Inspection method and control system based on inspection robot

Abstract

The present application relates to the technical field of robot inspection, in particular to an inspection method and control system based on an inspection robot.The present application has the beneficial effect that when the inspection robot identifies that the to-be-detected equipment has an abnormal situation under an offline perception model, a help-seeking instruction is generated according to the level parameter of the abnormal situation, the optimal network signal coverage area is screened, the inspection robot is guided to move to the optimal network signal coverage area, and after reaching the network signal coverage area, an alarm signal is sent to the cloud platform; after receiving the response instruction from the cloud platform, the inspection robot continues to perform an inspection task on the to-be-detected equipment that has not been detected according to a preset inspection path, overcomes the problem that the abnormal situation of the equipment cannot be fed back in time in an offline state, and can be autonomously judged by the inspection robot, timely moved to the network signal coverage area, and the help-seeking instruction is sent out, so that the control room can make correct decision instructions in time, and the loss of equipment is avoided.

Description

Technical Field

[0001] This invention relates to the field of robot inspection technology, specifically to inspection methods and control systems based on inspection robots. Background Technology

[0002] Inspection robots are intelligent devices with autonomous navigation, environmental perception, and data analysis capabilities, primarily used to replace manual labor in high-risk, high-intensity, or repetitive inspection tasks. With the maturity of industrial automation and robotics technology, their application scenarios have gradually expanded. Early inspection systems mainly consisted of track-mounted or magnetically attached robots, primarily used for preliminary equipment monitoring in high-risk scenarios such as nuclear power plants and substations. With the rapid development of sensor technology, artificial intelligence, and the Internet of Things, inspection robots have entered the intelligent stage, beginning to possess autonomous navigation, multimodal perception, and real-time analysis capabilities.

[0003] Existing inspection robots can make intelligent decisions for autonomous navigation based on multimodal perception and real-time analysis capabilities. For example, they can automatically make response decisions when abnormalities are detected in equipment. However, this function relies on real-time network transmission. If the inspection robot travels to an area with unstable signal, such as a tunnel, it will cause network instability and slow real-time data transmission, making it impossible to provide timely feedback on abnormalities and thus unable to make timely response decisions, leading to increased losses. Summary of the Invention

[0004] The purpose of this invention is to provide an inspection method and control system based on an inspection robot to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an inspection method based on an inspection robot.

[0006] Step S1: Obtain the target area map and inspection path planning data. The target area map includes spatial location markings of static obstacles, and the inspection path planning data includes preset inspection paths and inspection tasks.

[0007] Step S2: When the network communication of the inspection robot is abnormal, switch to offline perception mode, obtain environmental information of the inspection robot's forward direction through the visual sensor and perform dynamic obstacle detection.

[0008] Step S3: Select the operating plan based on the detection results:

[0009] If no dynamic obstacle is detected, the inspection robot will perform the inspection task along the preset inspection path;

[0010] If a dynamic obstacle is detected, a corrected path is generated and sent to the inspection robot. The corrected path guides the inspection robot to avoid contact with the dynamic obstacle.

[0011] Step S4: When the inspection robot detects an abnormality in the equipment to be inspected, it generates a task execution strategy based on the level parameter of the abnormality.

[0012] In the event of a low-level fault, maintain the current inspection path and record abnormal data;

[0013] In the event of a high-level failure, plan a network communication recovery path and guide the inspection robot to move towards an area with network signal coverage.

[0014] Step S5: Record offline operation data, which is all instructions executed by the inspection robot in offline perception mode; after the network is restored, synchronize the offline operation data to the remote server and update the local environment data.

[0015] Step S1 includes:

[0016] S11. Mark the area to be inspected by the inspection robot as the target environment, periodically collect images of the target area through the image acquisition device and generate a target area map, establish a three-dimensional coordinate system, and mark the spatial position of static obstacles in the target area map. The static obstacles are a certain type of item that is fixedly distributed at a certain place on the inspection path.

[0017] S12. Input the inspection start point and inspection end point, mark them on the target area map, and generate a set of paths connecting the start point to the end point. Select the shortest path in the path set as the preset inspection path, and formulate inspection tasks according to the equipment to be inspected along the preset inspection path. The inspection task is the inspection instruction performed on the equipment to be inspected.

[0018] Step S2 includes:

[0019] The inspection robot's network communication is monitored in real time. When the network communication transmission speed is detected to be lower than the transmission speed threshold and the delay time exceeds the time threshold, it automatically switches to offline sensing mode. In offline sensing mode:

[0020] The visual sensor is activated to acquire images of the inspection robot's forward direction. These images are then transmitted to the local central processing unit for image preprocessing. Dynamic obstacle detection is performed by analyzing the image data. The dynamic obstacle detection process involves: processing the image into grayscale; analyzing whether obstacles are present in the image; and calculating the location of obstacles if they are found.

[0021] Extract the spatial position of the labeled static obstacle. If the position of the static obstacle coincides with the position of the obstacle, the output detection result is "no dynamic obstacle detected"; if the position of the static obstacle does not coincide with the position of the obstacle, the output detection result is "dynamic obstacle detected".

[0022] The vision sensor includes a first binocular camera and a second binocular camera. The first binocular camera is used to capture images directly in front of the inspection robot, and the second binocular camera is located above the inspection robot. The second binocular camera is used to capture images in front of the inspection robot, and the acquisition direction of the first binocular camera forms an angle of less than 90 degrees with the acquisition direction of the second binocular camera.

[0023] The process of calculating the position of the obstacle is as follows: obtain the focal length of the camera and the shooting angle of the camera. In the image captured by the first binocular camera, calculate the horizontal distance from the object to the inspection robot by the position Y1 of the bottom of the object in the image as D1 = H1 × (f1 / Y1), where H1 is the vertical height of the first binocular camera, f1 is the focal length of the first binocular camera, and D1 is the horizontal distance from the object to the inspection robot.

[0024] In the image captured by the second binocular camera, the vertical distance from the object to the inspection robot is calculated by the position Y2 of the bottom of the object in the image as D2 = H2 × (f2 / Y2 × sinθ), where H2 is the vertical height of the second binocular camera, f2 is the focal length of the second binocular camera, D2 is the vertical height from the object to the inspection robot, and θ is the angle between the second binocular camera and the first binocular camera.

[0025] The position coordinates of the obstacle can be obtained from D1 and D2.

[0026] Step S3 includes:

[0027] When no dynamic obstacles are detected, a preset inspection path is obtained, the inspection robot moves along the preset inspection path, and executes the corresponding inspection instructions on the equipment to be inspected that it passes through, thus completing the inspection task.

[0028] When a dynamic obstacle is detected, the current position coordinates of the inspection robot and its distance from the obstacle are obtained in real time. When the distance between the dynamic obstacle and the inspection robot is less than a threshold, the inspection robot is given an avoidance command. The inspection robot generates a corrected path through its built-in algorithm logic to guide the inspection robot to avoid contact with the dynamic obstacle.

[0029] Step S4 includes:

[0030] S41. When an abnormal situation is detected in the equipment, the level parameter of the abnormal situation is calculated. The calculation process is as follows:

[0031] Obtain the fault type corresponding to the abnormal situation and the urgency level corresponding to the fault type. Each fault type has a preset urgency level, which includes low, medium and high levels. Each level of urgency corresponds to its own urgency weight. Also, score the urgency level of each level and use the score result as the urgency value of that level.

[0032] The urgency level is 1-4 for low-level urgency, 5-7 for medium-level urgency, and 8-10 for high-level urgency. The weight of low-level urgency is 0.2, the weight of medium-level urgency is 0.45, and the weight of high-level urgency is 0.68. This data is derived from multiple experiments.

[0033] S42. Retrieve the total number of all abnormal situations detected in this offline sensing mode, as well as the number of devices to be inspected that have been inspected. Divide the total number of abnormal situations by the number of devices to be inspected that have been inspected to calculate the offline failure rate.

[0034] S43. Calculate the level parameter of the abnormal situation that occurred this time. The calculation process is as follows: Level parameter = (urgency value × urgency weight) + (offline failure rate × failure rate weight);

[0035] If the level parameter is greater than the set parameter threshold, the output is a high-level fault; if the level parameter is less than the set parameter threshold, the output is a low-level fault. The parameter threshold can be preset in advance.

[0036] S44. If the output of S43 is a low-level fault, then obtain the preset inspection path and inspection task that the inspection robot has not performed, and continue to perform the inspection task according to the preset inspection path that has not been performed.

[0037] If the S43 output is an advanced fault, a distress command is generated, the optimal network signal coverage area is selected, and the inspection robot is guided to move to the optimal network signal coverage area. After reaching the network signal coverage area, an alarm signal is sent to the cloud platform. After receiving the response command from the cloud platform, the inspection robot continues to perform inspection tasks on the untested equipment according to the preset inspection path.

[0038] The process of selecting the optimal network signal coverage area is as follows:

[0039] Analysis of Candidate Area 1: Obtain the inspection status of the inspection robot during the inspection process. The inspection status includes a preset inspection mode and an offline perception mode. The preset inspection mode is the state when the inspection robot's network communication is normal. Mark the inspection status of the inspection robot on the inspection path it has passed. Divide the inspection path according to the inspection status. The inspection path passed under the preset inspection mode is divided into network communication segments, and the inspection path passed under the offline perception mode is divided into network abnormal segments. Obtain the network communication segment closest to the current position of the inspection robot as the target area. Obtain the position coordinates of the target area. Calculate the path length between the inspection robot and the target area using the coordinates. Calculate the distress call time required to reach the target area based on the inspection robot's travel speed. At the same time, obtain the failure rate of the inspection robot encountering dynamic obstacles during its journey from the current position to the target area. The failure rate is the number of times the inspection robot encountered dynamic obstacles during its journey from the current position to the target area in the historical inspection records divided by the total number of historical inspections.

[0040] Analysis of candidate area 2: Mark the unexplored preset inspection path as the target path, obtain the historical inspection records of the target path, mark the road segment that has not switched to offline perception mode in the historical inspection records as the historical network communication road segment, select the historical network communication road segment closest to the current position of the inspection robot as the target road segment, obtain the position coordinates of the target road segment, calculate the path length between the inspection robot and the target road segment using the coordinates, and calculate the second distress call time required to reach the target road segment based on the inspection robot's travel speed. At the same time, obtain the second failure rate of the inspection robot encountering dynamic obstacles during the process of reaching the target road segment from the current position. The second failure rate is the number of times dynamic obstacles are encountered during the process of reaching the target road segment from the current position in the historical inspection records divided by the total number of historical inspections.

[0041] Analysis of candidate area three: Obtain the location of each artificial base station marked on the target area map. The artificial base station is a service area that is fixed in a certain place and provides services to the inspection robot. Select the artificial base station that is closest to the current inspection robot as the target base station, obtain the location coordinates of the target base station, calculate the path length between the inspection robot and the target base station through the coordinates, calculate the distress call time required to reach the target base station based on the inspection robot's travel speed, and obtain the failure rate three of the inspection robot encountering dynamic obstacles during the process of reaching the target base station from the current position. The failure rate three is the number of times dynamic obstacles are encountered during the process of reaching the target base station from the current position in the historical inspection records divided by the total number of historical inspections.

[0042] S45. Calculate the grade value for each candidate area. The calculation process is as follows: Grade value A = Rescue duration A × Weight 1 + Failure rate A × Weight 2 + Area weight A; where Grade value A ∈ {Grade value of candidate area 1, Grade value of candidate area 2, Grade value of candidate area 3}; Rescue duration A ∈ {Rescue duration 1, Rescue duration 2, Rescue duration 3}; Failure rate A ∈ {Failure rate 1, Failure rate 2, Failure rate 3}; Area weight A ∈ {Area weight 1, Area weight 2, Area weight 3}, and each candidate area corresponds to one area weight.

[0043] The candidate area with the highest grade value is output, where the candidate area ∈ {candidate area one, candidate area two, candidate area three}. A task execution strategy is generated for the inspection robot to move to the candidate area. When communication is restored, the data of the detected abnormal situation is uploaded to the cloud platform and an alarm signal is issued.

[0044] A control system based on an inspection robot, the control system comprising:

[0045] Inspection planning module: acquires target area map and inspection path planning data. The target area map includes spatial location markings of static obstacles, and the inspection path planning data includes preset inspection paths and inspection tasks.

[0046] Mode switching module: When the network communication of the inspection robot is abnormal, it switches to offline perception mode, acquires environmental information of the inspection robot's forward direction through the vision sensor and performs dynamic obstacle detection;

[0047] Operation Decision Module: Selects an operation plan based on the detection results: If no dynamic obstacle is detected, the inspection robot performs the inspection task along the preset inspection path; if a dynamic obstacle is detected, a corrected path is generated and sent to the inspection robot, the corrected path being designed to guide the inspection robot to avoid contact with the dynamic obstacle;

[0048] Anomaly Detection Module: When the inspection robot detects an anomaly in the equipment under inspection, it generates a task execution strategy based on the anomaly severity level parameter.

[0049] In the event of a low-level fault, maintain the current inspection path and record abnormal data;

[0050] In the event of a high-level failure, plan a network communication recovery path and guide the inspection robot to move towards an area with network signal coverage.

[0051] Transmission module: Records offline operation data, which consists of all instructions executed by the inspection robot in offline perception mode; after the network is restored, it synchronizes the offline operation data to the remote server and updates the local environment data.

[0052] The mode switching module includes:

[0053] Image acquisition unit: used to acquire images of the inspection robot's forward direction, transmit the images to the local central processing unit for image preprocessing, and perform dynamic obstacle detection by analyzing the image data;

[0054] Dynamic obstacle analysis unit: The image is processed into grayscale and analyzed to see if there are any obstacles in the image. If there are obstacles, the position of the obstacles is calculated and the spatial position of the labeled static obstacles is extracted. If the position of the static obstacle coincides with the position of the obstacle, the detection result is output as no dynamic obstacle detected; if the position of the static obstacle does not coincide with the position of the obstacle, the detection result is output as a dynamic obstacle detected.

[0055] Network detection unit: Real-time detection of the inspection robot's network communication. When the network communication transmission speed is detected to be lower than the transmission speed threshold and the delay time exceeds the time threshold, it automatically switches to offline sensing mode.

[0056] The anomaly detection module includes:

[0057] Level determination unit: Obtain the fault type corresponding to the abnormal situation and the urgency of the fault type. Each fault type has a preset urgency and calculates the level parameter of the abnormal situation that occurred.

[0058] Output unit: If the level parameter of the abnormal situation corresponds to a low-level fault, then obtain the preset inspection path and inspection task that the inspection robot has not performed, and continue to perform the inspection task according to the preset inspection path that has not been performed.

[0059] If the level parameter of the abnormal situation corresponds to a high-level fault, a distress command is generated, the optimal network signal coverage area is selected, and the inspection robot is guided to move to the optimal network signal coverage area. After reaching the network signal coverage area, an alarm signal is sent to the cloud platform.

[0060] Compared with the prior art, the beneficial effects of the present invention are:

[0061] 1. When an abnormal network communication is detected in the inspection robot, it switches to offline sensing mode. In offline sensing mode, if a dynamic obstacle is detected, a corrected path is generated and sent to the inspection robot. The corrected path guides the inspection robot to avoid contact with the dynamic obstacle. Furthermore, in the offline sensing model, when the inspection robot identifies an abnormal situation in the device to be inspected, it generates a distress signal based on the level parameter of the abnormal situation, selects the optimal network signal coverage area, and guides the inspection robot to move towards the optimal network signal coverage area. After reaching the network signal coverage area, it sends an alarm signal to the cloud platform. After receiving the response command from the cloud platform, the inspection robot continues to perform inspection tasks on the uninspected devices to be inspected according to the preset inspection path. This overcomes the problem of not being able to promptly report device abnormalities in offline mode. The inspection robot can make its own judgment, move towards the network signal coverage area in a timely manner, and send out the distress signal so that the control room can make the correct decision command in a timely manner and avoid further equipment damage. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the inspection method based on inspection robots. Detailed Implementation

[0063] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Example 1: As Figure 1 As shown, this invention provides an inspection method based on an inspection robot:

[0065] Step S1: Obtain the target area map and inspection path planning data. The target area map includes spatial location markings of static obstacles, and the inspection path planning data includes preset inspection paths and inspection tasks.

[0066] Step S2: When the network communication of the inspection robot is abnormal, switch to offline perception mode, obtain environmental information of the inspection robot's forward direction through the visual sensor and perform dynamic obstacle detection.

[0067] Step S3: Select the operating plan based on the detection results:

[0068] If no dynamic obstacle is detected, the inspection robot will perform the inspection task along the preset inspection path;

[0069] If a dynamic obstacle is detected, a corrected path is generated and sent to the inspection robot. The corrected path guides the inspection robot to avoid contact with the dynamic obstacle.

[0070] Step S4: When the inspection robot detects an abnormality in the equipment to be inspected, it generates a task execution strategy based on the level parameter of the abnormality.

[0071] In the event of a low-level fault, maintain the current inspection path and record abnormal data;

[0072] In the event of a high-level failure, plan a network communication recovery path and guide the inspection robot to move towards an area with network signal coverage.

[0073] Step S5: Record offline operation data, which is all instructions executed by the inspection robot in offline perception mode; after the network is restored, synchronize the offline operation data to the remote server and update the local environment data.

[0074] Step S1 includes:

[0075] S11. Mark the area to be inspected by the inspection robot as the target environment, periodically collect images of the target area through the image acquisition device and generate a target area map, establish a three-dimensional coordinate system, and mark the spatial position of static obstacles in the target area map. The static obstacles are a certain type of item that is fixedly distributed at a certain place on the inspection path.

[0076] S12. Input the inspection start point and inspection end point, mark them on the target area map, and generate a set of paths connecting the start point to the end point. Select the shortest path in the path set as the preset inspection path, and formulate inspection tasks according to the equipment to be inspected along the preset inspection path. The inspection task is the inspection instruction performed on the equipment to be inspected.

[0077] Step S2 includes:

[0078] The inspection robot's network communication is monitored in real time. When the network communication transmission speed is detected to be lower than the transmission speed threshold and the delay time exceeds the time threshold, it automatically switches to offline sensing mode. In offline sensing mode:

[0079] The visual sensor is activated to acquire images of the inspection robot's forward direction. The images are then transmitted to the local central processing unit for image preprocessing. Dynamic obstacle detection is performed by analyzing the image data. The dynamic obstacle detection process is as follows: the image is processed into grayscale, and the presence of obstacles in the image is analyzed. If obstacles are present, their positions are calculated, and the spatial positions of marked static obstacles are extracted. If the positions of the static and static obstacles coincide, the detection result is output as "no dynamic obstacle detected." If the positions of the static and static obstacles do not coincide, the detection result is output as "dynamic obstacle detected."

[0080] The vision sensor includes a first binocular camera and a second binocular camera. The first binocular camera is used to capture images directly in front of the inspection robot, and the second binocular camera is located above the inspection robot. The second binocular camera is used to capture images in front of the inspection robot, and the acquisition direction of the first binocular camera forms an angle of less than 90 degrees with the acquisition direction of the second binocular camera.

[0081] Step S3 includes:

[0082] When no dynamic obstacles are detected, a preset inspection path is obtained, the inspection robot moves along the preset inspection path, and executes the corresponding inspection instructions on the equipment to be inspected that it passes through, thus completing the inspection task.

[0083] When a dynamic obstacle is detected, the current position coordinates of the inspection robot and its distance from the obstacle are obtained in real time. When the distance between the dynamic obstacle and the inspection robot is less than a threshold, the inspection robot is given an avoidance command. The inspection robot generates a corrected path through its built-in algorithm logic to guide the inspection robot to avoid contact with the dynamic obstacle.

[0084] Step S4 includes:

[0085] S41. When an abnormal situation is detected in the equipment, the level parameter of the abnormal situation is calculated. The calculation process is as follows:

[0086] Obtain the fault type corresponding to the abnormal situation and the urgency level corresponding to the fault type. Each fault type has a preset urgency level, which includes low, medium and high levels. Each level of urgency corresponds to its own urgency weight. Also, score the urgency level of each level and use the score result as the urgency value of that level.

[0087] S42. Retrieve the total number of all abnormal situations detected in this offline sensing mode, as well as the number of devices to be inspected that have been inspected. Divide the total number of abnormal situations by the number of devices to be inspected that have been inspected to calculate the offline failure rate.

[0088] S43. Calculate the level parameter of the abnormal situation that occurred this time. The calculation process is as follows: Level parameter = (urgency value × urgency weight) + (offline failure rate × failure rate weight);

[0089] If the level parameter is greater than the set parameter threshold, the output is a high-level fault; if the level parameter is less than the set parameter threshold, the output is a low-level fault.

[0090] S44. If the output of S43 is a low-level fault, then obtain the preset inspection path and inspection task that the inspection robot has not performed, and continue to perform the inspection task according to the preset inspection path that has not been performed.

[0091] If the S43 output is an advanced fault, a distress command is generated, the optimal network signal coverage area is selected, and the inspection robot is guided to move to the optimal network signal coverage area. After reaching the network signal coverage area, an alarm signal is sent to the cloud platform. After receiving the response command from the cloud platform, the inspection robot continues to perform inspection tasks on the untested equipment according to the preset inspection path.

[0092] The process of selecting the optimal network signal coverage area is as follows:

[0093] Analysis of Candidate Area 1: Obtain the inspection status of the inspection robot during the inspection process. The inspection status includes a preset inspection mode and an offline perception mode. The preset inspection mode is the state when the inspection robot's network communication is normal. Mark the inspection status of the inspection robot on the inspection path it has passed. Divide the inspection path according to the inspection status. The inspection path passed under the preset inspection mode is divided into network communication segments, and the inspection path passed under the offline perception mode is divided into network abnormal segments. Obtain the network communication segment closest to the current position of the inspection robot as the target area. Obtain the position coordinates of the target area. Calculate the path length between the inspection robot and the target area using the coordinates. Calculate the distress call time required to reach the target area based on the inspection robot's travel speed. At the same time, obtain the failure rate of the inspection robot encountering dynamic obstacles during its journey from the current position to the target area. The failure rate is the number of times the inspection robot encountered dynamic obstacles during its journey from the current position to the target area in the historical inspection records divided by the total number of historical inspections.

[0094] Analysis of candidate area 2: Mark the unexplored preset inspection path as the target path, obtain the historical inspection records of the target path, mark the road segment that has not switched to offline perception mode in the historical inspection records as the historical network communication road segment, select the historical network communication road segment closest to the current position of the inspection robot as the target road segment, obtain the position coordinates of the target road segment, calculate the path length between the inspection robot and the target road segment using the coordinates, and calculate the second distress call time required to reach the target road segment based on the inspection robot's travel speed. At the same time, obtain the second failure rate of the inspection robot encountering dynamic obstacles during the process of reaching the target road segment from the current position. The second failure rate is the number of times dynamic obstacles are encountered during the process of reaching the target road segment from the current position in the historical inspection records divided by the total number of historical inspections.

[0095] Analysis of candidate area three: Obtain the location of each artificial base station marked on the target area map. The artificial base station is a service area that is fixed in a certain place and provides services to the inspection robot. Select the artificial base station that is closest to the current inspection robot as the target base station, obtain the location coordinates of the target base station, calculate the path length between the inspection robot and the target base station through the coordinates, calculate the distress call time required to reach the target base station based on the inspection robot's travel speed, and obtain the failure rate three of the inspection robot encountering dynamic obstacles during the process of reaching the target base station from the current position. The failure rate three is the number of times dynamic obstacles are encountered during the process of reaching the target base station from the current position in the historical inspection records divided by the total number of historical inspections.

[0096] S45. Calculate the grade value for each candidate area. The calculation process is as follows: Grade value A = Rescue duration A × Weight 1 + Failure rate A × Weight 2 + Area weight A; where Grade value A ∈ {Grade value of candidate area 1, Grade value of candidate area 2, Grade value of candidate area 3}; Rescue duration A ∈ {Rescue duration 1, Rescue duration 2, Rescue duration 3}; Failure rate A ∈ {Failure rate 1, Failure rate 2, Failure rate 3}; Area weight A ∈ {Area weight 1, Area weight 2, Area weight 3}, each candidate area corresponds to one area weight, and the sum of Area weight 1 + Area weight 2 + Area weight 3 is 1;

[0097] The candidate area with the highest grade value is output, where the candidate area ∈ {candidate area one, candidate area two, candidate area three}. A task execution strategy is generated for the inspection robot to move to the candidate area. When communication is restored, the data of the detected abnormal situation is uploaded to the cloud platform and an alarm signal is issued.

[0098] Example 2: A control system based on an inspection robot, the control system comprising:

[0099] Inspection planning module: acquires target area map and inspection path planning data. The target area map includes spatial location markings of static obstacles, and the inspection path planning data includes preset inspection paths and inspection tasks.

[0100] Mode switching module: When the network communication of the inspection robot is abnormal, it switches to offline perception mode, acquires environmental information of the inspection robot's forward direction through the vision sensor and performs dynamic obstacle detection;

[0101] Operation Decision Module: Selects an operation plan based on the detection results: If no dynamic obstacle is detected, the inspection robot performs the inspection task along the preset inspection path; if a dynamic obstacle is detected, a corrected path is generated and sent to the inspection robot, the corrected path being designed to guide the inspection robot to avoid contact with the dynamic obstacle;

[0102] Anomaly Detection Module: When the inspection robot detects an anomaly in the equipment under inspection, it generates a task execution strategy based on the anomaly severity level parameter.

[0103] In the event of a low-level fault, maintain the current inspection path and record abnormal data;

[0104] In the event of a high-level failure, plan a network communication recovery path and guide the inspection robot to move towards an area with network signal coverage.

[0105] Transmission module: Records offline operation data, which consists of all instructions executed by the inspection robot in offline perception mode; after the network is restored, it synchronizes the offline operation data to the remote server and updates the local environment data.

[0106] The mode switching module includes:

[0107] Image acquisition unit: used to acquire images of the inspection robot's forward direction, transmit the images to the local central processing unit for image preprocessing, and perform dynamic obstacle detection by analyzing the image data;

[0108] Dynamic obstacle analysis unit: The image is processed into grayscale and analyzed to see if there are any obstacles in the image. If there are obstacles, the position of the obstacles is calculated and the spatial position of the labeled static obstacles is extracted. If the position of the static obstacle coincides with the position of the obstacle, the detection result is output as no dynamic obstacle detected; if the position of the static obstacle does not coincide with the position of the obstacle, the detection result is output as a dynamic obstacle detected.

[0109] Network detection unit: Real-time detection of the inspection robot's network communication. When the network communication transmission speed is detected to be lower than the transmission speed threshold and the delay time exceeds the time threshold, it automatically switches to offline sensing mode.

[0110] The anomaly detection module includes:

[0111] Level determination unit: Obtain the fault type corresponding to the abnormal situation and the urgency of the fault type. Each fault type has a preset urgency and calculates the level parameter of the abnormal situation that occurred.

[0112] Output unit: If the level parameter of the abnormal situation corresponds to a low-level fault, then obtain the preset inspection path and inspection task that the inspection robot has not performed, and continue to perform the inspection task according to the preset inspection path that has not been performed.

[0113] If the level parameter of the abnormal situation corresponds to a high-level fault, a distress command is generated, the optimal network signal coverage area is selected, and the inspection robot is guided to move to the optimal network signal coverage area. After reaching the network signal coverage area, an alarm signal is sent to the cloud platform.

[0114] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for inspection based on a patrol robot, characterized in that: Step S1, obtaining a target area map and inspection path planning data, wherein the target area map contains the spatial position label of static obstacles, and the inspection path planning data includes a preset inspection path and an inspection task; Step S2, when the network communication of the patrol robot is abnormal, switching to an offline sensing mode, obtaining environmental information in the forward direction of the patrol robot through a visual sensor and performing dynamic obstacle detection; Step S3, selecting a running scheme based on the detection result: If no dynamic obstacle is detected, the patrol robot performs the inspection task along the preset inspection path; If a dynamic obstacle is detected, a modified path is generated and sent to the patrol robot, wherein the modified path guides the patrol robot to avoid contact with the dynamic obstacle; Step S4, when the patrol robot identifies that an abnormal situation occurs in a device to be detected, generating a task execution strategy according to the level parameter of the abnormal situation: In the case of a low-level fault, maintaining the current inspection path and recording abnormal data; In the case of a high-level fault, planning a network communication recovery path to guide the patrol robot to move to a network signal coverage area; Step S5, recording offline operation data, which is all instructions executed by the patrol robot in the offline sensing mode; After the network is restored, synchronizing the offline operation data to a remote server and updating local environmental data; The step S4 includes: S41, when an abnormal situation is detected, calculating the level parameter of the abnormal situation, the calculation process being: Obtaining the fault type corresponding to the abnormal situation and the urgency degree corresponding to the fault type, each type of fault type being provided with a corresponding urgency degree, the urgency degree including low, medium and high levels, each level of the urgency degree corresponding to a respective urgency degree weight, and scoring each level of the urgency degree, the scoring result being taken as the urgency degree value of the level; S42, calling the total number of abnormal situations detected in the offline sensing mode and the number of devices to be detected that have been inspected, and calculating the offline fault rate by dividing the total number of abnormal situations by the number of devices to be detected that have been inspected; S43, calculating the level parameter of the abnormal situation, the calculation process being: level parameter = (urgency degree value × urgency degree weight) + (offline fault rate × fault rate weight); If the level parameter is greater than a set parameter threshold, the output is a high-level fault, and if the level parameter is less than the set parameter threshold, the output is a low-level fault; S44, if the output of S43 is a low-level fault, obtaining the preset inspection path and inspection task that have not been performed by the patrol robot, and continuing to perform the inspection task according to the preset inspection path that has not been performed; If the output of S43 is a high-level fault, generating a help-seeking instruction, selecting the optimal network signal coverage area, guiding the patrol robot to move to the optimal network signal coverage area, and after reaching the network signal coverage area, sending an alarm signal to a cloud platform; after receiving the response instruction from the cloud platform, the patrol robot continues to perform the inspection task on the devices to be detected that have not been detected according to the preset inspection path. 2.The method of claim 1, wherein: The step S1 includes: S11, mark the area responsible for the inspection of the inspection robot as a target environment, periodically collect images of the target area through an image collection device and generate a target area map, establish a three-dimensional coordinate system, and mark the spatial position of static obstacles in the target area map, the static obstacles being certain articles fixedly distributed at a certain place on the inspection path; S12, input the inspection starting point and the inspection ending point, mark them out on the target area map, generate a path set connecting the starting point to the ending point, select the shortest path in the path set as the preset inspection path, and formulate an inspection task according to the to-be-detected equipment passed through by the preset inspection path, the inspection task being a detection instruction performed on the to-be-detected equipment. 3.The method of claim 1, wherein: Step S2 includes: Real-time detection of network communication of the inspection robot, when the transmission speed of the network communication is detected to be lower than a transmission speed threshold and the delay time exceeds a time threshold, automatically switching to an offline perception mode, in the offline perception mode: starting a visual sensor to collect images of the forward direction of the inspection robot, transmitting the images to a local central processor for image preprocessing, and performing dynamic obstacle detection by analyzing the image data, the dynamic obstacle detection process being: performing grayscale processing on the images, analyzing whether there is an obstacle picture in the images, if there is an obstacle picture, calculating the position of the obstacle, extracting the spatial position of the labeled static obstacle, if the spatial position of the static obstacle coincides with the position of the obstacle, outputting a detection result that no dynamic obstacle is detected; if the spatial position of the static obstacle does not coincide with the position of the obstacle, outputting a detection result that a dynamic obstacle is detected; The visual sensor includes a first binocular camera and a second binocular camera, the first binocular camera is used to collect images of the front of the inspection robot, the second binocular camera is located above the inspection robot, the second binocular camera is used to collect images of the front of the inspection robot, and the collection direction of the first binocular camera and the collection direction of the second binocular camera form an included angle less than ninety degrees. 4.The method of claim 1, wherein: The step S3 includes: When no dynamic obstacle is detected, a preset inspection path is obtained, the inspection robot moves along the preset inspection path, and corresponding inspection instructions are executed on the to-be-detected equipment passed through, to complete the inspection task; When a dynamic obstacle is detected, the current position coordinates of the inspection robot and the distance between the dynamic obstacle and the inspection robot are obtained in real time, when the distance between the dynamic obstacle and the inspection robot is less than a threshold, an instruction for the inspection robot to avoid is triggered, the inspection robot generates a corrected path through the built-in algorithm logic to guide the inspection robot not to contact the dynamic obstacle.

5. The method of claim 1, wherein: The process of screening the optimal network signal coverage area is: analyze the inspection state of the inspection robot in the inspection process in the to-be-selected area: the inspection state includes a preset inspection mode and an offline perception mode, wherein the preset inspection mode is the state when the network communication of the inspection robot is normal; The inspection state of the inspection robot is marked on the passed inspection path, the inspection path is divided according to the inspection state, the inspection path passed in the preset inspection mode is divided into a network communication section, and the inspection path passed in the offline perception model is divided into a network anomaly section; A network communication section closest to the current position of the inspection robot is obtained as a target region, the position coordinates of the target region are obtained, the path length between the inspection robot and the target region is calculated through the coordinates, the rescue time required to reach the target region is calculated according to the driving speed of the inspection robot, and a failure rate one of encountering a dynamic obstacle in the process of the inspection robot from the current position to the target region is obtained, the failure rate one being the number of times of encountering the dynamic obstacle in the process of reaching the target region from the current position in the historical inspection record divided by the total number of historical inspections; The target path is marked as a target path, the historical inspection record of the target path is obtained, the section in the historical inspection record that does not appear to switch to the offline perception mode is marked as a historical network communication section, a historical network communication section closest to the current position of the inspection robot is selected as a target section, the position coordinates of the target section are obtained, the path length between the inspection robot and the target section is calculated through the coordinates, the rescue time required to reach the target section is calculated according to the driving speed of the inspection robot, and a failure rate two of encountering a dynamic obstacle in the process of the inspection robot from the current position to the target section is obtained, the failure rate two being the number of times of encountering the dynamic obstacle in the process of reaching the target section from the current position in the historical inspection record divided by the total number of historical inspections; The positions of the artificial base stations marked in the target region map are obtained, the artificial base station is a service region in which a fixed base station is arranged and provides services for the inspection robot, a closest artificial base station to the current inspection robot is selected as a target base station, the position coordinates of the target base station are obtained, the path length between the inspection robot and the target base station is calculated through the coordinates, the rescue time required to reach the target base station is calculated according to the driving speed of the inspection robot, and a failure rate three of encountering a dynamic obstacle in the process of the inspection robot from the current position to the target base station is obtained, the failure rate three being the number of times of encountering the dynamic obstacle in the process of reaching the target base station from the current position in the historical inspection record divided by the total number of historical inspections; S45, and the grade value of each candidate region is calculated, and the calculation process is: grade value A = rescue time A x weight one + failure rate A x weight two + region weight A; Wherein, the grade value A is in {the grade value of the candidate region one, the grade value of the candidate region two, and the grade value of the candidate region three}; the rescue time A is in {the rescue time one, the rescue time two, and the rescue time three}; the failure rate A is in {the failure rate one, the failure rate two, and the failure rate three}; and the region weight A is in {the region weight one, the region weight two, and the region weight three}, each candidate region corresponds to a region weight. Output the candidate region with the highest level value, and generate a task execution strategy for the inspection robot to move to the candidate region, and upload the detected abnormal condition data to the cloud platform and send a warning signal when the communication is restored.

6. A control system based on a patrol robot, applied to the patrol method based on a patrol robot according to any one of claims 1-5, characterized in that: The control system comprises: An inspection planning module for obtaining a target area map and inspection path planning data, wherein the target area map comprises a spatial position label of static obstacles, and the inspection path planning data comprises a preset inspection path and an inspection task; A mode switching module for switching to an offline sensing mode when the network communication of the inspection robot is abnormal, acquiring environmental information in the forward direction of the inspection robot through a visual sensor, and performing dynamic obstacle detection; An operation decision module for selecting an operation scheme based on the detection result: if no dynamic obstacle is detected, the inspection robot performs the inspection task along the preset inspection path; if a dynamic obstacle is detected, a correction path is generated and sent to the inspection robot, wherein the correction path guides the inspection robot to avoid contact with the dynamic obstacle; An abnormality judgment module for generating a task execution strategy according to the level parameter of the abnormal condition when the inspection robot identifies that the equipment to be detected has an abnormal condition: Maintain the current inspection path and record abnormal data when a low-level fault occurs; Plan a network communication recovery path and guide the inspection robot to move to a network signal coverage area when a high-level fault occurs; A transmission module for recording offline operation data, which is all instructions executed by the inspection robot in the offline sensing mode; and synchronizing the offline operation data to a remote server and updating local environmental data after the network is restored.

7. The control system based on a patrol robot according to claim 6, characterized in that: The mode switching module comprises: An image acquisition unit for acquiring images in the forward direction of the inspection robot, transmitting the images to a local central processor for image preprocessing, and performing dynamic obstacle detection by analyzing image data; A dynamic obstacle analysis unit for performing grayscale processing on the images, analyzing whether there is an obstacle in the images, calculating the position of the obstacle if there is an obstacle, extracting the spatial position of the labeled static obstacle, and outputting a detection result that no dynamic obstacle is detected if the position of the static obstacle coincides with the position of the obstacle; or outputting a detection result that a dynamic obstacle is detected if the position of the static obstacle does not coincide with the position of the obstacle; A network detection unit for detecting the network communication of the inspection robot in real time, and automatically switching to the offline sensing mode when the transmission speed of the network communication is lower than a transmission speed threshold and the delay time exceeds a time threshold.

8. The control system based on a patrol robot according to claim 6, characterized in that: The abnormality judgment module comprises: A level judgment unit for obtaining the fault type corresponding to the abnormal condition and the urgency corresponding to the fault type, presetting a corresponding urgency for each fault type, and calculating the level parameter of the current abnormal condition; An output unit for obtaining the preset inspection path and the inspection task that have not been performed by the inspection robot if the level parameter of the abnormal condition corresponds to an output of a low-level fault, and continuing to perform the inspection task according to the unperformed preset inspection path. If the level parameter corresponding to the output of the abnormal situation is a high-level fault, a help-seeking instruction is generated, the optimal network signal coverage area is screened, the inspection robot is guided to move to the optimal network signal coverage area, and after reaching the network signal coverage area, an alarm signal is sent to the cloud platform.

Images