image processing algorithm for identifying broken strands of the line to obtain multiple sub-inspection images; Determine the category of each pixel in each of the sub-inspection images according to the image processing algorithm for identifying broken strands of the line, and determine the sub-inspection images where the power supply line exists according to the category of each pixel; Judge whether there is a wire breakage fault in the power supply line according to the sub-inspection images where the power supply line exists; For the power supply line with the wire breakage fault, determine the position of the wire breakage fault in the power supply line.

[0105] Optionally, when the recognition module 330 determines the category of each pixel in the inspection image according to the image processing algorithm for identifying broken strands of the line, it specifically is used for: Obtain the temperature value in the sub-inspection image where the power supply line exists; Determine at least one sub-inspection image where the power supply line has a wire breakage fault according to the temperature value in the sub-inspection image and a preset temperature value, and record it as the first set of sub-inspection images; Determine at least one sub-inspection image where the power supply line has a wire breakage fault according to the image processing algorithm for identifying broken strands of the line, and record it as the second set of sub-inspection images; Judge whether there is a wire breakage fault in the power supply line according to the first set of sub-inspection images and the second set of sub-inspection images.

[0106] Optionally, when the target object is a drainage pipeline, when the recognition module 330 determines the fault of the target object according to the recognition algorithm corresponding to the target object, it specifically is used for: Determine whether there is a leakage fault in the drainage pipeline according to the inspection image and a target detection model corresponding to the leakage fault, where the target detection model corresponding to the leakage fault is obtained by training with images corresponding to drainage pipelines containing leakage faults; When there is a drainage pipeline with a leakage fault in the inspection image, determine the position of the leakage fault in the open-pit mine through the target detection model corresponding to the leakage fault.

[0107] Optionally, when the target object is a switch, when the recognition module 330 determines the fault of the target object according to the recognition algorithm corresponding to the target object, it specifically is used for: Obtain the temperature value in the inspection image including the electric switch; Determine whether there is a fault in the electric switch according to the temperature value and the electric switch temperature warning value corresponding to the electric switch.

[0108] Optionally, when the target is a fire, or personnel and vehicles, when the recognition module 330 determines the fault of the target according to the recognition algorithm corresponding to the target, it is specifically used for: Determine whether the target exists in the inspection image according to the inspection image and the target detection model corresponding to the target, wherein the target detection model corresponding to the target is obtained by training with images including the target; When the inspection image includes the target, determine the target detection model corresponding to the target and determine the position of the target in the open-pit mine.

[0109] Optionally, after the recognition module 330 determines the fault or information of the target according to the recognition algorithm corresponding to the target, the warning module 340 is used for: Output a warning message through the UAV.

[0110] It should be noted that the open-pit mine inspection device based on UAVs in this embodiment is Figure 2 The processing unit 220 corresponding to the corresponding embodiment.

[0111] For the open-pit mine inspection device based on UAVs provided in the embodiments of the present application, the specific implementation process can refer to the above method embodiments, and its implementation principle and technical effects are similar, which will not be elaborated here in this embodiment.

[0112] Figure 4 It is a schematic structural diagram of an electronic device provided in an embodiment of the present application. Among them, the electronic device can be a handheld terminal. For example, a mobile phone is installed with a corresponding application software for realizing the open-pit mine inspection work based on UAVs. As Figure 4 shown, the electronic device includes: a processor 410 and a memory 420.

[0113] Among them, the memory 420 stores computer execution instructions.

[0114] The processor 410 executes the computer execution instructions stored in the memory 420, so that the processor 410 executes the method described in any of the above embodiments.

[0115] For the electronic device provided in the embodiments of the present application, the specific implementation process can refer to the above method embodiments, and its implementation principle and technical effects are similar, which will not be elaborated here in this embodiment.

[0116] In the aboveFigure 4 In the illustrated embodiment, it should be understood that the processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the invention may be directly embodied as being executed and completed by a hardware processor, or may be executed and completed by a combination of hardware and software modules in the processor.

[0117] The memory may include high-speed RAM memory and may also include non-volatile storage NVM, such as at least one disk memory.

[0118] The bus may be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, the buses in the drawings of this application are not limited to only one bus or one type of bus.

[0119] The embodiments of the present application further provide a computer-readable storage medium, in which computer-executable instructions are stored. When the processor executes the computer-executable instructions, the method shown in the above method embodiments is implemented.

[0120] For the above-mentioned computer-readable storage medium, the above-readable storage medium may be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, a disk or an optical disc. The readable storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0121] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can be located in an Application Specific Integrated Circuits (ASIC). Of course, the processor and the readable storage medium can also exist as discrete components in a device.

[0122] Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to program instructions. The foregoing program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps including the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disks, or optical discs that can store program codes.

[0123] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An open pit inspection method based on drones, characterized in that: include: Obtaining an inspection image of the area to be inspected in the open-pit mine, the inspection image being captured by a data acquisition unit carried by the drone when the drone flies over the open-pit mine along a preset inspection route, the name of the preset inspection route including the name of a target object to be captured when the drone flies along the preset inspection route, the target object being at least one of the following: a slope, a power supply line, a drainage pipeline, a switch, a fire, a person, and a vehicle; According to the name of the preset inspection route, determining a recognition algorithm corresponding to at least one target object from a preset algorithm set, wherein the preset algorithm set includes at least one of the following algorithms: an image processing algorithm based on semantic segmentation corresponding to the slope, an image processing algorithm for line break recognition corresponding to the power supply line, an image processing algorithm for water leakage recognition corresponding to the drainage pipeline, a temperature monitoring algorithm corresponding to the switch, a target detection algorithm for flame and smoke recognition corresponding to the fire, and a target detection algorithm corresponding to the personnel and the vehicle; According to the recognition algorithm corresponding to the target object, the fault or information of the target object is determined.

2. The method according to claim 1, characterized in that When the target object is the slope, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Determining the category of each pixel in the inspection image according to the semantic segmentation-based image processing algorithm; According to the category of each pixel in the inspection image, it is determined whether there are ground cracks and landslide bodies on the slope.

3. The method according to claim 1, characterized in that When the target object is a power supply line, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: According to the image processing algorithm for line broken strand recognition, the inspection image is cropped to obtain a plurality of sub-inspection images; Determine the category of each pixel in each of the sub-inspection images according to the image processing algorithm for line broken strand identification, and determine the sub-inspection image in which the power supply line exists according to the category of each pixel; According to the sub-inspection image of the power supply line, determining whether there is a wire breakage fault in the power supply line; For a power supply line having the wire strand break fault, a location of the wire strand break fault in the power supply line is determined.

4. The method according to claim 3, characterized in that The determining, based on the sub-inspection image of the power supply line, whether there is a wire breakage fault in the power supply line includes: Acquire a temperature value in the sub-inspection image where the power supply line exists; According to the temperature value in the sub-inspection image and the preset temperature value, at least one sub-inspection image is determined to indicate that the power supply line has a wire breakage fault, which is recorded as a first sub-inspection image set; According to the image processing algorithm for line broken strand identification, at least one sub-inspection image is determined to indicate that the power supply line has a wire broken strand fault, which is recorded as a second sub-inspection image set; It is determined whether there is a wire breakage fault in the power supply line according to the first sub-inspection image set and the second sub-inspection image set.

5. The method according to claim 1, characterized in that When the target object is a drainage pipe, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Determine whether the drainage pipeline has a water leakage fault according to the inspection image and the target detection model corresponding to the water leakage fault, wherein the target detection model corresponding to the water leakage fault is obtained by training with images corresponding to the drainage pipeline containing the water leakage fault; When there is a drainage pipeline with a water leakage fault in the inspection image, the location of the water leakage fault in the open pit is determined by using the target detection model corresponding to the water leakage fault.

6. The method according to claim 1, characterized in that When the target object is a switch, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Acquiring a temperature value in the inspection image including the switch; It is determined whether the switch has a fault according to the temperature value and a switch temperature warning value corresponding to the switch.

7. The method according to claim 1, characterized in that When the target object is a fire, or a person or a vehicle, determining the fault of the target object according to the recognition algorithm corresponding to the target object includes: Determining whether the target object exists in the inspection image according to the inspection image and a target detection model corresponding to the target object, wherein the target detection model corresponding to the target object is obtained by training an image containing the target object; When the inspection image contains the target object, a target detection model corresponding to the target object is determined to determine the position of the target object in the open pit.

8. The method according to any one of claims 1 to 7, characterized in that: After determining the fault or information of the target object according to the recognition algorithm corresponding to the target object, the method further includes: The warning information is outputted through the UAV.

9. An open pit inspection system based on drones, characterized in that: include: A drone body, a processing unit, a data acquisition unit and a communication unit, wherein the data acquisition unit and the communication unit are both connected to the processing unit, and the data acquisition unit and the communication unit are both mounted on the drone body; The data acquisition unit is used to photograph the area to be inspected in the open pit under the control of the processing unit to obtain an inspection image; The communication unit is used for communication between the drone body, the data acquisition unit and the processing unit; The processing unit is used to execute the method according to any one of claims 1 to 8.

10. An open pit inspection device based on drones, characterized in that: include: an acquisition module, for acquiring an inspection image of the area to be inspected in the open-pit mine, wherein the inspection image is acquired by a data acquisition unit carried on the drone when the drone flies over the open-pit mine along a preset inspection route, wherein the name of the preset inspection route includes the name of a target object to be collected when the drone flies along the preset inspection route, and the target object is at least one of the following: a slope, a power supply line, a drainage pipeline, a switch, a fire, a person, and a vehicle; a determination module, configured to determine, according to the name of the preset inspection route, a recognition algorithm corresponding to at least one target object from a preset algorithm set, wherein the preset algorithm set includes at least one of the following algorithms: an image processing algorithm based on semantic segmentation corresponding to the slope, an image processing algorithm for line break recognition corresponding to the power supply line, an image processing algorithm for water leakage recognition corresponding to the drainage pipeline, a temperature monitoring algorithm corresponding to the switch, a target detection algorithm for flame and smoke recognition corresponding to the fire, and a target detection algorithm corresponding to the personnel and the vehicle; The identification module is used to determine the fault or information of the target object according to the identification algorithm corresponding to the target object.

Images