Building facade three-dimensional surveying and mapping system and method based on unmanned aerial vehicle and AI

Abstract

The invention discloses a building facade three-dimensional surveying and mapping system and method based on an unmanned aerial vehicle and AI, and relates to the technical field of building surveying and mapping, and the system comprises a scene analysis module which is used for executing the information analysis of calibration scene information, and marking each surveying and mapping sub-region as a to-be-collected node; the acquisition planning module is used for calculating the node transition probability of each node to be acquired; the surveying and mapping search module is used for configuring a virtual unmanned aerial vehicle and performing selective iteration of surveying and mapping sub-regions by using pheromone concentration and heuristic factors; the clustering sharing module is used for carrying out three-dimensional surveying and mapping on the exterior facade of the building and establishing a shared information pool in a clustering manner; and the surveying and mapping optimization module is used for carrying out collaborative decision optimization on the local path and updating the acquisition path according to a collaborative decision optimization result. The technical problems of low surveying and mapping efficiency and large measurement error caused by difficulty in meeting large-scale and high-precision surveying and mapping requirements of modern buildings in the prior art are solved, and the technical effect of improving the surveying and mapping efficiency and precision of the buildings is achieved.

Description

Technical Field

[0001] This application relates to the technical field of building surveying and mapping, and specifically relates to a three-dimensional building facade surveying and mapping system and method based on drones and AI. Background Art

[0002] With the acceleration of the urbanization process, the construction industry has developed vigorously, and the demand for accurate surveying and mapping of building facades has been increasing. Traditional building facade surveying and mapping methods, such as manually using simple surveying instruments for on-site surveying, for example, professional personnel need to carry equipment and climb to various parts of the building facade for measurement. This is not only inefficient, time-consuming and laborious, but also has large measurement errors, is limited by human physical strength and operation accuracy, and is significantly affected by environmental factors. For example, in some high-rise or complex-structured buildings, it is difficult for humans to reach certain positions, resulting in data loss. In addition, conventional instrument surveying, such as total stations, although can improve a certain accuracy, but need to set up stations at multiple positions, the process is cumbersome, has high requirements for the surrounding environment, and is difficult to be effectively carried out in narrow spaces or where there are many surrounding obstacles. Moreover, the data obtained is difficult to process, and it is impossible to quickly generate an intuitive three-dimensional model, which is difficult to meet the requirements of large-scale and high-precision surveying and mapping of modern buildings.

[0003] Therefore, in the current related technologies, there are technical problems such as being difficult to meet the requirements of large-scale and high-precision surveying and mapping of modern buildings, resulting in low surveying and mapping efficiency and large measurement errors. Summary of the Invention

[0004] This application provides a three-dimensional building facade surveying and mapping system and method based on drones and AI, solves the technical problems in the prior art that it is difficult to meet the requirements of large-scale and high-precision surveying and mapping of modern buildings, resulting in low surveying and mapping efficiency and large measurement errors, and achieves the technical effect of improving the efficiency and accuracy of building surveying and mapping.

[0005] This application provides a 3D mapping system for building facades based on drones and AI. The system includes: a scene analysis module for reading the calibrated scene information of the scene to be mapped, performing information analysis on the calibrated scene information, establishing N mapping sub-regions, and marking each mapping sub-region as a node to be collected; a collection planning module for distributing the initial pheromone values on each mapping path and configuring a heuristic factor after determining the collection starting points of the distributed drone swarm, and calculating the node transition probability of each node to be collected; a mapping search module for configuring a virtual drone to start from the collection starting point, calculating the pheromone concentration based on the initial pheromone value, and performing iterative selection of mapping sub-regions using the pheromone concentration and the heuristic factor to establish an optimization result. Among them, after each global path planning is completed, the initial pheromone value is accumulated and updated; a clustering sharing module for controlling the drone swarm to perform 3D mapping of the building facade based on the optimization result, clustering and establishing a shared information pool according to the real-time mapping positions of the drone swarm, and uploading the real-time mapping results to the shared information pool; a mapping optimization module for performing collaborative decision optimization of local paths through the shared information pool and the optimization result, and updating the collection path according to the collaborative decision optimization result.

[0006] In a possible implementation, the 3D mapping system for building facades based on drones and AI further performs the following processing: a drawing reading module for reading the scene drawing of the scene to be mapped and using the scene drawing as the calibrated scene information; a grid segmentation module for establishing a median grid granularity, performing grid division on the building facade of the scene to be mapped based on the median grid granularity, and generating an initial grid segmentation result; a grid update module for identifying the integrity of elements within the grid for the initial grid segmentation result, performing grid migration clustering according to the integrity identification result, generating a grid migration clustering result, and establishing N mapping sub-regions based on the grid migration clustering result.

[0007] In a possible implementation, the 3D mapping system for building facades based on drones and AI further performs the following processing: obtaining a set of elements not covered by the initial grid segmentation result and marking it as an updated element set; obtaining the size data of each element in the updated element set, and selecting clustering elements according to the matching degree between the size data and the median grid granularity; using the clustering elements to complete grid migration clustering.

[0008] In a possible implementation, the 3D mapping system for building facades based on drones and AI further performs the following processing: establishing a collection node transfer cost function as follows: ; where represents the transfer from collection node to collection node The acquisition node transfer cost function characterizes the acquisition node and the acquisition node distance and respectively characterize the risk factors of the acquisition node and the acquisition node ; and respectively characterize the node priorities of the acquisition node and the acquisition node ; , , are the distance weight factor, risk weight factor, and node priority weight factor respectively; configure the heuristic factor according to the acquisition node transfer cost function, and calculate the node transfer probability using the heuristic factor as follows: ; wherein characterizes the node transfer probability from the acquisition node to the acquisition node ; characterizes the pheromone concentration from the acquisition node to the acquisition node ; is the heuristic factor, specifically representing the heuristic information from the acquisition node to the acquisition node ; , characterizes the set of optional acquisition nodes not visited by the UAV characterizes any optional acquisition node characterizes the pheromone concentration from the acquisition node to the acquisition node ; characterizes the heuristic information from the acquisition node to the acquisition node ; characterizes the pheromone influence factor characterizes the heuristic factor influence factor

[0009] In a possible implementation, the drone and AI-based three-dimensional surveying and mapping system for building facades further performs the following processing: obtaining the spatial position coordinates of each drone in the drone swarm at the current time node; obtaining the next surveying sub-region of each drone in the drone swarm; performing real-time surveying position clustering based on the next surveying sub-region and the spatial position coordinates to generate a real-time surveying position clustering result; identifying the drone with the largest storage space in each cluster set in the real-time surveying position clustering result, creating a shared information pool in the identified drone, and receiving the real-time surveying results of other drones within the real-time surveying position clustering result.

[0010] In a possible implementation, the drone and AI-based three-dimensional surveying and mapping system for building facades further performs the following processing: obtaining the environmental shared information in the shared information pool, reconstructing the local environmental mutation constraints based on the environmental shared information; obtaining surveying anomalies based on the real-time surveying results in the shared information pool, and creating additional surveying tasks using the surveying anomalies; performing collaborative decision optimization of the local path based on the local environmental mutation constraints, the additional surveying tasks, and the optimization result.

[0011] In a possible implementation, the drone and AI-based three-dimensional surveying and mapping system for building facades further performs the following processing: an anomaly identification module for recording the acquisition anomalies of the drone swarm, performing anomaly level discrimination, matching the response plans in the response plan library according to the anomaly level discrimination result, and performing anomaly management according to the response plan matching result. The response plan library includes an anomaly marking response plan and a temporary task creation response plan.

[0012] This application also provides a drone and AI-based three-dimensional surveying and mapping method for building facades, including: reading the calibrated scene information of the scene to be surveyed, performing information parsing of the calibrated scene information, establishing N surveying sub-regions, and marking each surveying sub-region as a node to be collected; after the starting points of the distributed drone swarm, assigning initial pheromone values and configuring heuristic factors on each surveying path, and calculating the node transition probability of each node to be collected; configuring a virtual drone to start from the starting point of collection, calculating the pheromone concentration based on the initial pheromone value, and performing iterative selection of the surveying sub-region using the pheromone concentration and the heuristic factor to establish an optimization result. Among them, after each global path planning is completed, the initial pheromone value is accumulated and updated; controlling the drone swarm to perform three-dimensional surveying of the building facade based on the optimization result, establishing a shared information pool according to the real-time surveying position clustering of the drone swarm, and uploading the real-time surveying results to the shared information pool; performing collaborative decision optimization of the local path through the shared information pool and the optimization result, and updating the acquisition path according to the collaborative decision optimization result.

[0013] The three-dimensional mapping system and method for building facades based on drones and AI proposed in this application, a scene analysis module, is used to perform information analysis for calibrating scene information and mark each mapping sub-region as a node to be collected; a collection planning module is used to calculate the node transfer probability of each node to be collected; a mapping search module is used to configure a virtual drone and perform iterative selection of mapping sub-regions using pheromone concentration and heuristic factors; a clustering sharing module is used to perform three-dimensional mapping of the building facade and cluster to establish a shared information pool; a mapping optimization module is used to perform collaborative decision optimization of local paths and update the collection path according to the collaborative decision optimization results. It solves the technical problems in the prior art that it is difficult to meet the requirements of large-scale and high-precision mapping of modern buildings, resulting in low mapping efficiency and large measurement errors, and achieves the technical effect of improving the efficiency and accuracy of building mapping. Brief Description of the Drawings

[0014] To more clearly illustrate the technical solutions of the embodiments of the present disclosure, the accompanying drawings of the embodiments of the present disclosure will be briefly introduced below. Flowcharts are used in this application to illustrate the operations performed by the systems according to the embodiments of the application. It should be understood that the operations described above or below do not necessarily need to be executed precisely in order. On the contrary, as needed, various steps can be executed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several operations can be removed from these processes.

[0015] Figure 1 It is a schematic structural diagram of a three-dimensional mapping system for building facades based on drones and AI provided by an embodiment of the present application.

[0016] Figure 2 It is a schematic flowchart of a three-dimensional mapping method for building facades based on drones and AI provided by an embodiment of the present application.

[0017] Description of the reference numerals: scene analysis module 10, collection planning module 20, mapping search module 30, clustering sharing module 40, mapping optimization module 50. Detailed Embodiments

[0018] The above description is only an overview of the technical solutions of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically illustrates the specific embodiments of this application.

[0019] In order to make the purpose, technical solutions and advantages of this application clearer, the following will further describe this application in detail with reference to the accompanying drawings. The described embodiments should not be regarded as limitations of this application. All other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application.

[0020] In the following description, "some embodiments" are involved, which describe a subset of all possible embodiments. However, it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The terms "first / second" involved are only used to distinguish similar objects and do not represent a specific order for the objects. The terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, system, product or server comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules not clearly listed or inherent to these processes, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application.

[0021] The embodiments of this application provide a three-dimensional mapping system for building facades based on drones and AI, as Figure 1 shown. The system includes: A scene analysis module 10, configured to read the calibrated scene information of the scene to be mapped, perform information analysis on the calibrated scene information, establish N mapping sub-regions, and mark each mapping sub-region as a node to be collected.

[0022] Preferably, reading the calibrated scene information of the scene to be mapped may include obtaining detailed information such as the basic dimensions, shape, structural layout, floor height, and door and window positions of the building from the design drawings of the building; then using GIS data to obtain the geographical location information of the building, including longitude, latitude, altitude, etc., as well as environmental information such as the surrounding terrain and features, to determine the position of the mapping area in the entire geographical space and consider the impact of the surrounding environment on the mapping work, such as the occlusion of surrounding buildings and the restrictions on the flight airspace; obtaining key dimension data and feature point information through on-site measurement, such as the actual height of the building, the inclination of the facade, and the dimensions of special structures; and at the same time using sensor devices such as laser rangefinders and total stations to obtain the distance information and spatial coordinate data of the building facade.

[0023] Preferably, information parsing is to parse and convert the read original calibration scene information into actionable data. Specifically, various read data are subjected to format conversion and normalization processing so that they can be processed by unified algorithms and models. For example, coordinate data from different sources are converted into a unified coordinate system, image data are adjusted to standard resolutions and formats, noise and outliers in the data are removed, and laser ranging data, sensor measurement data, etc. are processed through algorithms such as filtering and smoothing to improve the quality and reliability of the data; computer vision algorithms and machine learning models are used to extract feature information of the building from the image data, such as edges, corners, textures, etc. For example, the Canny edge detection algorithm is used to extract the contour edges of the building facade, and the Harris corner detection algorithm is used to identify key corners in the building structure. For laser ranging data and spatial coordinate data, geometric features of the building are extracted through methods such as clustering analysis and surface fitting, such as the parameters and position information of structures such as planes, curved surfaces, and cylinders. For example, the DBSCAN clustering algorithm is used to cluster laser point cloud data into different structural units, and the least squares method is used to fit the plane equation to determine the plane area of the building facade; then, combined with knowledge and semantic models in the building field, semantic understanding and annotation of the extracted features are performed. For example, by training deep learning models such as convolutional neural networks (CNNs) or semantic segmentation models, different building components such as windows, doors, balconies, columns, etc. are identified and their types and position information are annotated to achieve semantic segmentation and recognition of building images and point cloud data.

[0024] Preferably, N mapping sub-regions are established, where N is a positive integer representing the total number of mapping sub-regions. Specifically, according to the structural characteristics of the building and the requirements of mapping accuracy, a division rule is formulated. For example, the building facade is divided into several horizontal sub-regions according to the floor height. The height of each sub-region can be determined according to the flight performance of the drone and the mapping accuracy, generally about 5-10 meters. Or, according to the facade shape and structural unit of the building, it is divided into several rectangular or polygonal sub-regions, and each sub-region preferably contains complete building components or structural features. Or, an image segmentation algorithm or a spatial clustering algorithm is used to automatically divide the feature data of the building. For example, using a region-growing-based image segmentation algorithm, starting from the seed points of the building image, different sub-regions are gradually grown according to the similarity of pixels. Or, the K-means clustering algorithm is used to cluster the laser point cloud data, and the point cloud data is divided into N different sub-regions. Each sub-region corresponds to a mapping sub-region, and the initially divided sub-regions are optimized and adjusted to ensure that the boundaries between the sub-regions are clear and seamlessly connected, and the size and shape of each sub-region are relatively uniform, avoiding the appearance of overly large or overly small sub-regions that affect the mapping efficiency and accuracy. At the same time, considering the flight path and acquisition angle of the drone, some special-shaped or difficult-to-acquire sub-regions are appropriately merged or split. And a unique identifier is assigned to each mapping sub-region, marking it as a point to be acquired. The annotation content may also include relevant attribute information of the mapping sub-region, such as the position coordinates, size, shape, and the type of building structure to which it belongs, which helps the drone to accurately locate and identify each sub-region during the acquisition process. For example, record the coordinates of the upper left and lower right corners of the sub-region, as well as the corresponding floor and building facade information.

[0025] Furthermore, the specific configuration of the scene analysis module 10 also includes a drawing reading module for reading the scene drawing of the scene to be mapped and using the scene drawing as the calibrated scene information; a grid segmentation module for establishing a median grid granularity and performing grid division on the building facade of the scene to be mapped based on the median grid granularity to generate an initial grid segmentation result; and a grid update module for identifying the integrity of the elements within the grid for the initial grid segmentation result, performing grid migration clustering according to the integrity identification result to generate a grid migration clustering result, and establishing N mapping sub-regions according to the grid migration clustering result.

[0026] Preferably, the main function of the drawing reading module is to read the scene drawings of the scene to be surveyed and mapped, and regard these drawings as calibrated scene information. Among them, the scene drawings contain various detailed information of the building to be surveyed and mapped. Specifically, the drawings of the scene to be surveyed and mapped are obtained from different data sources, such as the electronic documents provided by the construction unit, the scanned version of the drawings in the archive, etc. The drawing formats are uniformly processed. For example, for CAD format drawings, professional CAD libraries may be needed for reading and conversion. For image format drawings, image preprocessing may be required, such as adjusting the resolution and grayscale. The information related to surveying and mapping, such as the external dimensions, structural layout, door and window positions, floor height, etc. of the building, is extracted from the drawings and used as the calibrated scene information.

[0027] Preferably, the core task of the grid segmentation module is to establish the median grid granularity, and based on this granularity, the building facade of the scene to be surveyed and mapped is divided into grids, and finally the initial grid segmentation result is generated. Specifically, considering multiple factors, such as the surveying and mapping accuracy requirements, the complexity of the building, the flight performance of the drone, etc., the median grid granularity is determined. Among them, the median grid granularity is the key parameter that determines the grid size. Generally speaking, for scenes with high accuracy requirements or complex building structures, the median grid granularity should be set smaller, otherwise it can be appropriately increased. Then, according to the determined median grid granularity, the building facade is divided into several grids. For example, regular grid division methods, such as square grids or rectangular grids, can be used, or irregular grid division methods can be used according to the shape and structural characteristics of the building. During the division process, it is ensured that the grids can cover the entire building facade and there are no overlaps or omissions between the grids. Finally, the information of the divided grids is sorted and recorded to form the initial grid segmentation result, which usually contains information such as the position, size, and number of each grid.

[0028] Preferably, the grid update module mainly identifies the integrity of elements within the grid for the initial grid segmentation result, performs grid migration clustering based on the identification result, finally generates the grid migration clustering result, and establishes N mapping sub-regions accordingly. Specifically, the integrity of elements is judged by analyzing the information in the drawing or combining image recognition technology, that is, each grid in the initial grid segmentation result is checked to identify whether the building elements (such as doors and windows, walls, decorative components, etc.) contained in the grid are complete. For example, if a grid only contains some doors and windows, the elements in that grid are incomplete; according to the identification result of element integrity, migration and clustering operations are performed on the grid. For grids with incomplete elements, they are migrated to adjacent grids containing complete elements, or multiple adjacent grids with incomplete elements are merged into a new grid, so as to make the elements in each grid as complete as possible, improve the accuracy and efficiency of mapping, and organize and record the grid information after migration and clustering operations to form the grid migration clustering result, reflecting the optimized grid division situation, and the elements in each grid are more complete and reasonable; finally, according to the grid migration clustering result, the building facade is divided into N mapping sub-regions. For example, the sub-regions are divided according to factors such as the position of the grid, element type, and mapping difficulty. Each mapping sub-region should have relatively independent mapping tasks and goals, so that the unmanned aerial vehicle can collect data more efficiently.

[0029] Furthermore, the specific configuration of the scene analysis module 10 further includes obtaining a set of elements not covered by the initial grid segmentation result, marked as the updated element set; obtaining the size data of each element in the updated element set, and selecting clustering elements according to the matching degree between the size data and the median grid granularity; using the clustering elements to complete grid migration clustering.

[0030] Preferably, when performing grid division on the building facade, due to the complexity of the building structure, data errors, or limitations of the division rules, the initial grid segmentation result may not fully cover all building elements (such as doors and windows, decorative components, special texture areas, etc. on the building facade). By comparing the initial grid segmentation result with the original scene drawing, identify the elements that are not included in any of the initial grids, mark them as the updated element set, which includes the elements missed during the grid division process; for each element in the updated element set, obtain its detailed dimension data, which may include extracting from the original design drawing or obtaining information such as the length, width, and area of the element through image measurement, 3D modeling, etc.; then compare the dimension data of each element with the median grid granularity to evaluate their matching degree. For example, if the area of an element is similar to the area of the grid corresponding to the median grid granularity, or the length and width dimensions of the element can be well integrated into a grid based on the median grid granularity, then it is considered that the matching degree of this element with the median grid granularity is high; according to the result of the matching degree evaluation, select those elements with a high matching degree from the updated element set as clustering elements, thereby improving the rationality and integrity of the grid division.

[0031] Preferably, considering the position of the grid, element type, integrity of the included elements, etc., in the initial grid segmentation result, find a suitable migration target grid for each clustering element. For example, if a clustering element is a window, grids that contain part of the window element or are adjacent to the position of the window can be preferentially selected as the migration target, and migrate the clustering element to the selected target grid, which may include updating the attribute information of the target grid, such as area, included element type and quantity, etc., while ensuring that the migrated grid meets the integrity of the element and the connectivity of the grid; after completing the migration of the clustering elements, there may be some adjacent grids with similar elements, then perform a clustering and merging operation on these grids. Through the grid migration and clustering method, further optimize the grid division result, make the elements in each grid more complete and unified, ensure that all elements of the building facade can be reasonably included in the division of the surveying sub-region, thereby improving the accuracy and efficiency of the surveying.

[0032] Preferably, the grid migration clustering of the initial grid segmentation result specifically includes: First, clarify the attribute information of each grid in the initial grid segmentation result, including the position (coordinates), size, type of building elements contained, and related features (such as the number of doors and windows, area ratio, etc.) of the grid, and establish a two-dimensional array to represent the grid set, with each element corresponding to a grid; then set an integrity threshold to evaluate the integrity of each grid. If the area ratio of key elements such as doors and windows in the grid is lower than this threshold, the grid elements are considered incomplete; or judge according to whether the grid contains a complete building structure unit (such as the wall area corresponding to a complete room); then determine its adjacent grids for each grid by defining neighborhood rules. For example, for rectangular grids on a two-dimensional plane, define the grids adjacent to the current grid in the horizontal, vertical, or diagonal directions as its adjacent grids, and then establish an adjacency list or adjacency matrix to represent the grid adjacency relationship for quickly finding the adjacent grids of each grid.

[0033] Preferably, then perform grid migration clustering, that is, traverse all grids to judge their integrity. If the current grid elements are complete, skip this grid and continue to process the next grid; if the current grid elements are incomplete, according to the adjacent grid relationship and the element integrity evaluation result, find the grid with complete elements among the adjacent grids of the current grid. If multiple adjacent grids with complete elements are found, select the grid with the shortest distance (calculated based on the distance between the grid centers) or the most similar element features (such as similar door and window types, wall materials, etc.) as the target grid; then merge the attribute information (such as position, partial elements contained, etc.) of the current incomplete element grid into the target grid, and then delete the current incomplete element grid from the grid set, and at the same time update the adjacent grid relationship data structure to ensure that the adjacent relationship between other grids and the deleted grid is correctly processed (for example, update the adjacency list or adjacency matrix). Iterate until all grids are traversed. After the grid migration is completed, there may be some adjacent grids with complete elements but high similarity, so cluster and merge these grids. For example, by calculating the similarity between adjacent grids (comprehensively evaluating based on factors such as element type, position relationship, area ratio, etc.), merge the grids with similarity higher than a certain threshold into a larger grid, and repeat the clustering optimization until there are no more grids to be merged, and finally obtain the grid migration clustering result.

[0034] The acquisition planning module 20 is used to allocate initial pheromone values on each mapping path and configure the heuristic factor after the acquisition starting points of the distributed unmanned aerial vehicle group, and calculate the node transfer probability of each node to be acquired.

[0035] Furthermore, the specific configuration of the acquisition planning module 20 also includes establishing an acquisition node transfer cost function as follows: ; Among them, represents the acquisition node transfer cost function from transferring to the acquisition node , represents the distance between the acquisition node and the acquisition node , and respectively represent the risk factors of the acquisition node and the acquisition node , and respectively represent the node priorities of the acquisition node and the acquisition node ; , , are the distance weight factor, risk weight factor, and node priority weight factor respectively; configure the heuristic factor according to the acquisition node transfer cost function, and calculate the node transfer probability using the heuristic factor as follows: ; Among them, represents the node transfer probability from the acquisition node to the acquisition node , represents the pheromone concentration from the acquisition node to the acquisition node , is the heuristic factor, specifically representing the heuristic information from the acquisition node to the acquisition node , , represents the set of optional acquisition nodes not visited by the UAV, represents any optional acquisition node, represents the pheromone concentration from the acquisition node to the acquisition node , represents the heuristic information from the acquisition node to the acquisition node , represents the pheromone influence factor, represents the heuristic factor influence factor.

[0036] Preferably, in the three-dimensional mapping scenario of the building facade, considering various factors such as the location, shape, surrounding environment of the building, and the overall mapping requirements, etc., determine the location where the drone swarm starts data collection, that is, the collection starting point. The collection starting point can be determined according to the geometric center of the building, the main entrance, or other representative locations. For example, for a building with a regular shape, the starting point can be set in front of or above the building. For a building with a complex shape, multiple starting points may need to be distributed according to multiple key feature points to ensure that the drones can comprehensively and efficiently map the building facade; assign the same initial pheromone value to all mapping paths, usually a small constant, to give all paths the opportunity to be selected and avoid some paths being overly preferred or ignored at the beginning. Among them, pheromone is an important concept in the ant colony algorithm. In the path planning of drone mapping, the pheromone value represents the attractiveness or quality of a certain mapping path, and the initial pheromone value is the initial value assigned to each possible mapping path at the beginning of the algorithm.

[0037] Preferably, the heuristic factor is a parameter used to guide the drone to select a path, reflecting the visibility or intuitive attractiveness of the path. Specifically, determine the heuristic factor according to the specific mapping scenario and requirements (such as the distance, angle, obstacle situation of the path, etc.). For example, a shorter path and a path with a better perpendicular angle to the building facade may have a higher heuristic factor; then calculate the node transition probability of each node to be collected. Among them, the node transition probability refers to the likelihood that the drone transfers from the current node to the next node to be collected. According to the principle of the ant colony algorithm, the node transition probability is calculated by the formula and is calculated by comprehensively considering the pheromone value and heuristic factor on the path to simulate the behavior of ants choosing paths according to factors such as pheromone and distance when looking for food. By calculating the transition probability of each node to be collected, the drone can randomly select the next node to be collected according to the probability size, thereby gradually planning a mapping path and achieving a more efficient and reasonable mapping path planning to improve the efficiency and quality of drone mapping. Among them, represents the node transition probability from the collection node to the collection node , represents the pheromone concentration of the collection node to the collection node , is the heuristic factor, specifically representing the heuristic information from the collection node to the collection node , , represents the set of optional collection nodes not visited by the drone, represents any one of the optional collection nodes, represents the collection node to the collection node pheromone concentration of, characterizing the collection node to the collection node heuristic information of, characterizing the pheromone influence factor, characterizing the heuristic factor influence factor, used to adjust the influence degrees of the pheromone concentration and the heuristic factor on the transition probability.

[0038] Preferably, characterizing the collection node transfer cost function from transferring to the collection node for quantifying the cost paid by the UAV for transferring from the collection node to the collection node and its formula is , where characterizes the distance between the collection node and the collection node , and respectively characterize the risk factors of the collection node and the collection node , and respectively characterize the node priorities of the collection node and the collection node . , , are the distance weight factor, the risk weight factor, and the node priority weight factor respectively. Specifically, the farther the distance, the greater the proportion in the cost function, that is, the transfer cost will increase with the increase of the distance. For example, in actual surveying and mapping, a long flight distance means that the UAV consumes more power and may also increase the risk of being affected by external interference. Therefore, the distance is an important factor affecting the transfer cost; the risk factor may cover various situations, such as strong airflows, signal interference, building blockages, etc. in the area that may affect the flight safety of the UAV. The higher the flight risk, the greater the value of the risk factor, and the transfer cost will increase; the nodes with high priority may be locations containing important surveying and mapping information, and appear in the cost function in the form of subtracting the corresponding product, indicating that when transferring to high-priority nodes, the transfer cost will be reduced, guiding the UAV to preferentially go to these key nodes for data collection.

[0039] The surveying and mapping search module 30 is used to configure the virtual UAV to start from the collection starting point, calculate the pheromone concentration according to the initial pheromone value, and use the pheromone concentration and the heuristic factor to perform iterative selection of the surveying and mapping sub-regions to establish an optimization result. Among them, after each global path planning is completed, the initial pheromone value is accumulated and updated.

[0040] Preferably, a virtual drone is configured to execute a task starting from a determined acquisition starting point. Among them, the virtual drone is used to simulate the behavior of a real drone in the mapping area. Through path planning simulation, different mapping paths and sub-region selection schemes are explored; the virtual drone calculates the pheromone concentration on each path at the current moment according to the initial pheromone value. In the ant colony algorithm, pheromone volatilizes over time and is also enhanced when the drone passes by. Specifically, a pheromone volatilization coefficient ρ is set, which represents the volatilization ratio of pheromone per unit time, and its value range is between 0 and 1. For example, ρ = 0.1 means that every time a unit time passes, 10% of the pheromone on the path will volatilize. A constant Q is set to control the degree of pheromone enhancement.

[0041] Preferably, since pheromone will naturally volatilize, within each time interval (which can be regarded as an iteration cycle), pheromone volatilization calculation is performed, that is, at time t + 1, the pheromone concentration on path (i, j) (t + 1) first considers the remaining amount after volatilization, and the calculation formula is: τ(t + 1) = (1 - ρ) τ(t), where τ(t) is the pheromone concentration on path (i, j) at time t. At the beginning of the iteration, τ(t) = τ(0), that is, after a time interval, the remaining pheromone on the path is the pheromone concentration at the previous moment multiplied by (1 - ρ). When the virtual drone passes through a certain path, pheromone will be left on this path, thereby enhancing its concentration. Assuming that in one iteration, m virtual drones participate in path planning, when the k-th virtual drone passes through path (i, j), the amount of pheromone added to this path is Δ τ, then after all virtual drones pass by, the total amount of pheromone enhancement on path (i, j) is: , usually, Δ τ is related to the performance of the virtual drone in this path planning. For example, it can be calculated according to the flight distance of the drone : If the k-th drone passes through path (i, j), then Δ τ; if the k-th drone does not pass through path (i, j), then Δ τ = 0, that is, the shorter the flight distance of the drone, the more pheromone it adds to the path it passes through. Finally, adding the remaining amount after pheromone volatilization and the enhancement amount can obtain the final pheromone concentration on path (i, j) at time t + 1. Through continuous iterative calculation, as time goes by and the virtual drone conducts multiple path planning simulations, the pheromone concentration on each path will change dynamically, reflecting the quality of path selection, and further guiding the subsequent selection of mapping sub-regions and path planning.

[0042] Preferably, based on the pheromone concentration and the heuristic factor, use a virtual unmanned aerial vehicle (UAV) to simulate optimization and perform iterative selection of the mapping sub-regions. Specifically, determine parameters such as the pheromone evaporation coefficient ρ, the pheromone influence factor δ, the heuristic factor influence factor φ, and the upper limit N of the number of iterations, and assign an initial pheromone value to the paths between all mapping sub-regions (nodes to be collected). (0), and calculate the heuristic factor according to the transfer cost function of the collection node Then calculate the probability of the virtual UAV transferring to each unvisited node. According to the calculated transfer probability, use the roulette wheel algorithm to randomly select the next mapping sub-region to be visited, that is, regard the transfer probability of each node as a sector area on the roulette wheel. The larger the probability, the larger the sector area, and then determine the selected node by randomly rotating the roulette wheel; the virtual UAV continuously moves to the next unvisited sub-region until all mapping sub-regions are visited, completing a global path planning. Record the total path length L or other evaluation indicators (such as the total collection time, collection cost, etc.) of this path planning, and then perform pheromone accumulation and update (including evaporation operation and pheromone enhancement operation), and calculate the updated pheromone concentration; then perform multiple iterations. After each iteration, compare the evaluation indicator of the current path planning with the previously recorded optimal evaluation indicator. If the current result is better, update the optimal path and the optimal evaluation indicator. When the number of iterations reaches the preset upper limit, stop the iteration, and take the recorded optimal path as the final optimization result. This path is the recommended path for the UAV swarm to perform 3D mapping of the building facade.

[0043] The clustering sharing module 40 is used to control the UAV swarm to perform 3D mapping of the building facade based on the optimization result, cluster and establish a shared information pool according to the real-time mapping positions of the UAV swarm, and upload the real-time mapping results to the shared information pool.

[0044] Preferably, use the optimization result (optimal mapping path planning) to control the UAV swarm to perform 3D mapping of the building facade. Specifically, each UAV starts from the collection starting point according to the path segment assigned to itself and flies to each mapping sub-region (node to be collected) in a predetermined order. For example, the UAV sequentially collects images or data from the sub-regions of different floors and different facades of the building according to the flight route determined by the optimization result to ensure that the entire building facade can be comprehensively and efficiently covered; and during the flight, the UAV uses its own navigation system (such as GPS, inertial navigation, etc.) and flight control algorithms to accurately control the flight attitude and speed to fly along the path of the optimization result. At the same time, perform real-time adjustment according to the actual environmental conditions (such as wind speed, air flow, etc.) to ensure flight stability and safety. For example, when encountering strong winds, the UAV automatically adjusts the flight parameters to stay on the predetermined mapping path.

[0045] Preferably, each drone is equipped with a positioning device (such as a GPS module), which can obtain its own position information (longitude, latitude, altitude, etc.) in real time. The position information is continuously updated as the drone flies and is transmitted to the ground control center or a designated server through a wireless communication module (such as 4G, 5G or Wi-Fi). After receiving the real-time position information of the drone swarm, the ground control center or the server uses a clustering algorithm (such as the K-Means clustering algorithm, the DBSCAN clustering algorithm, etc.) to perform clustering analysis on these position points. The purpose is to divide the positions of drones that are close to each other in space into a group, reflecting the concentration of drones in different areas. For example, if multiple drones are performing mapping near a specific floor or facade of a building, they are clustered into a cluster, representing that this area is the key area for current mapping. Based on the clustering results, a shared information pool is established, that is, a data storage and management platform, which is used to store the real-time position clustering information of the drones and other relevant data (such as flight status, battery power, etc.), ensuring the collaborative operation of the drone swarm, so that each drone can obtain the position and working status information of other drones. The drones use the sensors carried (such as high-definition cameras, lidar, etc.) to collect data from the building facade (which may include image data, point cloud data, 3D model data, etc., recording detailed information of the building facade, such as texture, shape, size, etc.). After preliminary processing (such as data compression, format conversion, etc.), it is uploaded to the shared information pool in real time through a wireless communication link. The shared information pool stores and manages the real-time mapping results, ensuring the security and accessibility of the data, thereby improving the efficiency of drone swarm mapping and the accuracy of data processing.

[0046] Furthermore, the specific configuration of the clustering and sharing module 40 also includes obtaining the spatial position coordinates of each drone in the drone swarm at the current time node; obtaining the next mapping sub-region of each drone in the drone swarm; performing real-time mapping position clustering according to the next mapping sub-region and the spatial position coordinates to generate a real-time mapping position clustering result; identifying the drone with the largest storage space in each cluster concentration in the real-time mapping position clustering result, creating a shared information pool on the identified drone, and receiving the real-time mapping results of other drones within the real-time mapping position clustering result.

[0047] Preferably, each drone is equipped with a corresponding positioning device (such as GPS, Beidou satellite navigation system, etc.), which can determine its own position in space in real time and present it in the form of coordinates (such as longitude, latitude, altitude, etc.). Through wireless communication technology (such as 4G, 5G or a dedicated drone communication link), these spatial position coordinate information is transmitted to the ground control center in real time to obtain the specific position of each drone in space at the current time node; clarify the sub-region (to-be-acquired node) that each drone will go to for surveying and mapping in the next round. For example, if a certain drone has completed the surveying and mapping task of the current sub-region, determine its next-round surveying and mapping sub-region according to the preset path sequence so that the drone can continue the surveying and mapping work according to the plan; then combine the information of the next-round surveying and mapping sub-region of each drone obtained with the current spatial position coordinates, and use clustering algorithms (such as K-Means clustering, DBSCAN clustering, etc.) to analyze the positions of the drones, and group the drones with similar spatial positions or those about to go to the same or similar surveying and mapping sub-regions into one group to form different clustering sets. For example, if the next-round surveying and mapping sub-regions of multiple drones are all concentrated on a certain side of a certain floor of a building and their current spatial positions are also relatively close, these drones are clustered into the same set, and then a real-time surveying and mapping position clustering result is generated.

[0048] Preferably, compare the storage spaces of the drones, find the drone with the largest storage space, and mark it. Select the drone with the largest storage space as the information storage center of this clustering set, and create a shared information pool. Other drones in the clustering set transmit their real-time surveying and mapping results (including acquired image data, point cloud data, surveying and mapping parameters, etc.) to the shared information pool of the marked drone through wireless communication, aiming to centrally store and manage surveying and mapping data within a local range, reduce the burden of data transmission, and facilitate data sharing and collaborative operations among the drones in the same clustering set, improving the surveying and mapping efficiency and data processing ability of the entire drone group. For example, when there are doubts about the data acquired by a certain drone, the surveying and mapping results of other drones at similar positions can be obtained from the shared information pool for comparative analysis.

[0049] The surveying and mapping optimization module 50 is used to perform collaborative decision-making optimization of the local path through the shared information pool and the optimization result, and update the acquisition path according to the collaborative decision-making optimization result.

[0050] The specific configuration of the surveying and mapping optimization module 50 further includes obtaining the environmental shared information in the shared information pool, reconstructing the local environmental mutation constraint according to the environmental shared information; obtaining the surveying and mapping anomaly according to the real-time surveying and mapping results in the shared information pool, and creating an additional surveying and mapping task by using the surveying and mapping anomaly; performing collaborative decision-making optimization of the local path according to the local environmental mutation constraint, the additional surveying and mapping task and the optimization result.

[0051] Preferably, during the three-dimensional mapping of the building facade by the UAV swarm, based on the data in the shared information pool and the previous optimization results, the local path is dynamically adjusted and optimized to cope with environmental changes and anomalies that occur during the mapping process. Specifically, obtain the environmental shared information in the shared information pool, such as meteorological conditions (wind speed, wind direction, temperature, etc.), obstacle information (newly added temporary obstacles, local structural changes of buildings, etc.), then judge whether there is an environmental mutation situation according to the environmental shared information, and determine the factors causing the environmental mutation, and then determine the corresponding constraint parameters. Finally, according to the determined constraint parameters, construct specific local environmental mutation constraint conditions, that is, if situations such as sudden strong winds and newly emerged obstacles occur, reconstruct the local environmental mutation constraints. For example, for wind speed mutation, the maximum allowable flight speed, flight height range, and wind direction areas to be avoided under the current wind speed can be determined according to the performance parameters of the UAV; for newly added obstacles, the no-fly area range around them should be clarified, that is, the area within a certain distance from the obstacle where the UAV cannot enter. The local environmental mutation constraints cover various environmental change situations that may affect the UAV flight and mapping tasks, which can ensure that the UAV can comply with these constraints during flight and complete the mapping task safely and efficiently.

[0052] Preferably, by analyzing the real-time mapping results in the shared information pool, anomalies that occur during the mapping process are discovered, such as blurred image data, missing point cloud data, and excessive deviation of mapping data from the expected values. If mapping anomalies are found, new additional mapping tasks are created for these anomalies. For example, if the image data of a certain area is blurred, high-definition image acquisition of this area needs to be carried out again. If the point cloud data is missing, additional laser scanning tasks for this area need to be added. These additional mapping tasks are entered into the path planning to ensure the integrity and accuracy of the mapping data.

[0053] Preferably, the local environmental mutation constraints, additional mapping tasks, and optimization results are used as the basis for collaborative decision-making. The optimization results provide the initial global optimal path planning, while the local environmental mutation constraints and additional mapping tasks reflect the changes in the current actual situation, and are used to optimize the collaborative decision-making of local paths, that is, to make adjustments on the basis of the original path planning. Specifically, path optimization algorithms (such as ant colony algorithm, genetic algorithm, etc.) are used to combine the above information to optimize the collaborative decision-making of local paths, comprehensively consider the new constraint conditions and additional tasks, and complete the additional mapping tasks as efficiently as possible under the premise of meeting the environmental constraints, while minimizing the impact on the original planned path. For example, through algorithm calculation, how to reasonably arrange the drones to complete the additional mapping tasks while avoiding new obstacles, so as to minimize the overall mapping time and cost. Finally, according to the optimization results of the collaborative decision-making of local paths, the acquisition paths of the drones are updated, and the optimized path information is sent to the corresponding drones, so that they can perform mapping operations according to the new paths. The updated acquisition paths can better adapt to environmental changes and mapping anomalies, improve the efficiency and quality of mapping, and ensure the smooth completion of mapping tasks.

[0054] Furthermore, the specific configuration of the three-dimensional building facade mapping system based on drones and AI further includes an anomaly recognition module, which is used to record the acquisition anomalies of the drone swarm, perform anomaly level discrimination, match the response plans in the response plan library according to the anomaly level discrimination results, and perform anomaly management according to the response plan matching results. The response plan library includes an anomaly marking response plan and a temporary task creation response plan.

[0055] Preferably, the anomaly recognition module is used to continuously monitor the data collection work of the UAV, record various possible data collection anomalies, such as incomplete collected data, blurred images, data errors caused by sensor failures, abnormal distances from the target building, etc. Then, the recorded data collection anomalies are classified according to their severity and impact on the mapping task. For example, a slightly blurred image may be determined as an anomaly of a lower level, while a complete sensor failure resulting in a large amount of missing data may be determined as an anomaly of a higher level. Then, for different anomaly levels, response plans are matched in the anomaly marking response plan library. The anomaly marking response plan library contains anomaly marking response plans and temporary task creation response plans. Specifically, the anomaly marking response plan mainly marks and records the anomaly situation, recording information such as the time, location, and type of the anomaly occurrence, which is convenient for technicians to conduct detailed inspections and analyses after the task is completed. The temporary task creation response plan creates a temporary task according to the specific situation of the anomaly to solve the anomaly problem. For example, if the data collection in a certain area is missing due to a UAV failure, a new task is created according to the discrimination result of the anomaly level, and other UAVs are arranged to re-collect the area to ensure the integrity of the mapping data. Finally, the anomalies are managed according to the matching results. If an anomaly marking response plan is matched, the anomaly is marked and recorded according to the plan. If a temporary task creation response plan is matched, the corresponding temporary task will be automatically created and executed to solve the anomaly situation and ensure that the mapping task can continue smoothly. Through the anomaly recognition module, various data collection anomaly situations can be effectively managed and processed, improving the reliability and accuracy of UAV swarm mapping.

[0056] In the foregoing, with reference to Figure 1 a three-dimensional mapping system for building facades based on UAVs and AI according to an embodiment of the present invention was described in detail. Next, with reference to Figure 2 a three-dimensional mapping method for building facades based on UAVs and AI according to an embodiment of the present invention will be described.

[0057] The three-dimensional mapping method for building facades based on UAVs and AI, as Figure 2As shown, the method includes: reading the calibration scene information of the scene to be surveyed and mapped, performing information parsing on the calibration scene information, establishing N surveying and mapping sub-regions, and marking each surveying and mapping sub-region as a node to be collected; after the collection starting point of the distributed drone swarm, assigning an initial pheromone value to each surveying path, configuring a heuristic factor, and calculating the node transition probability of each node to be collected; configuring a virtual drone to start from the collection starting point, calculating the pheromone concentration according to the initial pheromone value, and using the pheromone concentration and the heuristic factor to perform iterative selection of surveying and mapping sub-regions to establish an optimization result, where, after each global path planning is completed, the initial pheromone value is accumulated and updated; controlling the drone swarm to perform three-dimensional surveying and mapping of the building facade based on the optimization result, clustering according to the real-time surveying and mapping positions of the drone swarm to establish a shared information pool, and uploading the real-time surveying and mapping results to the shared information pool; performing collaborative decision-making optimization of the local path through the shared information pool and the optimization result, and updating the collection path according to the collaborative decision-making optimization result.

[0058] In a possible implementation manner, the three-dimensional surveying and mapping method of the building facade based on drones and AI further includes: a drawing reading module, configured to read the scene drawing of the scene to be surveyed and mapped, and use the scene drawing as the calibration scene information; a grid segmentation module, configured to establish a median grid granularity, perform grid division on the building facade of the scene to be surveyed and mapped based on the median grid granularity, and generate an initial grid segmentation result; a grid update module, configured to perform integrity identification of elements within the grid on the initial grid segmentation result, perform grid migration clustering according to the integrity identification result, generate a grid migration clustering result, and establish N surveying and mapping sub-regions according to the grid migration clustering result.

[0059] In a possible implementation manner, the three-dimensional surveying and mapping method of the building facade based on drones and AI further includes: obtaining a set of elements not covered by the initial grid segmentation result, and marking it as an updated element set; obtaining the size data of each element in the updated element set, and selecting clustering elements according to the matching degree between the size data and the median grid granularity; using the clustering elements to complete grid migration clustering.

[0060] In a possible implementation manner, the three-dimensional surveying and mapping method of the building facade based on drones and AI further includes: establishing a collection node transfer cost function as follows: ; where represents the collection node transfer cost function from collection node to collection node , represents the distance between collection node and collection node , and respectively characterize the acquisition nodes and the acquisition nodes risk factors, and respectively characterize the node priorities of the acquisition nodes and the acquisition nodes ; , , are respectively the distance weight factor, the risk weight factor, and the node priority weight factor; configure the heuristic factor according to the acquisition node transfer cost function, and calculate the node transfer probability using the heuristic factor as follows: ; wherein, represents the node transfer probability from the acquisition node to the acquisition node , represents the pheromone concentration from the acquisition node to the acquisition node , is the heuristic factor, specifically representing the heuristic information from the acquisition node to the acquisition node , , represents the set of optional acquisition nodes not visited by the UAV, represents any one of the optional acquisition nodes, represents the pheromone concentration from the acquisition node to the acquisition node , represents the heuristic information from the acquisition node to the acquisition node , represents the pheromone influence factor, represents the heuristic factor influence factor.

[0061] In a possible implementation manner, the three-dimensional mapping method for the building facade based on the UAV and AI further includes: obtaining the spatial position coordinates of each UAV in the UAV group at the current time node; obtaining the next-round mapping sub-region of each UAV in the UAV group; performing real-time mapping position clustering according to the next-round mapping sub-region and the spatial position coordinates to generate a real-time mapping position clustering result; identifying the UAV with the largest storage space in each cluster set in the real-time mapping position clustering result, creating a shared information pool on the identified UAV, and receiving the real-time mapping results of other UAVs in the real-time mapping position clustering result.

[0062] In a possible implementation manner, the three-dimensional surveying and mapping method for the building facade based on an unmanned aerial vehicle (UAV) and artificial intelligence (AI) further includes: obtaining environmental sharing information in a shared information pool, and reconstructing local environmental mutation constraints according to the environmental sharing information; obtaining surveying and mapping anomalies based on real-time surveying and mapping results in the shared information pool, and creating additional surveying and mapping tasks by using the surveying and mapping anomalies; and performing collaborative decision optimization of local paths according to the local environmental mutation constraints, the additional surveying and mapping tasks, and the optimization result.

[0063] In a possible implementation manner, the three-dimensional surveying and mapping method for the building facade based on an unmanned aerial vehicle (UAV) and artificial intelligence (AI) further includes: dynamically collecting real-time microscopic images of target magnetic flocs in the target water body; performing feature evaluation and analysis on the real-time microscopic images according to a microscopic state evaluation mechanism to obtain a real-time microscopic state; obtaining any information value coefficient of the real-time microscopic state for any water quality index; combining the any information value coefficient to perform predictive analysis on the real-time water quality parameters to obtain predicted real-time water quality parameters; and calibrating the real-time water quality parameters with the predicted real-time water quality parameters.

[0064] In a possible implementation manner, the three-dimensional surveying and mapping method for the building facade based on an unmanned aerial vehicle (UAV) and artificial intelligence (AI) further includes: collecting real-time image feature information of the real-time microscopic images; performing normalization weighting on real-time image color parameters and real-time image structure parameters in the real-time image feature information to obtain real-time image feature values; and combining microscopic state coefficients stored in the microscopic state evaluation mechanism to perform standardization correction on the real-time image feature values to obtain the real-time microscopic state.

[0065] In a possible implementation manner, the three-dimensional surveying and mapping method for the building facade based on an unmanned aerial vehicle (UAV) and artificial intelligence (AI) further includes: extracting a first historical data set from a magnetic coagulation precipitation database; performing mutual information analysis on a first historical microscopic state of a first magnetic floc and first historical index parameters corresponding to a first historical any water quality index in the first historical data set to obtain a first historical mutual information coefficient; and using the first historical mutual information coefficient as the any information value coefficient.

[0066] The three-dimensional surveying and mapping system for the building facade based on an unmanned aerial vehicle (UAV) and artificial intelligence (AI) provided by an embodiment of the present invention can execute the three-dimensional surveying and mapping method for the building facade based on an unmanned aerial vehicle (UAV) and artificial intelligence (AI) provided by any embodiment of the present invention, and has function modules and beneficial effects corresponding to the execution of the method.

[0067] Although this application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server. The included individual units and modules are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of mutual distinction and are not used to limit the protection scope of the present invention.

[0068] The above specific implementation manners do not constitute a limitation to the protection scope of this application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principle of this application shall be included within the protection scope of this application.

Claims

1. A three-dimensional mapping system for building facades based on drones and AI, characterized in that, The system includes: A scene parsing module, which is used to read the calibration scene information of the scene to be surveyed and mapped, perform information parsing on the calibration scene information, establish N surveying and mapping sub-regions, and mark each surveying and mapping sub-region as a node to be collected; An acquisition planning module, which is used to allocate an initial pheromone value on each surveying and mapping path after the acquisition starting point of the distributed drone swarm, configure a heuristic factor, and calculate the node transition probability of each node to be collected; A surveying and mapping search module, which is used to configure a virtual drone to start from the acquisition starting point, calculate the pheromone concentration according to the initial pheromone value, and perform iterative selection of surveying and mapping sub-regions by using the pheromone concentration and the heuristic factor to establish an optimization result. Among them, after each global path planning is completed, the initial pheromone value is accumulated and updated; A clustering and sharing module, which is used to control the drone swarm to perform three-dimensional surveying and mapping of the building facade based on the optimization result, cluster and establish a shared information pool according to the real-time surveying and mapping positions of the drone swarm, and upload the real-time surveying and mapping results to the shared information pool; A surveying and mapping optimization module, which is used to perform collaborative decision-making optimization of local paths through the shared information pool and the optimization result, and update the acquisition path according to the collaborative decision-making optimization result.

2. The 3D mapping system for building facades based on drones and AI according to claim 1, characterized in that, The scene parsing module includes: A drawing reading module, which is used to read the scene drawing of the scene to be surveyed and mapped, and use the scene drawing as the calibration scene information; A grid segmentation module, which is used to establish a median grid granularity, perform grid division on the building facade of the scene to be surveyed and mapped based on the median grid granularity, and generate an initial grid segmentation result; A grid update module, which is used to identify the integrity of elements within the grid for the initial grid segmentation result, perform grid migration clustering according to the integrity identification result, generate a grid migration clustering result, and establish N surveying and mapping sub-regions according to the grid migration clustering result.

3. The three-dimensional surveying and mapping system for building facades based on drones and AI according to claim 2, characterized in that, In the grid update module, performing grid migration clustering according to the integrity identification result includes: Obtaining a set of elements not covered by the initial grid segmentation result, and marking it as an updated element set; Obtaining the size data of each element in the updated element set, and selecting clustering elements according to the matching degree between the size data and the median grid granularity; Using the clustering elements to complete grid migration clustering.

4. The 3D mapping system for building facades based on drones and AI according to claim 1, characterized in that, In the acquisition planning module, calculating the node transition probability of each node to be collected includes: Establishing an acquisition node transition cost function as follows: ； Among them, represents the acquisition node transfer cost function from the acquisition node to the acquisition node . represents the distance between the acquisition node and the acquisition node . and respectively represent the risk factors of the acquisition node and the acquisition node . and respectively represent the node priorities of the acquisition node and the acquisition node . , , are the distance weight factor, risk weight factor, and node priority weight factor respectively; Configuring a heuristic factor according to the acquisition node transition cost function, and using the heuristic factor to calculate the node transition probability as follows: ； Among them, represents the node transition probability from the collection node to the collection node ; represents the pheromone concentration from the collection node to the collection node ; is the heuristic factor, specifically representing the heuristic information from the collection node to the collection node ; , represents the set of optional collection nodes not visited by the UAV, represents any optional collection node, represents the pheromone concentration from the collection node to the collection node ; represents the heuristic information from the collection node to the collection node ; represents the pheromone influence factor, represents the heuristic factor influence factor.

5. The 3D mapping system for building facades based on drones and AI according to claim 1, characterized in that, In the clustering and sharing module, clustering and establishing a shared information pool according to the real-time surveying and mapping positions of the drone swarm includes: Obtaining the spatial position coordinates of each drone in the drone swarm at the current time node; Obtaining the next surveying and mapping sub-region of each drone in the drone swarm; Performing clustering of real-time surveying and mapping positions according to the next surveying and mapping sub-region and the spatial position coordinates to generate a real-time surveying and mapping position clustering result; Identifying the drone with the largest storage space in each cluster set in the real-time surveying and mapping position clustering result, creating a shared information pool on the identified drone, and receiving the real-time surveying and mapping results of other drones within the real-time surveying and mapping position clustering result.

6. The 3D mapping system for building facades based on drones and AI according to claim 1, characterized in that, In the mapping optimization module, collaborative decision-making optimization of local paths is performed through the shared information pool and the optimization results, including: Obtain the environmental shared information in the shared information pool, and reconstruct the local environmental mutation constraints according to the environmental shared information; Obtain mapping anomalies based on the real-time mapping results in the shared information pool, and create additional mapping tasks using the mapping anomalies; Perform collaborative decision-making optimization of local paths according to the local environmental mutation constraints, the additional mapping tasks, and the optimization results.

7. The 3D mapping system for building facades based on drones and AI according to claim 1, characterized in that, The system further includes: An anomaly recognition module, which is used to record the acquisition anomalies of the drone swarm, perform anomaly level discrimination, match the response plans in the response plan library according to the anomaly level discrimination results, and perform anomaly management according to the response plan matching results. The response plan library includes anomaly marking response plans and temporary task creation response plans.

8. A three-dimensional surveying method for building facades based on drones and AI, characterized in that, The method is applied to any building facade three-dimensional mapping system based on drones and AI described in claims 1-7. The method includes: Read the calibration scene information of the scene to be mapped, perform information parsing of the calibration scene information, establish N mapping sub-regions, and mark each mapping sub-region as a node to be collected; After the starting points of the distributed drone swarm for acquisition, allocate initial pheromone values on each mapping path, configure the heuristic factor, and calculate the node transition probability of each node to be collected; Configure a virtual drone to start from the acquisition starting point, calculate the pheromone concentration according to the initial pheromone value, and perform iterative selection of mapping sub-regions using the pheromone concentration and the heuristic factor to establish an optimization result. Among them, after each global path planning is completed, the initial pheromone value is accumulated and updated; Control the drone swarm to perform three-dimensional mapping of the building facade based on the optimization result, cluster and establish a shared information pool according to the real-time mapping positions of the drone swarm, and upload the real-time mapping results to the shared information pool; Perform collaborative decision-making optimization of local paths through the shared information pool and the optimization result, and update the acquisition path according to the collaborative decision-making optimization result.

Images