An automatic road network generation method and system

Abstract

This invention relates to the field of traffic planning technology, and in particular to an automatic road network generation method and system. The method involves acquiring design drawing data of a target building and collecting user location heatmaps within that building; based on the design drawing data, extracting a first route information set, and extracting neighboring design data and user location information for each first route to obtain a first neighboring design dataset and multiple first user location information sets; fitting a second route information set based on the user location heatmaps; predicting the location signal impact error for each first route based on the first neighboring design dataset to obtain a location error parameter set; verifying the location error parameter set and the multiple first user location information sets, configuring generation weights, and fusing the first and second route information sets to obtain a road network information set. This achieves the technical effect of generating a high-precision, highly applicable road network.

Description

Technical Field

[0001] This invention relates to the field of traffic planning technology, and in particular to an automatic road network generation method and system. Background Technology

[0002] Existing indoor road network generation technologies generally suffer from problems such as a single data source, insufficient physical constraint modeling, coarse error assessment, and rigid fusion methods. They rely solely on design drawings or sparse positioning data, failing to form a multi-dimensional data loop. This results in deviations between the generated road network and the real-world scenario, especially in complex architectural environments where accuracy is insufficient. Consequently, these technologies suffer from low accuracy and poor applicability in the generated road networks. Summary of the Invention

[0003] This invention addresses the technical problems of low accuracy and poor applicability of road networks generated in existing technologies by providing an automatic road network generation method and system.

[0004] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0005] In a first aspect, the present invention provides an automatic road network generation method, comprising: acquiring design drawing data of a target building and collecting user location heatmaps within the target building; extracting a first route information set based on the design drawing data, and extracting neighboring design data and user location information for each first route information to obtain a first neighboring design dataset and multiple first user location information sets; fitting a second route information set based on the user location heatmaps; predicting the location signal impact error for each first route information based on the first neighboring design dataset to obtain a location error parameter set; verifying the location error parameter set and the multiple first user location information sets, configuring generation weights, and fusing the first route information set and the second route information set to obtain a road network information set.

[0006] Optionally, the design drawing data of the target building is obtained, and user location heatmaps within the target building are collected, including: obtaining the design drawing data of the target building; collecting user location information within the target building within a preset time range, and generating user location heatmaps, wherein each user location information includes location coordinates.

[0007] Optionally, based on the design drawing data, a first route information set is extracted, and neighboring design data and user location information for each first route information are extracted to obtain a first neighboring design dataset and multiple first user location information sets, including: extracting the first route information set based on the design drawing data; extracting neighboring design data within a preset distance for each first route information within the design drawing data to obtain a first neighboring design dataset; and extracting user location information within a preset distance for each first route information within the first user location heatmap to obtain multiple first user location information sets.

[0008] Optionally, based on the user location heatmap, a second route information set is obtained by fitting the data, including: clustering user location information whose distance to the user location heatmap is less than a preset route distance to obtain a location information clustering result set; randomly fitting second route information within each location information clustering result, calculating the average distance between all user location information and the second route information within the location information clustering result to obtain the fitted route distance; and optimizing the fitting of the second route information within each location information clustering result to obtain multiple second route information sets with the smallest fitted route distance, thereby obtaining the second route information set.

[0009] Optionally, based on the first proximity design dataset, the positioning signal impact error prediction for each first route information is performed to obtain a positioning error parameter set, including: calling a positioning error impact analyzer, wherein the positioning error impact analyzer is obtained by machine learning training, and the training data includes a sample proximity design dataset and a sample positioning error parameter set; inputting each first proximity design data in the first proximity design dataset into the positioning error impact analyzer, and identifying the output to obtain a positioning error parameter set, wherein each positioning error parameter includes the error distance of the user positioning information.

[0010] Optionally, the positioning error parameter set and multiple first user positioning information sets are verified and weights are configured, including: calculating the average distance between each first user positioning information set and the corresponding first route information based on the multiple first user positioning information sets to obtain multiple actual positioning error parameters; calculating the similarity between the multiple positioning error parameters in the positioning error parameter set and the multiple actual positioning error parameters to obtain error similarity; using the error similarity to correct the preset positioning weights to obtain positioning weights, and calculating the design weights.

[0011] The process of fusing the first route information set and the second route information set to obtain a road network information set includes: calculating the route distance between each first route information set in the first route information set and each second route information set in the second route information set; combining the first route information set with the second route information set with the smallest route distance to obtain multiple route information groups; and weighting and fusing the route coordinates of the second route information set and the first route information set within each route information group according to the positioning weight and design weight to obtain multiple fused route information sets as the road network information set.

[0012] In a second aspect, the present invention provides an automatic road network generation system, comprising:

[0013] The user location acquisition module is used to acquire the design drawing data of the target building and collect the user location heat map within the target building;

[0014] The route information acquisition module is used to extract a first route information set based on the design drawing data, and extract the neighboring design data and user location information of each first route information to obtain a first neighboring design dataset and multiple first user location information sets, and to fit a second route information set based on the user location heat map.

[0015] The positioning error prediction module is used to predict the positioning signal impact error of each first route information according to the first neighbor design dataset, and obtain the positioning error parameter set.

[0016] The route information fusion module is used to verify the positioning error parameter set and multiple first user positioning information sets, configure the generation weights, and fuse the first route information set and the second route information set to obtain a road network information set.

[0017] By implementing this invention, it is possible to obtain the design drawing data of the target building and collect the user location hotspot map within the target building. The user location hotspot map reflects the actual traffic situation and provides a basis for road network correction.

[0018] By implementing this invention, a first route information set can be extracted from the design drawing data, and adjacent design data and user location information of each first route information can be extracted to obtain a first adjacent design dataset and multiple first user location information sets. A second route information set can be obtained by fitting the user location heat map. By associating it with its adjacent physical structure, a spatial reference can be provided for analyzing location signal obstruction. By clustering and fitting the user location heat map, the actual user travel habits can be reflected, providing a basis for the subsequent generation of road network information.

[0019] By implementing this invention, it is possible to predict the positioning signal impact error of each first route information based on the first proximity design dataset, obtain a positioning error parameter set, establish the correlation between proximity design data and positioning error, and achieve quantitative assessment of signal interference.

[0020] By implementing this invention, the positioning error parameter set and multiple first user positioning information sets can be verified and weights can be configured. The first route information set and the second route information set can be fused to obtain a road network information set. By quantifying the error similarity, the weight allocation can be ensured to conform to the real scenario, and a road network that conforms to the structural logic and is close to real traffic can be generated.

[0021] In summary, by implementing this invention, the technical effect of generating a high-precision and highly applicable road network can be achieved. Attached Figure Description

[0022] Figure 1 A flowchart illustrating an automatic road network generation method provided by the present invention;

[0023] Figure 2 This is a schematic diagram of the automatic road network generation system provided by the present invention.

[0024] In the attached diagram, the components represented by each number are as follows:

[0025] User location acquisition module 11, route information acquisition module 12, location error prediction module 13, and route information fusion module 14. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0028] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0029] Example 1, as Figure 1 As shown, this embodiment of the invention provides an automatic road network generation method, including:

[0030] S100: Obtain the design drawings of the target building and collect user location hotspot maps within the target building;

[0031] S200: Based on the design drawing data, extract the first route information set, and extract the neighboring design data and user positioning information of each first route information to obtain the first neighboring design dataset and multiple first user positioning information sets. Based on the user positioning heat map, fit to obtain the second route information set.

[0032] S300: Based on the first neighboring design dataset, predict the positioning signal impact error of each first route information to obtain a positioning error parameter set;

[0033] S400: Verify the positioning error parameter set and multiple first user positioning information sets, configure generation weights, and fuse the first route information set and the second route information set to obtain a road network information set.

[0034] In step S100 of this application embodiment, the design drawing data of the target building is obtained, and the user location hotspot map within the target building is collected, including:

[0035] Obtain the design drawings of the target building;

[0036] Collect user location information within the target building within a preset time range and generate a user location heat map, where each user location information includes location coordinates.

[0037] In this embodiment, obtaining the design drawing data of the target building is used as a basis to draw a user location heatmap within the target building based on user location information. The design drawing data of the target building can be a BIM file or CAD file from the architectural design phase, from which two-dimensional plan route information is extracted as the design drawing data. This drawing data should include geometric data such as route coordinates, walls, beams, and columns.

[0038] The user location information within the target building can be obtained through the location information of the user's mobile phone upon entering the building, such as WiFi fingerprint positioning and Bluetooth beacon positioning. It can also be obtained through visual positioning methods using cameras and image recognition. Specifically, when using mobile phone positioning, user location permissions can be obtained through app pop-ups or scene prompts to ensure compliance with privacy regulations. The historical time period within the aforementioned preset time range for collecting user location information within the target building covers the time period for drawing user location heatmaps. This historical time can be set to the operating hours of the target building for the most recent 100 working days (e.g., 9:00-21:00). The frequency of collecting user location data within the historical time period can be once per second, with the addition of metadata such as timestamps and device IDs.

[0039] Furthermore, in creating the heatmap, the building plan can be divided into a fixed-size grid (e.g., 0.5m × 0.5m), and the number of location points within each grid can be counted to form a density matrix. Then, a Gaussian kernel function is used to smooth the discrete location points, and the density value of each coordinate point is calculated to generate a continuous heatmap (the darker the color, the denser the location points). Finally, GIS tools (such as Leaflet and QGIS) can be used to overlay the heatmap onto the building plan (i.e., the design drawing data of the aforementioned target building) to visually display high-frequency traffic areas (such as the main passageway of a shopping mall or the area near the hospital registration desk).

[0040] In addition, heatmap data can be stored as raster files (such as GeoTIFF) or structured data (such as coordinate-density mapping tables in JSON format) for subsequent cluster analysis.

[0041] In step S200 of this application embodiment, a first route information set is obtained based on the design drawing data, and adjacent design data and user location information of each first route information are extracted to obtain a first adjacent design dataset and multiple first user location information sets, including:

[0042] Based on the design drawing data, the first route information set is extracted;

[0043] Within the design drawing data, extract the neighboring design data within a preset distance for each first route information to obtain the first neighboring design dataset;

[0044] Within the first user location heatmap, user location information within a preset distance of each first route information is extracted to obtain multiple sets of first user location information.

[0045] In this embodiment, the first route information set is a collection of existing route information obtainable from design drawing data. It can be obtained by directly identifying the geometric boundaries of passable areas (such as corridors, staircases, and elevator lobbies) from design drawings (CAD / BIM) and extracting centerlines or central axes. Furthermore, the routes need to undergo topology processing, dividing continuous paths into nodes (such as intersections and turning points) and edges (line segments connecting two nodes) to form a structured route network. Each route is assigned a unique identifier and its attributes (such as length, width, and direction of travel) are labeled. The coordinates of the route's start point, end point, and intermediate control points are extracted to form a complete geometric description.

[0046] In this embodiment of the application, the neighboring design data refers to the walls, beams and other structures adjacent to the line. These structures can block the positioning signal and cause positioning deviation. Therefore, it is necessary to include the influence of the neighboring design data on the positioning deviation in the parameters required for automatic road network generation.

[0047] Specifically, a preset distance threshold (e.g., 2 meters) can be set for each route to define its neighborhood range. Then, physical structures within the preset distance, such as walls, columns, and fixed equipment, are searched in the design drawings. Then, the association between the route and the adjacent structure is established (e.g., "Route R1, adjacent wall W1, distance 0.8 meters"). The above-mentioned first neighbor design data is collected for each first route to form a structured first neighbor design dataset.

[0048] Next, user location information within a preset distance for each first route needs to be extracted from the first user location heatmap. In one possible implementation, for each first route, location points within a preset distance range (e.g., 5 meters) can be selected from the user location heatmap. For curved or polygonal routes, the minimum distance from a point to a line segment is used to determine correlation. The location points that meet the criteria are grouped according to their respective routes to form multiple sets of first user location information.

[0049] In step S200 of this application embodiment, a second route information set is obtained by fitting the user location heatmap, including:

[0050] Cluster the location information of users whose location hotspots are less than the preset route distance to obtain a location information clustering result set;

[0051] Within each location information clustering result, second route information is randomly fitted, and the average distance between all user location information and second route information within the location information clustering result is calculated to obtain the fitted route distance.

[0052] Within each location information clustering result, the fitting optimization of the second route information is performed to obtain multiple second route information with the smallest fitting route distance, thus obtaining the second route information set.

[0053] In this embodiment of the application, clustering user location information whose distance to the user location hotspot map is less than the preset route distance is for the purpose of fitting and obtaining second route information, which will be used to correct the first route information (i.e., the existing route).

[0054] In one implementation, DBSCAN (density-based spatial clustering) or OPTICS algorithms can be used to group location points within a preset route distance (e.g., 2 meters) into the same cluster. During classification, the neighborhood radius parameter can be dynamically adjusted; for example, increasing the radius in sparse areas (e.g., at the end of a corridor) and decreasing the radius in dense areas (e.g., in a lobby) to avoid over-segmentation or under-segmentation. During analysis, isolated points not assigned to any cluster are considered noise (e.g., temporary stops, location error points) and excluded from subsequent analysis. Clusters that overlap spatially or are too close (e.g., pedestrian trajectories in adjacent corridors) are merged by calculating inter-cluster similarity (e.g., Jaccard coefficient). The location information clustering result set is obtained using the above methods.

[0055] Furthermore, for each clustering result, multiple candidate polylines (e.g., 100) are randomly generated within its convex hull. Each polyline is formed by connecting several control points (the number of control points is adaptively adjusted according to the cluster size). The constraint for generating path polylines is that the starting and ending points of the polyline must be located at the entrance and exit positions of the cluster boundary (identified by the density gradient of the heat map) to avoid generating invalid paths. Here, the convex hull is a geometric concept, referring to the smallest convex polygon (in a two-dimensional scene) containing a given set of points; while control points are key coordinate points used to define the shape of polylines or curves. In polyline fitting, control points are connected sequentially by straight line segments to form piecewise linear polylines.

[0056] For each candidate polyline, calculate the vertical distance from all points within the cluster to the polyline, and take the average value as the fitted route distance.

[0057] Next, route fitting optimization is required. Simulated annealing can be used, which randomly perturbs the control point positions of the current optimal polyline, accepting changes that reduce the average distance, and with a certain probability, accepting deteriorating changes to escape local optima. The termination condition for iteration can be set to stop optimization when the average distance improvement is less than a threshold (e.g., 0.1 meters) after N consecutive iterations or when the maximum number of iterations is reached. Here, N can be a positive integer value determined according to actual needs. Using the above method, multiple second route information sets with the minimum fitted route distance can be obtained iteratively, resulting in a second route information set.

[0058] In step S300 of this application embodiment, based on the first proximity design dataset, the positioning signal impact error prediction for each first route information is performed to obtain a positioning error parameter set, including:

[0059] The positioning error impact analyzer is invoked, wherein the positioning error impact analyzer is obtained by machine learning training, and the training data includes a sample nearest neighbor design data set and a sample positioning error parameter set;

[0060] Each first neighbor design data in the first neighbor design dataset is input into the positioning error impact analyzer, and the positioning error parameter set is obtained by identifying the output, wherein each positioning error parameter includes the error distance of the user positioning information.

[0061] In this embodiment, the sample neighbor design data set can be obtained through the method described in step S200, for example, "Route R1, adjacent wall W1, distance 0.8 meters". The sample positioning error parameter set is a set of error parameters for positioning in the route of the sample neighbor design data. Specifically, it can be obtained by collecting user positioning points in the corresponding building area using multimodal positioning technology (such as WiFi + inertial navigation), comparing the actual distance between the positioning point and the design route, and calculating the error distance (e.g., if a point is 1.5 meters away from the design route, the error parameter is +1.5 meters). The neighbor design data and positioning error parameters are matched one-to-one to form the training samples of the positioning error impact analyzer, namely the first neighbor design dataset and the positioning error parameter set. The aforementioned dataset is divided into a training set, a validation set, and a test set in an 8:1:1 ratio to ensure that the building types in each set are evenly distributed, and is used to train the positioning error impact analyzer.

[0062] Next, feature extraction needs to be performed on each record in the first neighbor design dataset to generate a feature vector isomorphic to the training data. For example, the neighbor structure of route R1 includes a 24cm thick brick wall (0.5 meters away), which can be converted into a feature vector: [0.24, 0.5], and so on.

[0063] In one possible implementation, since the relationship between localization error and neighboring design data is typically non-linear, the localization error impact analyzer can be constructed using gradient boosting trees. Regarding the parameter settings for the localization error impact analyzer, the learning rate can be set to 0.05-0.1; the number of trees can be set to 100-300 (the number of decision trees in an ensemble, adjusted using a validation set); the tree depth can be set to 3-6 to limit the complexity of a single tree and prevent overfitting; the subsampling ratio can be set to 0.8-1.0 to control the proportion of samples used by each tree and enhance generalization ability; and the regularization parameter can be an L1 / L2 regularization coefficient to suppress overfitting.

[0064] In training the positioning error impact analyzer, training is stopped if the mean absolute error (MSE) shows no improvement after 10 consecutive iterations on the validation set to prevent overfitting. The parameter combination with the smallest MSE on the validation set is selected as the optimal parameter. The evaluation metric for the positioning error impact analyzer can be the mean absolute error (MAE). For example, the model can be set to converge when the MAE < 0.3 meters, thus obtaining the positioning error impact analyzer.

[0065] Finally, the feature vector extracted from the first neighbor design data is input into the trained positioning error impact analyzer to output the predicted error distance value for the route (e.g., 1.2 meters). Inputting each first neighbor design data point from the first neighbor design dataset into the positioning error impact analyzer allows for the identification and output of the positioning error parameter set.

[0066] In step S400 of this application embodiment, the positioning error parameter set and multiple first user positioning information sets are verified, and weights are configured and generated, including:

[0067] Based on the multiple first user positioning information sets, calculate the average distance between each first user positioning information set and the corresponding first route information to obtain multiple actual positioning error parameters;

[0068] Calculate the similarity between multiple positioning error parameters in the positioning error parameter set and multiple actual positioning error parameters to obtain the error similarity.

[0069] The preset positioning weights are corrected and calculated using error similarity to obtain the positioning weights, and the design weights are then calculated.

[0070] In this embodiment, the average distance between the first user location information set and the corresponding first route information is actually calculated by taking the shortest distance from each point in each first user location information set (containing multiple location points) to the corresponding first route information, obtaining multiple shortest distances, and averaging these multiple shortest distances, which is the average distance between each first user location information set and the corresponding first route information. If the route is a polyline, the route is divided into multiple line segments, the perpendicular distance from the point to each line segment is calculated, and the minimum value is taken as the distance from the point to the route.

[0071] Furthermore, points exceeding a preset threshold (e.g., 5 meters) can be defined as outliers and excluded from the calculation. The arithmetic mean of the distances to the remaining valid positioning points is then taken to obtain the actual positioning error parameters for the route. Multiple actual positioning error parameters can be calculated using this method.

[0072] Further analysis requires calculating the similarity between multiple positioning error parameters and multiple actual positioning error parameters within the positioning error parameter set. Essentially, this compares the closeness between the "predicted positioning error (positioning error parameter)" and the "actual measured positioning error (actual positioning error parameter)." This can be measured using a value between 0 and 1 (similarity). First, calculate the average deviation between the two values, which can be expressed as the absolute value of the difference between the positioning error parameter and the actual positioning error parameter, i.e., average deviation = |positioning error parameter - actual positioning error parameter|. If the positioning error parameter = 1 meter and the actual positioning error parameter = 1.2 meters, then the average deviation can be calculated as |1 - 1.2| = 0.2 meters. Next, a maximum allowable error difference (e.g., 1.5 meters) can be set, indicating that when the difference between the predicted error and the actual error exceeds this value, the similarity is 0. Then, divide the average deviation by the maximum allowable error difference to obtain the calculated deviation ratio, for example, 0.2 / 1.5 ≈ 0.13.

[0073] Finally, subtract the deviation ratio from 1 to obtain the error similarity. For example, error similarity = 1 - 0.13 = 0.87 (meaning a similarity of 87%).

[0074] In the application embodiment, error similarity is used to correct the preset positioning weight to obtain the positioning weight, and the design weight is calculated. The preset positioning weight is an initial reference value used to measure the basic reliability of "user positioning data" in route fusion or error correction. The value range is typically 0-1; a larger value indicates greater reliance on positioning data, while a smaller value indicates greater reliance on design data (such as route planning in architectural drawings). If more trust is placed in design data (such as high-precision building BIM models), the preset value can be set to 0.3-0.4; if more reliance is placed on positioning data, it can be set to 0.6-0.7.

[0075] The specific correction method can be as follows: if the preset positioning weight is 0.5 (that is, the credibility of "positioning data" and "design data" is equal) and the error similarity is 0.87, the corrected positioning weight can be: Positioning weight W1 = preset positioning weight × (1 + error similarity - 0.5) = 0.5 × (1 + 0.87 - 0.5) = 0.685.

[0076] In step S400 of this embodiment, the first route information set and the second route information set are fused to obtain a road network information set, including:

[0077] Calculate the route distance between each first route information in the first route information set and each second route information in the second route information set, and combine the first route information with the second route information with the smallest route distance to obtain multiple route information groups;

[0078] Based on the positioning weight and design weight, the route coordinates of the second route information and the first route information in each route information group are weighted and fused to obtain multiple fused route information sets as road network information sets.

[0079] In this embodiment, route distance is used to measure the spatial proximity of two routes (e.g., first route information A and second route information B). The calculation method can be the average of the distance between the starting point of route A and the starting point of route B plus the distance between the ending point of route A and the ending point of route B. Alternatively, the route can be divided into several key control points (e.g., turning points), and the average distance between the corresponding control points can be calculated. Then, for each designed route in the first route information set, all fitted routes in the second route information set are traversed, and the route distance between the two routes is calculated. The fitted route B with the smallest distance is selected as the paired route of A, forming a route information group (A, B). If the minimum distance exceeds a preset threshold (e.g., 5 meters), the designed route is considered to have no effective fitted route and is not merged (retained as an independent route).

[0080] Next, the route coordinates of the second route information and the first route information within each route information group need to be weighted and fused. For example, in the first route information A, the coordinates of point a are (Xa, Ya), representing the location extracted from the architectural drawings. In the second route information B, the coordinates of the point b closest to point a are (Xb, Yb), representing the location fitted based on the user's location data and the design weight calculated in the previous steps.

[0081] Optionally, the X-coordinate (horizontal axis position) of the merged new point is: New X = Xa × Design Weight + Xb × Positioning Weight. If Xa of design point A is 10 meters, Xb of positioning point B is 12 meters, design weight S = 0.6, and positioning weight D = 0.4, then New X = 10 × 0.6 + 12 × 0.4 = 6 + 4.8 = 10.8 meters.

[0082] Optionally, the Y-coordinate (vertical axis position) of the new point after fusion is new Y = Ya × design weight + Yb × positioning weight. If the Ya of design point A is 20 meters, the Yb of positioning point B is 21 meters, the design weight S = 0.6, and the positioning weight D = 0.4, then new Y = 20 × 0.6 + 21 × 0.4 = 12 + 8.4 = 20.4 meters.

[0083] The above calculations yield a new coordinate point (new X, new Y), namely (10.8, 20.4). Multiple new coordinate points are calculated using this method, and connecting them yields the fused route information.

[0084] If the design weight is higher (e.g., design weight = 0.8, positioning weight = 0.2), the new point will be closer to the coordinates of design point A. If the positioning weight is higher (e.g., positioning weight = 0.7, design weight = 0.3), the new point will be closer to the coordinates of positioning point B.

[0085] If the two routes have different numbers of turning points (e.g., the design route has 3 points, and the positioning route has 5 points), you can first add or delete points to make their number of points the same, and then calculate the fused coordinates point by point as described above. For example, if the design route has 3 points (start point → midpoint → end point) and the positioning route has 4 points (start point → midpoint 1 → midpoint 2 → end point), you can add a point in the middle of the design route or delete a midpoint of the positioning route to make both routes have 4 points, and then calculate the fused coordinates point by point.

[0086] The above method allows for the weighted fusion of the route coordinates of the second and first route information within each route information group, resulting in multiple fused route information sets as a road network information set. Example 2, as follows... Figure 2 As shown, based on the same inventive concept as the automatic road network generation method provided in Embodiment 1, this embodiment of the invention also provides an automatic road network generation system, including:

[0087] User location acquisition module 11 is used to acquire the design drawing data of the target building and collect user location heatmaps within the target building;

[0088] The route information acquisition module 12 is used to extract a first route information set based on the design drawing data, and extract the adjacent design data and user location information of each first route information to obtain a first adjacent design dataset and multiple first user location information sets, and to fit a second route information set based on the user location heat map.

[0089] The positioning error prediction module 13 is used to predict the positioning signal impact error of each first route information according to the first neighbor design dataset, and obtain a positioning error parameter set.

[0090] The route information fusion module 14 is used to verify the positioning error parameter set and multiple first user positioning information sets, configure the generation weights, and fuse the first route information set and the second route information set to obtain a road network information set.

[0091] Furthermore, the user location acquisition module 11 includes the following execution steps:

[0092] Obtain the design drawings of the target building;

[0093] Collect user location information within the target building within a preset time range and generate a user location heat map, where each user location information includes location coordinates.

[0094] Furthermore, the route information acquisition module 12 includes the following execution steps:

[0095] Based on the design drawing data, the first route information set is extracted;

[0096] Within the design drawing data, extract the neighboring design data within a preset distance for each first route information to obtain the first neighboring design dataset;

[0097] Within the first user location heatmap, user location information within a preset distance of each first route information is extracted to obtain multiple sets of first user location information.

[0098] Cluster the location information of users whose location hotspots are less than the preset route distance to obtain a location information clustering result set;

[0099] Within each location information clustering result, second route information is randomly fitted, and the average distance between all user location information and second route information within the location information clustering result is calculated to obtain the fitted route distance.

[0100] Within each location information clustering result, the fitting optimization of the second route information is performed to obtain multiple second route information with the smallest fitting route distance, thus obtaining the second route information set.

[0101] Furthermore, the positioning error prediction module 13 includes the following execution steps:

[0102] The positioning error impact analyzer is invoked, wherein the positioning error impact analyzer is obtained by machine learning training, and the training data includes a sample nearest neighbor design data set and a sample positioning error parameter set;

[0103] Each first neighbor design data in the first neighbor design dataset is input into the positioning error impact analyzer, and the positioning error parameter set is obtained by identifying the output, wherein each positioning error parameter includes the error distance of the user positioning information.

[0104] Furthermore, the route information fusion module 14 includes the following execution steps:

[0105] Based on the multiple first user positioning information sets, calculate the average distance between each first user positioning information set and the corresponding first route information to obtain multiple actual positioning error parameters;

[0106] Calculate the similarity between multiple positioning error parameters in the positioning error parameter set and multiple actual positioning error parameters to obtain the error similarity.

[0107] The preset positioning weights are corrected and calculated using error similarity to obtain the positioning weights, and the design weights are then calculated.

[0108] Calculate the route distance between each first route information in the first route information set and each second route information in the second route information set, and combine the first route information with the second route information with the smallest route distance to obtain multiple route information groups;

[0109] Based on the positioning weight and design weight, the route coordinates of the second route information and the first route information in each route information group are weighted and fused to obtain multiple fused route information sets as road network information sets.

[0110] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0111] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0112] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0113] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0114] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0115] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.

[0116] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for automatic road network generation, characterized in that, The method includes: Obtain the design drawings of the target building and collect user location heatmaps within the target building; Based on the design drawing data, a first route information set is extracted, and the neighboring design data and user location information of each first route information are extracted to obtain a first neighboring design dataset and multiple first user location information sets. Based on the user location heat map, a second route information set is obtained by fitting. Based on the first neighbor design dataset, the positioning signal impact error prediction is performed for each first route information to obtain a positioning error parameter set; The positioning error parameter set and multiple first user positioning information sets are verified and weights are configured. The first route information set and the second route information set are fused to obtain a road network information set. The configuration of the generation weights includes configuring positioning weights and design weights, including: Based on the multiple first user positioning information sets, calculate the average distance between each first user positioning information set and the corresponding first route information to obtain multiple actual positioning error parameters; Calculate the similarity between multiple positioning error parameters in the positioning error parameter set and multiple actual positioning error parameters to obtain the error similarity. The preset positioning weights are corrected using error similarity calculations to obtain the positioning weights, and the design weights are then calculated. Among them, location weight measures the credibility of user location data in route fusion, and design weight measures the credibility of design data in route fusion.

2. The automatic road network generation method of claim 1, wherein, Obtain the design drawings of the target building and collect user location heatmaps within the target building, including: Obtain the design drawings of the target building; Collect user location information within the target building within a preset time range and generate a user location heat map, where each user location information includes location coordinates.

3. The automatic road network generation method of claim 1, wherein, Based on the design drawing data, a first route information set is extracted, and adjacent design data and user location information for each first route information are extracted to obtain a first adjacent design dataset and multiple first user location information sets, including: Based on the design drawing data, the first route information set is extracted; Within the design drawing data, extract the neighboring design data within a preset distance for each first route information to obtain the first neighboring design dataset; Within the first user location heatmap, user location information within a preset distance of each first route information is extracted to obtain multiple sets of first user location information.

4. The automatic road network generation method of claim 1, wherein, Based on the user location heatmap, a second route information set is obtained by fitting, including: Cluster the location information of users whose location hotspots are less than the preset route distance to obtain a location information clustering result set; Within each location information clustering result, second route information is randomly fitted, and the average distance between all user location information and second route information within the location information clustering result is calculated to obtain the fitted route distance. Within each location information clustering result, the fitting optimization of the second route information is performed to obtain multiple second route information with the smallest fitting route distance, thus obtaining the second route information set.

5. The automatic road network generation method of claim 1, wherein, Based on the first neighbor design dataset, the positioning signal impact error is predicted for each first route information to obtain a positioning error parameter set, including: The positioning error impact analyzer is invoked, wherein the positioning error impact analyzer is obtained by machine learning training, and the training data includes a sample nearest neighbor design data set and a sample positioning error parameter set; Each first neighbor design data in the first neighbor design dataset is input into the positioning error impact analyzer, and the positioning error parameter set is obtained by identifying the output, wherein each positioning error parameter includes the error distance of the user positioning information.

6. The automatic road network generation method of claim 1, wherein, The first route information set and the second route information set are fused to obtain a road network information set, including: Calculate the route distance between each first route information in the first route information set and each second route information in the second route information set, and combine the first route information with the second route information with the smallest route distance to obtain multiple route information groups; Based on the positioning weight and design weight, the route coordinates of the second route information and the first route information in each route information group are weighted and fused to obtain multiple fused route information sets as road network information sets.

7. An automatic road network generation system characterized by comprising: The system includes: The user location acquisition module is used to acquire the design drawing data of the target building and collect the user location heat map within the target building; The route information acquisition module is used to extract a first route information set based on the design drawing data, and extract the neighboring design data and user location information of each first route information to obtain a first neighboring design dataset and multiple first user location information sets, and to fit a second route information set based on the user location heat map. The positioning error prediction module is used to predict the positioning signal impact error of each first route information according to the first neighbor design dataset, and obtain the positioning error parameter set. The route information fusion module is used to verify the positioning error parameter set and multiple first user positioning information sets, configure generation weights, and fuse the first route information set and the second route information set to obtain a road network information set. The configuration of generation weights includes configuring positioning weights and design weights, including: Based on the multiple first user positioning information sets, calculate the average distance between each first user positioning information set and the corresponding first route information to obtain multiple actual positioning error parameters; Calculate the similarity between multiple positioning error parameters in the positioning error parameter set and multiple actual positioning error parameters to obtain the error similarity. The preset positioning weights are corrected using error similarity calculations to obtain the positioning weights, and the design weights are then calculated. Among them, location weight measures the credibility of user location data in route fusion, and design weight measures the credibility of design data in route fusion.

Images