A comprehensive inspection path intelligent planning method, system and storage medium

Abstract

The present invention discloses a comprehensive patrol route intelligent planning method, system and storage medium, which relates to the field of data processing technology, including determining the grid size according to the automatic patrol equipment structure information, patrol target information, road signal interference and building information, and obtaining the grid interference information around the road; screening multiple patrol routes based on the starting point coordinates, interference information and end point coordinate lists; the automatic patrol equipment patrols according to the patrol route with the lowest cost; and optimizing the neural network model by comparing the predicted value and the actual value of the automatic patrol equipment position by checking nodes, and replanning the patrol route. The accuracy of the grid size is improved, and the interference information around the road is effectively considered. The intelligent planning of the path is realized. When the path taken is not suitable, another suitable path can be selected, which improves the adaptability and accuracy of the path planning and ensures the safe and reliable operation of the automatic patrol equipment.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and more specifically, to a comprehensive inspection path intelligent planning method, system and storage medium. Background Art

[0002] With the development of industrial automation and intelligence, inspection work has become increasingly important in various fields. Traditional inspection methods rely on manual operations, which have problems such as low efficiency, high cost, and poor safety. In order to solve these problems, researchers have proposed inspection path planning methods based on artificial intelligence. These methods use computer algorithms and sensor technology to automatically identify inspection targets, predict the location of inspection equipment, and optimize inspection paths.

[0003] In the prior art, before planning the inspection route, the interference factors near the inspection area will be considered, and the influencing data in the area will be gridded, such as signal interference, road congestion, building interference, etc. However, if all the influencing data are considered, the grid size will be too small, which will make the processing more complicated, and there will be a lot of invalid data. Invalid data means that the influencing data will not interfere with the selection of the automatic inspection route. If only a part of the influencing data is considered, the grid size will be too large, and there may be incomplete analysis of the influencing data, resulting in low accuracy of the influencing data analysis. Therefore, choosing a suitable grid size is crucial to the accuracy of planning. Furthermore, when judging the inspection end point, the location of the inspection target farthest away is often selected as the end point. Such a preconceived choice has a certain degree of blindness and may not be the most suitable end point. Finally, after the automatic inspection equipment selects the route, it can only inspect according to the route, but because the road conditions are complicated, the predicted inspection path may not be suitable for the actual situation. At this time, it may be necessary to update the prediction model and re-judge and select the path based on the updated prediction model. The above problems will lead to low adaptability and accuracy of inspection path planning.

[0004] Therefore, how to improve the adaptability and accuracy of inspection path planning is a technical problem that needs to be solved. Summary of the invention

[0005] The present invention provides a comprehensive inspection path intelligent planning method to solve the technical problems of poor adaptability and low accuracy of the inspection path in the prior art. The method is applied to automatic inspection equipment, and the method includes:

[0006] Obtain complete map information within the inspection area, establish a two-dimensional coordinate plane map, and collect interference information around the road based on the plane map. The interference information includes road congestion, road signal interference, inspection target information, and building information;

[0007] The grid size is determined according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, and the center radius is determined according to the road congestion. The center line of the road on the plane map is used as the circle point, and the area within the center radius is divided according to the grid size, and the interference information at the corresponding position is filled in, so as to obtain the grid interference information around the road;

[0008] Confirm the starting point coordinates of the automatic inspection equipment, establish a destination coordinate list according to the inspection target information, filter multiple inspection paths based on the starting point coordinates, interference information and destination coordinate list, and determine the change nodes in the multiple inspection paths;

[0009] The automatic inspection equipment conducts inspections along the inspection path with the lowest total cost. A neural network model is established based on the grid interference information around the road and the previous location data of the automatic inspection equipment to predict the location of the automatic inspection equipment.

[0010] Multiple proofreading nodes are set in the inspection path. The predicted value and actual value of the automatic inspection equipment position are compared through the proofreading nodes to optimize the neural network model and re-plan the inspection path.

[0011] In some embodiments of the present application, the grid size is determined according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, including:

[0012] The automatic inspection equipment structure information includes the inspection equipment area and the observable range of the inspection equipment. The inspection target information includes multiple target locations, areas and attributes. The road signal interference includes the signal interference coverage, signal interference intensity and signal fluctuation. The building information includes multiple building locations and areas.

[0013] Determine a comprehensive inspection target index through multiple target locations, areas and attributes;

[0014]

[0015] Among them, Q1 is the comprehensive index of the inspection target, n is the number of target locations, G1 i →G2 i is the weight of the automatic inspection device identifying the target mapped from the i-th target area to the i-th target area, m is the number of paths for the automatic inspection device corresponding to the i-th target to move to the target location, M j is the distance of the sum of the paths of the jth path, the target attributes include type and importance level, B is the importance obtained by weighted summation of type and importance level, exp is an exponential function, and k1 is a first preset constant;

[0016] A signal interference comprehensive index Q2 is determined according to the signal interference coverage, signal interference strength and signal fluctuation, and a building comprehensive index Q3 is determined according to the locations and areas of multiple buildings;

[0017] The grid size is determined by the inspection equipment area, the observable range of the inspection equipment, the inspection target comprehensive index Q1, the signal interference comprehensive index Q2, and the building comprehensive index Q3;

[0018]

[0019] Among them, L i K is the partial grid coefficient corresponding to the inspection target comprehensive index, signal interference comprehensive index or building comprehensive index, i is the constant corresponding to the inspection target comprehensive index, the signal interference comprehensive index or the building comprehensive index, s is the grid area, α1 is the partial weight corresponding to the inspection target comprehensive index, L1 is the partial grid coefficient of the inspection target comprehensive index, α2 is the partial weight corresponding to the signal interference comprehensive index, L2 is the partial grid coefficient of the signal interference comprehensive index, α3 is the partial weight corresponding to the building comprehensive index, L3 is the partial grid coefficient of the building comprehensive index, s1 is the area of ​​the observable range of the inspection equipment, and s2 is the inspection equipment area.

[0020] In some embodiments of the present application, the center radius is determined according to the road congestion, the center line of the road on the plane map is used as the circle point, the area within the center radius is divided according to the grid size, and the interference information at the corresponding position is filled, including:

[0021] A circle center radius is determined according to the average value of the road congestion. Different average values ​​of different road congestion correspond to different circle center radii. A circle is drawn according to the center line of the road on the plane map as the circle point and the circle center radius so that the circular area completely contains the road, thereby obtaining a circular area.

[0022] The circular area is meshed according to the grid size, and the grids that exceed the circular area are also planned into the circular area;

[0023] The coordinate position information of each grid area is determined, and the road congestion, road signal interference and building information are filled in accordingly through the coordinate position information.

[0024] In some embodiments of the present application, a destination coordinate list is established according to the inspection target information, including:

[0025] Determine the coordinates of multiple target locations, and calculate the straight-line distance and path distance between the starting point coordinates of the automatic inspection device and the coordinates of the multiple target locations;

[0026] The weights of the straight-line distance and the path distance are assigned according to the importance of each target, and the measured distance is obtained by weighted summation;

[0027] Find the coordinates of the diagonal points of the starting point coordinates of the automatic inspection device on the plane map, calculate the straight-line distance, determine the first measurement distance threshold and the second measurement distance threshold according to the straight-line distance, and put multiple target positions whose measurement distance exceeds the first measurement distance threshold as end points into the end point coordinate list;

[0028] A plurality of target positions whose measured distances exceed the second measured distance threshold and do not exceed the first measured distance threshold are put into the endpoint coordinate list as candidate endpoints.

[0029] In some embodiments of the present application, multiple inspection paths are screened based on the starting point coordinates, interference information, and the end point coordinate list, and the change nodes in the multiple inspection paths are determined, including:

[0030] Step 1: define nodes and edges on a plane map, where nodes are starting points, road intersections, and end points, and edges are moving paths from the starting point to the road intersection, moving paths between road intersections, and moving paths between the road intersection and the end point;

[0031] Step 2: Adjust the heuristic function according to the interference information and estimate the heuristic estimated cost h from the current node to the end point;

[0032] Step 3: Initialize the open list and the closed list, add the starting point to the open list, and set f = g + h, where g is the actual cost from the current node to the end point;

[0033] Step 4: Perform a cyclic search and update the values ​​of g, h and f;

[0034] Step 5: If the current node is the end point, determine a path; otherwise, add the current node to the closed list and repeat step 4.

[0035] Step 6: Repeat steps 3-5 to obtain multiple paths;

[0036] The paths that meet the preset f threshold are retained to obtain the preferred path, and the change nodes on the preferred path are determined. The change nodes are nodes that can be transferred from a certain node on one preferred path to another preferred path.

[0037] In some embodiments of the present application, multiple proofreading nodes are set in the inspection path, including:

[0038] The plane map is divided into multiple areas, and the corresponding key areas are determined according to road congestion, road signal interference, inspection target information and building information. Several locations are randomly selected in multiple key areas, and the time nodes for the automatic inspection equipment to arrive at these locations are predicted and used as proofreading nodes.

[0039] In some embodiments of the present application, the neural network model is optimized by comparing the predicted value and the actual value of the position of the automatic inspection equipment through the calibration node, and the inspection path is replanned, including:

[0040]

[0041] Among them, x t is the actual value of the position of the automatic inspection equipment under the t calibration node, is the predicted value of the automatic inspection equipment position under the t calibration node, W old is the weight of the neural network model before updating, X1 and X2 are the preset position error values, and W new is the updated neural network model weight, is the partial derivative of the loss function MSE with respect to the weight W, MSE is the loss function, exp is the exponential function, is the learning factor of the difference mapping;

[0042] Re-predict the subsequent time-position situation of the automatic inspection equipment under the inspection path according to the optimized neural network model, and determine whether the inspection path is still the optimal solution. If it is the optimal solution, continue the inspection of the inspection path;

[0043] Otherwise, the inspection path corresponding to the optimal solution is changed according to the changed nodes on the inspection path.

[0044] Correspondingly, the present application also provides a comprehensive inspection path intelligent planning system, which is applied to automatic inspection equipment, including:

[0045] The collection module is used to obtain the complete map information of the inspection area, establish a two-dimensional coordinate plane map, and collect interference information around the road based on the plane map. The interference information includes road congestion, road signal interference, inspection target information and building information;

[0046] The grid module is used to determine the grid size according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, determine the center radius according to the road congestion, take the center line of the road on the plane map as the circle point, divide the area within the center radius according to the grid size, and fill in the interference information at the corresponding position, so as to obtain the grid interference information around the road;

[0047] The optimization module is used to confirm the starting point coordinates of the automatic inspection equipment, establish a list of end point coordinates according to the inspection target information, screen multiple inspection paths based on the starting point coordinates, interference information and the end point coordinate list, and determine the change nodes in the multiple inspection paths;

[0048] A construction module is used for the automatic inspection equipment to inspect along the inspection path with the lowest total cost, and a neural network model is established based on the grid interference information around the road and the previous location data of the automatic inspection equipment to predict the location of the automatic inspection equipment;

[0049] The optimization module is used to set multiple proofreading nodes within the inspection path, compare the predicted value and actual value of the automatic inspection equipment position through the proofreading nodes, optimize the neural network model, and re-plan the inspection path.

[0050] A computer-readable storage medium stores one or more programs, which, when executed by an electronic device including a plurality of application programs, cause the electronic device to execute the above method.

[0051] By applying the above technical solutions, the complete map information of the inspection area is obtained, a two-dimensional coordinate plane map is established, and interference information around the road is collected based on the plane map. The interference information includes road congestion, road signal interference, inspection target information and building information; the grid size is determined according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, and the center radius is determined according to the road congestion. The center line of the road on the plane map is used as the circle point, and the area within the center radius is divided according to the grid size, and the interference information at the corresponding position is filled in, so as to obtain the grid interference around the road. information; confirm the starting point coordinates of the automatic inspection equipment, establish a list of end point coordinates according to the inspection target information, screen multiple inspection paths based on the starting point coordinates, interference information and end point coordinate list, and determine the change nodes in the multiple inspection paths; the automatic inspection equipment inspects according to the inspection path with the lowest total cost, and establishes a neural network model for predicting the position of the automatic inspection equipment based on the grid interference information around the road and the previous position data of the automatic inspection equipment; set up multiple proofreading nodes in the inspection path, compare the predicted value and actual value of the automatic inspection equipment position through the proofreading nodes to optimize the neural network model, and re-plan the inspection path.

[0052] Compared with the prior art, the present invention has the following significant improvements:

[0053] This application determines the appropriate grid size based on the inspection equipment conditions and road interference conditions, thereby effectively gridding interference information, improving data accuracy, and avoiding redundant data. Multiple endpoints and alternative endpoints are screened out based on the straight-line distance and path distance in the inspection target location information to ensure the reliability of endpoint selection. The neural network model is optimized by comparing the predicted value and actual value of the automatic inspection equipment position through proofreading nodes, and the inspection path is replanned. Intelligent path planning is achieved. When the path taken is not suitable, another suitable path can be selected, which improves the adaptability and accuracy of path planning and ensures the safe and reliable operation of automatic inspection equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0055] Figure 1 A schematic diagram of a process flow of a comprehensive inspection path intelligent planning method proposed in an embodiment of the present invention is shown;

[0056] Figure 2 A structural schematic diagram of a comprehensive inspection path intelligent planning system proposed in an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0058] The present application embodiment provides a comprehensive inspection path intelligent planning method, which is applied to automatic inspection equipment, such as Figure 1 As shown, the method comprises the following steps:

[0059] Step S101, obtain complete map information within the inspection area, establish a two-dimensional coordinate plane map, and collect interference information around the road based on the plane map, the interference information includes road congestion, road signal interference and building information.

[0060] In this embodiment, the interference information includes road congestion, road signal interference and building information, which may have an impact on the inspection. Therefore, these factors are considered as much as possible to improve the reliability of the inspection.

[0061] Step S102, determine the grid size according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, determine the center radius according to the road congestion, take the center line of the road on the plane map as the circle point, divide the area within the center radius according to the grid size, and fill in the interference information at the corresponding position, so as to obtain the grid interference information around the road.

[0062] In this embodiment, because the information involved in the plane map is large, the interference information cannot be marked on the map on a large scale, and the interference information in some locations is invalid for the inspection path. Here, only the interference information within a certain range around the road is gridded, and gridding can also improve the accuracy of information processing. The size of the grid is crucial for subsequent modeling and inspection.

[0063] In some embodiments of the present application, the grid size is determined according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, including:

[0064] The automatic inspection equipment structure information includes the inspection equipment area and the observable range of the inspection equipment. The inspection target information includes multiple target locations, areas and attributes. The road signal interference includes the signal interference coverage, signal interference intensity and signal fluctuation. The building information includes multiple building locations and areas.

[0065] Determine a comprehensive inspection target index through multiple target locations, areas and attributes;

[0066]

[0067] Among them, Q1 is the comprehensive index of the inspection target, n is the number of target locations, G1 i →G2 i is the weight of the automatic inspection device identifying the target mapped from the i-th target area to the i-th target area, m is the number of paths for the automatic inspection device corresponding to the i-th target to move to the target location, M j is the distance of the sum of the paths of the jth path, the target attributes include type and importance level, B is the importance obtained by weighted summation of type and importance level, exp is an exponential function, and k1 is a first preset constant;

[0068] A signal interference comprehensive index Q2 is determined according to the signal interference coverage, signal interference strength and signal fluctuation, and a building comprehensive index Q3 is determined according to the locations and areas of multiple buildings;

[0069] The grid size is determined by the inspection equipment area, the observable range of the inspection equipment, the inspection target comprehensive index Q1, the signal interference comprehensive index Q2, and the building comprehensive index Q3;

[0070]

[0071] Among them, L i K is the partial grid coefficient corresponding to the inspection target comprehensive index, signal interference comprehensive index or building comprehensive index, i is the constant corresponding to the inspection target comprehensive index, the signal interference comprehensive index or the building comprehensive index, s is the grid area, α1 is the partial weight corresponding to the inspection target comprehensive index, L1 is the partial grid coefficient of the inspection target comprehensive index, α2 is the partial weight corresponding to the signal interference comprehensive index, L2 is the partial grid coefficient of the signal interference comprehensive index, α3 is the partial weight corresponding to the building comprehensive index, L3 is the partial grid coefficient of the building comprehensive index, s1 is the area of ​​the observable range of the inspection equipment, and s2 is the inspection equipment area.

[0072] In this embodiment, the inspection target location and area may have a certain impact on the inspection of the automatic inspection equipment. Different locations or areas are easy to identify by the inspection equipment at different levels. The inspection target type and importance level will affect the identification. Correct the identification situation. Determine a signal interference comprehensive index Q2 based on the signal interference coverage, signal interference strength and signal fluctuation, and determine a building comprehensive index Q3 based on the locations and areas of multiple buildings. The process of obtaining the comprehensive index can be weighted summation or other comprehensive considerations to obtain the comprehensive index, which is not the focus of protection of this plan and will not be elaborated here.

[0073] In this embodiment, the size of the grid area is obtained by the ratio between the inspection equipment area and the observable range of the inspection equipment and the inspection target comprehensive index, the signal interference comprehensive index, and the building comprehensive index.

[0074] In some embodiments of the present application, the center radius is determined according to the road congestion, the center line of the road on the plane map is used as the circle point, the area within the center radius is divided according to the grid size, and the interference information at the corresponding position is filled, including:

[0075] A circle center radius is determined according to the average value of the road congestion. Different average values ​​of different road congestion correspond to different circle center radii. A circle is drawn according to the center line of the road on the plane map as the circle point and the circle center radius so that the circular area completely contains the road, thereby obtaining a circular area.

[0076] The circular area is meshed according to the grid size, and the grids that exceed the circular area are also planned into the circular area;

[0077] The coordinate position information of each grid area is determined, and the road congestion, road signal interference and building information are filled in accordingly through the coordinate position information.

[0078] In this embodiment, the radius of the circle center is determined by comprehensively considering the average congestion of the road in the past, and the size of the circular area corresponding to different roads is different. The grids that partially exceed the circular area are also planned into the circular area. After gridding, some grid areas may exceed the circular area. In order to ensure the integrity of the information, the exceeding part is included in the circular area.

[0079] Step S103, confirming the starting coordinates of the automatic inspection device, establishing a destination coordinate list according to the inspection target information, screening multiple inspection paths based on the starting coordinates, interference information and destination coordinate list, and determining the change nodes in the multiple inspection paths.

[0080] In this embodiment, the distance between the starting point and each target position is calculated, so as to select multiple target positions as the end points, and then select multiple inspection paths to determine the nodes that can be changed.

[0081] In some embodiments of the present application, a destination coordinate list is established according to the inspection target information, including:

[0082] Determine the coordinates of multiple target locations, and calculate the straight-line distance and path distance between the starting point coordinates of the automatic inspection device and the coordinates of the multiple target locations;

[0083] The weights of the straight-line distance and the path distance are assigned according to the importance of each target, and the measured distance is obtained by weighted summation;

[0084] Find the coordinates of the diagonal points of the starting point coordinates of the automatic inspection device on the plane map, calculate the straight-line distance, determine the first measurement distance threshold and the second measurement distance threshold according to the straight-line distance, and put multiple target positions whose measurement distance exceeds the first measurement distance threshold as end points into the end point coordinate list;

[0085] A plurality of target positions whose measured distances exceed the second measured distance threshold and do not exceed the first measured distance threshold are put into the endpoint coordinate list as candidate endpoints.

[0086] In this embodiment, the importance of different targets corresponds to two weights of straight-line distance and path distance. The coordinates of the diagonal points of the starting point coordinates of the automatic inspection device are found on the plane map, and the point on the map with the longest distance extending from the starting point is found according to the preset angle. The alternative end point is used when the end point of the current edge is used up and no suitable path is found. At this time, the alternative end point can be used.

[0087] In some embodiments of the present application, multiple inspection paths are screened based on the starting point coordinates, interference information, and the end point coordinate list, and the change nodes in the multiple inspection paths are determined, including:

[0088] Step 1: define nodes and edges on a plane map, where nodes are starting points, road intersections, and end points, and edges are moving paths from the starting point to the road intersection, moving paths between road intersections, and moving paths between the road intersection and the end point;

[0089] Step 2: Adjust the heuristic function according to the interference information and estimate the heuristic estimated cost h from the current node to the end point;

[0090] Step 3: Initialize the open list and the closed list, add the starting point to the open list, and set f = g + h, where g is the actual cost from the current node to the end point;

[0091] Step 4: Perform a cyclic search and update the values ​​of g, h and f;

[0092] Step 5: If the current node is the end point, determine a path; otherwise, add the current node to the closed list and repeat step 4.

[0093] Step 6: Repeat steps 3-5 to obtain multiple paths;

[0094] The paths that meet the preset f threshold are retained to obtain the preferred path, and the change nodes on the preferred path are determined. The change nodes are nodes that can be transferred from a certain node on one preferred path to another preferred path.

[0095] In this embodiment, the heuristic function is adjusted according to the interference information, and the heuristic estimation cost h from the current node to the end point is estimated, which means that the predicted cost generated by the interference information is determined by integrating the interference information, and this cost is added to the original heuristic estimation cost after the distance function is converted, so as to ensure the accuracy of the cost estimation.

[0096] In the present embodiment, step 4, perform cyclic search, select the node with minimum f value from the open list as the current node. Expand all neighbor nodes of the current node. For each neighbor node, calculate its g value, h value and f value. If the neighbor node is already in the closed list, or the path to the neighbor node through the current node is not optimal, then ignore the neighbor node. Otherwise, if the g value is less than the g value of the neighbor node, then update the parent node of the neighbor node to be the current node, and update its g value and f value.

[0097] In this embodiment, f is the total cost, and the heuristic function may be Manhattan distance or Euclidean distance.

[0098] Step S104: The automatic inspection device performs inspection along the inspection path with the lowest total cost, and establishes a neural network model for predicting the position of the automatic inspection device based on the grid interference information around the road and the previous position data of the automatic inspection device.

[0099] In this embodiment, a neural network model for predicting the location of the automatic inspection equipment is established based on the grid interference information around the road in the same period of time and the previous location data of the automatic inspection equipment. The grid interference information here is historical data. The process is as follows:

[0100] 1. Data Collection:

[0101] - Collect grid interference information around the road, including road congestion, road signal interference, inspection target information and building information.

[0102] -These data can come from sensors, cameras, GIS systems, etc.

[0103] 2. Data preprocessing:

[0104] - Clean, denoise and standardize the collected data.

[0105] -Convert the data into a form suitable for neural network input, such as scaling and normalizing image data.

[0106] 3. Feature extraction:

[0107] -Extract features related to the location of inspection equipment from the preprocessed data.

[0108] -These features may include an index of road congestion, the strength and fluctuation of signal interference, the location and attributes of the target, the size and location of buildings, etc.

[0109] 4. Model design:

[0110] -Choose an appropriate neural network architecture, such as a Convolutional Neural Network (CNN) or a Recurrent Neural Network (RNN).

[0111] -Design the number of layers, nodes, and activation functions of the neural network based on the dimensions and characteristics of the data.

[0112] 5. Training Model:

[0113] -The model is trained using a dataset containing actual device locations.

[0114] - Improve the predictive accuracy of the model by tuning network parameters and optimization algorithms such as Stochastic Gradient Descent (SGD) and Adam.

[0115] 6. Model Validation:

[0116] - Use a dataset that was not used in training to validate the model and evaluate its predictive performance.

[0117] - Check the model's performance indicators such as accuracy, recall, F1 score, etc.

[0118] The above process is completed, and some adjustments can be made to the model.

[0119] Step S105, multiple proofreading nodes are set in the inspection path, and the predicted value and actual value of the position of the automatic inspection equipment are compared through the proofreading nodes to optimize the neural network model and re-plan the inspection path.

[0120] In some embodiments of the present application, multiple proofreading nodes are set in the inspection path, including:

[0121] The plane map is divided into multiple areas, and the corresponding key areas are determined according to road congestion, road signal interference, inspection target information and building information. Several locations are randomly selected in multiple key areas, and the time nodes for the automatic inspection equipment to arrive at these locations are predicted and used as proofreading nodes.

[0122] In this embodiment, the plane map is divided into multiple areas, and the corresponding key areas are determined according to the road congestion, road signal interference, inspection target information and building information, that is, the road congestion is the largest or the road signal interference intensity is the largest or the inspection target information is the most complex or the building is the densest, to obtain several key areas, and the inspection situation is detected through the key areas. Several locations are randomly selected in multiple key areas, and the time nodes of the automatic inspection equipment to these locations are predicted and used as proofreading nodes. This is to facilitate subsequent position comparison, so the time node is selected as the proofreading node.

[0123] In some embodiments of the present application, the neural network model is optimized by comparing the predicted value and the actual value of the position of the automatic inspection equipment through the calibration node, and the inspection path is replanned, including:

[0124]

[0125] Among them, x t is the actual value of the position of the automatic inspection equipment under the t calibration node, is the predicted value of the automatic inspection equipment position under the t calibration node, W old is the weight of the neural network model before updating, X1 and X2 are the preset position error values, and W new is the updated neural network model weight, is the partial derivative of the loss function MSE with respect to the weight W, MSE is the loss function, exp is the exponential function, is the learning factor of the difference mapping;

[0126] Re-predict the subsequent time-position situation of the automatic inspection equipment under the inspection path according to the optimized neural network model, and determine whether the inspection path is still the optimal solution. If it is the optimal solution, continue the inspection of the inspection path;

[0127] Otherwise, the inspection path corresponding to the optimal solution is changed according to the changed nodes on the inspection path. In this embodiment, MSE is a loss function, which is essentially a mean square error. represents the learning rate corresponding to the difference, Indicates the update amount of the weight.

[0128] In this embodiment, when determining whether the inspection path is still the optimal solution, the cost, time, etc. of the inspection path are recalculated, and the cost, time, etc. of other paths that can be changed are calculated, so as to compare multiple paths. For example, the total cost and time remaining of the inspection path are calculated, and the total cost of other inspection paths that can be changed (which may include detours) is calculated. When the total cost of other inspection paths is lower than that of the inspection path, the optimal solution is replaced.

[0129] By applying the above technical solutions, the complete map information of the inspection area is obtained, a two-dimensional coordinate plane map is established, and interference information around the road is collected based on the plane map. The interference information includes road congestion, road signal interference, inspection target information and building information; the grid size is determined according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, and the center radius is determined according to the road congestion. The center line of the road on the plane map is used as the circle point, and the area within the center radius is divided according to the grid size, and the interference information at the corresponding position is filled in, so as to obtain the grid interference around the road. Information; confirm the starting coordinates of the automatic inspection equipment, establish a list of end coordinates according to the inspection target information, select multiple inspection paths based on the starting coordinates, interference information and end coordinate list, and determine the change nodes in the multiple inspection paths; the automatic inspection equipment inspects according to the inspection path with the lowest total cost, and establishes a neural network model for predicting the position of the automatic inspection equipment based on the grid interference information around the road and the previous position data of the automatic inspection equipment; multiple proofreading nodes are set in the inspection path, and the predicted value and actual value of the automatic inspection equipment position are compared by proofreading nodes to optimize the neural network model and re-plan the inspection path. This application determines the appropriate grid size based on the inspection equipment situation and road interference situation, thereby effectively gridding interference information, improving data accuracy, and avoiding redundant data. Multiple end points and alternative end points are selected based on the straight-line distance and path distance in the inspection target location information, ensuring the reliability of the end point selection. The predicted value and actual value of the automatic inspection equipment position are compared by proofreading nodes to optimize the neural network model and re-plan the inspection path. It realizes intelligent path planning. When the path taken is not suitable, another suitable path can be selected, which improves the adaptability and accuracy of path planning and ensures the safe and reliable operation of automatic inspection equipment.

[0130] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present invention can be implemented by hardware, or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present invention.

[0131] In order to further explain the technical idea of ​​the present invention, the technical solution of the present invention is now described in combination with specific application scenarios.

[0132] Correspondingly, the present application also provides a comprehensive inspection path intelligent planning system, which is applied to automatic inspection equipment, including:

[0133] The collection module is used to obtain complete map information within the inspection area, establish a two-dimensional coordinate plane map, and collect interference information around the road based on the plane map. The interference information includes road congestion, road signal interference and building information;

[0134] The grid module is used to determine the grid size according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, determine the center radius according to the road congestion, take the center line of the road on the plane map as the circle point, divide the area within the center radius according to the grid size, and fill in the interference information at the corresponding position, so as to obtain the grid interference information around the road;

[0135] The optimization module is used to confirm the starting point coordinates of the automatic inspection equipment, establish a list of end point coordinates according to the inspection target information, screen multiple inspection paths based on the starting point coordinates, interference information and the end point coordinate list, and determine the change nodes in the multiple inspection paths;

[0136] A construction module is used for the automatic inspection equipment to inspect along the inspection path with the lowest total cost, and a neural network model is established based on the grid interference information around the road and the previous location data of the automatic inspection equipment to predict the location of the automatic inspection equipment;

[0137] The optimization module is used to set multiple proofreading nodes within the inspection path, compare the predicted value and actual value of the automatic inspection equipment position through the proofreading nodes, optimize the neural network model, and re-plan the inspection path.

[0138] Those skilled in the art will appreciate that the modules in the device in the implementation scenario can be distributed in the device in the implementation scenario according to the implementation scenario description, or can be changed accordingly and located in one or more devices different from the implementation scenario. The modules in the above implementation scenario can be combined into one module, or can be further split into multiple sub-modules.

[0139] A computer-readable storage medium stores one or more programs, which, when executed by an electronic device including a plurality of application programs, cause the electronic device to execute the above method.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A comprehensive inspection path intelligent planning method, applied to automatic inspection equipment, characterized in that: include: Obtain complete map information within the inspection area, establish a two-dimensional coordinate plane map, and collect interference information around the road based on the plane map. The interference information includes road congestion, road signal interference, and building information; The grid size is determined according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, and the center radius is determined according to the road congestion. The center line of the road on the plane map is used as the circle point, and the area within the center radius is divided according to the grid size, and the interference information at the corresponding position is filled in, so as to obtain the grid interference information around the road; Confirm the starting point coordinates of the automatic inspection equipment, establish a destination coordinate list according to the inspection target information, filter multiple inspection paths based on the starting point coordinates, interference information and destination coordinate list, and determine the change nodes in the multiple inspection paths; The automatic inspection equipment conducts inspections along the inspection path with the lowest total cost. A neural network model is established based on the grid interference information around the road and the previous location data of the automatic inspection equipment to predict the location of the automatic inspection equipment. Set up multiple proofreading nodes in the inspection path, compare the predicted value and actual value of the automatic inspection equipment position through the proofreading nodes to optimize the neural network model and re-plan the inspection path; in, The grid size is determined based on the automatic inspection equipment structure information, inspection target information, road signal interference and building information, including: The automatic inspection equipment structure information includes the inspection equipment area and the observable range of the inspection equipment. The inspection target information includes multiple target locations, areas and attributes. The road signal interference includes the signal interference coverage, signal interference intensity and signal fluctuation. The building information includes multiple building locations and areas. Determine a comprehensive inspection target index through multiple target locations, areas and attributes; ； in, is the comprehensive index of the inspection target, n is the number of target locations, is the weight of the automatic inspection device identifying the target mapped from the i-th target area to the i-th target area, m is the number of paths for the automatic inspection device corresponding to the i-th target to move to the target location, is the sum of the paths of the jth path, the target attributes include type and importance level, B is the importance obtained by weighted summation of type and importance level, exp is an exponential function, is the first preset constant; Determine a comprehensive signal interference index based on signal interference coverage, signal interference strength and signal fluctuation , determine a building comprehensive index based on multiple building locations and areas ; Through the inspection equipment area, inspection equipment observable range, inspection target comprehensive index , signal interference comprehensive index , Building Comprehensive Index To determine the grid size; ； in, is the partial grid coefficient corresponding to the inspection target comprehensive index, signal interference comprehensive index or building comprehensive index, is the constant corresponding to the inspection target comprehensive index or signal interference comprehensive index or building comprehensive index, s is the grid area, is the partial weight corresponding to the comprehensive index of the inspection target, is the partial grid coefficient of the comprehensive index of the inspection target, is the partial weight corresponding to the comprehensive index of signal interference, is the partial grid coefficient of the signal interference comprehensive index, is the partial weight corresponding to the comprehensive index of buildings, is the partial grid coefficient of the building comprehensive index, is the area within the observable range of the inspection equipment. The area for inspection equipment.

2. A comprehensive inspection path intelligent planning method as claimed in claim 1, characterized in that: The radius of the circle is determined according to the road congestion. The center line of the road on the plane map is used as the circle point. The area within the radius of the circle is divided according to the grid size and the interference information at the corresponding position is filled in, including: A circle center radius is determined according to the average value of the road congestion. Different average values ​​of different road congestion correspond to different circle center radii. A circle is drawn according to the center line of the road on the plane map as the circle point and the circle center radius so that the circular area completely contains the road, thereby obtaining a circular area. The circular area is meshed according to the grid size, and the grids that exceed the circular area are also planned into the circular area; The coordinate position information of each grid area is determined, and the road congestion, road signal interference and building information are filled in accordingly through the coordinate position information.

3. A comprehensive inspection path intelligent planning method as claimed in claim 1, characterized in that: Create a list of endpoint coordinates based on inspection target information, including: Determine the coordinates of multiple target locations, and calculate the straight-line distance and path distance between the starting point coordinates of the automatic inspection device and the coordinates of the multiple target locations; The weights of the straight-line distance and the path distance are assigned according to the importance of each target, and the measured distance is obtained by weighted summation; Find the coordinates of the diagonal points of the starting point coordinates of the automatic inspection device on the plane map, calculate the straight-line distance, determine the first measurement distance threshold and the second measurement distance threshold according to the straight-line distance, and put multiple target positions whose measurement distance exceeds the first measurement distance threshold as end points into the end point coordinate list; A plurality of target positions whose measured distances exceed the second measured distance threshold and do not exceed the first measured distance threshold are put into the endpoint coordinate list as candidate endpoints.

4. A comprehensive inspection path intelligent planning method as claimed in claim 1, characterized in that: Based on the starting point coordinates, interference information and the end point coordinate list, multiple inspection paths are screened, and the changed nodes in the multiple inspection paths are determined, including: Step 1: define nodes and edges on a plane map, where nodes are starting points, road intersections, and end points, and edges are moving paths from the starting point to the road intersection, moving paths between road intersections, and moving paths between the road intersection and the end point; Step 2: adjust the heuristic function according to the interference information and estimate the heuristic estimated cost h from the current node to the end point; Step 3: Initialize the open list and closed list, add the starting point to the open list, and set f = g + h, where g is the actual cost from the current node to the end point; Step 4: Perform a cyclic search and update the g, h and f values; Step 5: If the current node is the end point, determine a path; otherwise, add the current node to the closed list and repeat step 4. Step 6: Repeat steps 3-5 to obtain multiple paths; The paths that meet the preset f threshold are retained to obtain the preferred path, and the change nodes on the preferred path are determined. The change nodes are nodes that can be transferred from a certain node on one preferred path to another preferred path.

5. The method for intelligent planning of a comprehensive inspection path according to claim 1, characterized in that: Set up multiple proofreading nodes within the inspection path, including: The plane map is divided into multiple areas, and the corresponding key areas are determined according to road congestion, road signal interference, inspection target information and building information. Several locations are randomly selected in multiple key areas, and the time nodes for the automatic inspection equipment to arrive at these locations are predicted and used as proofreading nodes.

6. A comprehensive inspection path intelligent planning method as claimed in claim 5, characterized in that: By comparing the predicted and actual values ​​of the automatic inspection equipment position through the calibration node, the neural network model is optimized and the inspection path is replanned, including: ； in, is the actual value of the position of the automatic inspection equipment under the t calibration node, is the predicted value of the automatic inspection equipment position under the t calibration node, is the weight of the neural network model before updating, is the preset position error value, is the updated neural network model weight, is the partial derivative of the loss function MSE with respect to the weight W, MSE is the loss function, exp is the exponential function, is the learning factor of the difference mapping; Re-predict the subsequent time-position situation of the automatic inspection equipment under the inspection path according to the optimized neural network model, and determine whether the inspection path is still the optimal solution. If it is the optimal solution, continue the inspection of the inspection path; Otherwise, the inspection path corresponding to the optimal solution is changed according to the changed nodes on the inspection path.

7. A comprehensive inspection path intelligent planning system, applied to automatic inspection equipment, for implementing the comprehensive inspection path intelligent planning method according to any one of claims 1 to 6, characterized in that: include: The collection module is used to obtain the complete map information of the inspection area, establish a two-dimensional coordinate plane map, and collect interference information around the road based on the plane map. The interference information includes road congestion, road signal interference, and building information; The grid module is used to determine the grid size according to the automatic inspection equipment structure information, inspection target information, road signal interference and building information, determine the center radius according to the road congestion, take the center line of the road on the plane map as the circle point, divide the area within the center radius according to the grid size, and fill in the interference information at the corresponding position, so as to obtain the grid interference information around the road; The optimization module is used to confirm the starting point coordinates of the automatic inspection equipment, establish a list of end point coordinates according to the inspection target information, screen multiple inspection paths based on the starting point coordinates, interference information and the end point coordinate list, and determine the change nodes in the multiple inspection paths; A construction module is used for the automatic inspection equipment to inspect along the inspection path with the lowest total cost, and a neural network model is established based on the grid interference information around the road and the previous location data of the automatic inspection equipment to predict the location of the automatic inspection equipment; The optimization module is used to set multiple proofreading nodes in the inspection path, compare the predicted value and actual value of the automatic inspection equipment position through the proofreading nodes, optimize the neural network model, and re-plan the inspection path.

8. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, enables the electronic device to execute any one of the methods of claims 1 to 6.

Images