Indoor simulation three-dimensional model optimal point path planning method and system

Abstract

The invention belongs to the technical field of path planning, and particularly relates to an indoor simulation three-dimensional model optimal point path planning method and system, and the method comprises the following steps: A1, analyzing a DWG / DXF format drawing, extracting vector data containing engineering coordinates and attribute information, and generating a two-dimensional / three-dimensional vector file in an SHP / SVG format; a2, carrying out thresholding segmentation on the grid plane graph, extracting an edge contour and an image coordinate of a non-connected region, and establishing a materialized plane partition; a3, based on an engineering measurement coordinate system and a proportional scale, unifying the vector, the entity and the three-dimensional data to the same coordinate system, and completing coordinate conversion through a monomerized building reference point; a4, combining the two-dimensional vector skeleton line with a set height by using a PyVista library, constructing a three-dimensional white film structure, adding textures, and generating an indoor three-dimensional model; according to the method, the adaptability and flexibility of path planning are improved in a matched mode, and it is ensured that the planned path is feasible and optimal in different indoor environments.

Description

Technical Field

[0001] The present invention belongs to the technical field of path planning, and in particular to a method and system for optimal path planning of an indoor simulated three-dimensional model. Background Art

[0002] With the acceleration of urbanization and the continuous expansion of building scale, indoor spaces are becoming more and more complex, such as large shopping malls, hospitals, airports, etc. These indoor buildings usually contain numerous floors, rooms, passages and obstacles, which bring challenges to people's passage and navigation. Traditional indoor path planning methods are mainly based on static maps and fixed path algorithms. For example, in early indoor navigation systems, they often simply use Dijkstra algorithm or A[N] ... * Algorithms are used to find the shortest path between two points. This method meets basic navigation needs to a certain extent, but it has obvious limitations when facing complex and changing indoor environments.

[0003] Traditional indoor path planning methods are mostly based on static map information, that is, the location of obstacles, traffic areas and other information in the map are fixed. However, in actual indoor environments, the distribution of obstacles may change at any time, such as temporary promotional booths in shopping malls and mobile medical equipment in hospitals. This static path planning cannot adapt to environmental changes in real time and may cause the planned path to pass through obstacles, making the path infeasible and reducing the accuracy and reliability of navigation.

[0004] Existing path planning algorithms typically only consider a single objective, such as the shortest distance. In indoor environments, simply pursuing the shortest distance may not be the optimal choice. For example, in crowded places, the shortest path may pass through congested passages, resulting in increased walking time; or the shortest path may contain many turning points, increasing the difficulty and time cost of walking. In addition, there may be some high-risk areas in indoor environments, such as areas under construction and areas with electrical equipment failures. Traditional path planning algorithms cannot automatically avoid these high-risk areas, posing a potential threat to pedestrian safety.

[0005] To this end, the present invention provides a method and system for optimal path planning of an indoor simulated three-dimensional model. Summary of the Invention

[0006] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0007] The technical solution adopted by the present invention to solve the technical problem is: a method for optimal path planning of an indoor simulated three-dimensional model according to the present invention comprises the following steps:

[0008] A1. Parse DWG / DXF format drawings, extract vector data containing engineering coordinates and attribute information, and generate 2D / 3D vector files in SHP / SVG format;

[0009] A2. Threshold segmentation is performed on the grid plane image to extract the edge contours and image coordinates of the non-connected areas and establish the substantive plane partitions;

[0010] A3. Based on the engineering survey coordinate system and scale, vector, solid, and 3D data are unified into the same coordinate system, and coordinate conversion is completed through individual building reference points;

[0011] A4. Use the PyVista library to combine the 2D vector skeleton line with the set height to construct a 3D white membrane structure and add texture to generate a 3D indoor model.

[0012] A5. Based on the starting and ending points input by the user, the optimal path is calculated in the planar grid map using the Dijkstra algorithm and converted into 2D coordinates for visualization.

[0013] Preferably, the implementation of Dijkstra's algorithm includes:

[0014] B1. Configure the parameters csgraph, directed, and indices through the Dijkstra function of the Scipy library to dynamically adjust the path weight;

[0015] B2. Update the coordinates of the starting and ending points in real time based on user interaction events, and optimize the vertical projection accuracy of the path by combining the 3D model height information.

[0016] Preferably, in the coordinate matching conversion step:

[0017] C1. Generate a multi-level coordinate mapping relationship by superimposing scale information and image icons;

[0018] C2. Use the reference point coordinates to dynamically correct the path planning results to ensure the topological consistency of the path and the 3D model.

[0019] Preferably, the three-dimensional reconstruction and visualization step specifically includes:

[0020] D1. Associate the path planning results with the door and window entities in the 3D model to dynamically avoid obstacle areas;

[0021] D2. Adjust the model height parameters based on user interaction and update the Z-axis projection of the path in real time.

[0022] A system for planning an optimal path for an indoor physical building, which adopts the above-mentioned method for planning an optimal path for an indoor physical building, is characterized by comprising:

[0023] Vector parsing module: used to parse drawings and output standardized vector files;

[0024] Entity segmentation module: extracts plane entity areas and coordinates through image segmentation technology;

[0025] Coordinate transformation engine: Integrates engineering coordinate systems, scales, and reference points to unify coordinates of multi-source data;

[0026] 3D modeling unit: Generates textured 3D models based on the PyVista library and dynamically binds them to path planning results;

[0027] Path optimizer: uses Dijkstra algorithm to calculate the optimal path and supports real-time interactive adjustment;

[0028] Visual interface: Overlays the path and 3D model, and provides multi-view rendering and parameter configuration functions;

[0029] Integrated satellite modeling: Integrate OSGB / OBJ format 3D models generated by satellite imagery with indoor models to expand the global coverage of path planning and optimize path continuity in indoor and outdoor connection areas using satellite point cloud data.

[0030] Preferably, the path optimizer includes: a dynamic weight allocation module and a multi-objective optimization algorithm;

[0031] Dynamic weight allocation module: adjusts the path weight matrix based on the obstacle density and traffic area priority in the 3D model;

[0032] Multi-objective optimization algorithm: supports multi-condition path planning such as shortest distance, least turning points, and avoiding high-risk areas.

[0033] Preferably, the calculation method of the dynamic weight allocation module is:

[0034] Assume there are n nodes in the three-dimensional model, then:

[0035] Path weight matrix W = [w ij ], where w ij represents the initial weight from node i to node j;

[0036] Obstacle density matrix O = [o ij ], where o ij represents the obstacle density on the path from node i to node j;

[0037] Passage area priority matrix P = [p ij ], where p ij Indicates the priority of the passage area from node i to node j;

[0038] Define the adjustment coefficient function f(o) of obstacle density to weight ij );

[0039] Using the linear function f(o ij )=1+k2·o ij , where k1 is a constant and k1>0, which is used to control the influence of obstacle density on the weight;

[0040] Define the adjustment coefficient function g(p ij );

[0041] Using inverse proportional function Wherein, k2 is a constant, and k2>0, which is used to control the influence of the priority of the pass area on the weight;

[0042] Dynamic weight matrix update:

[0043] The updated dynamic weight matrix W'=[w' ij ], where w' ij =w ij ·f(o ij )·g(p ij ).

[0044] Preferably, the multi-objective optimization algorithm is calculated as follows:

[0045] Let the starting point be s, the end point be t, and the path r be a node sequence {r0=s,r1,…,r m =t};

[0046] The shortest distance objective function d(r): The sum of the dynamic weights between all adjacent nodes on path r is

[0047] Minimum inflection point objective function c(r): The number of direction changes on path r. The inflection point can be determined by calculating the vector angle between three adjacent nodes;

[0048] Avoid high-risk area objective function h(r): If the path r passes through a high-risk area node, a larger penalty value is assigned; if not, the value is 0;

[0049] The multi-objective problem is transformed into a single-objective problem by weighting. Let the weight of d(r) be α, the weight of c(r) be β, the weight of h(r) be γ, and α+β+γ=1, then the comprehensive objective function is:

[0050] F(r)=α·d(r)+β·c(r)+γ·h(r).

[0051] Preferably, the coordinate transformation engine achieves high-precision matching in the following manner:

[0052] Introducing laser point cloud data as an auxiliary benchmark to calibrate the coordinate deviation between vector data and 3D models;

[0053] During the path planning process, coordinate consistency is verified in real time to trigger the automatic correction mechanism.

[0054] Preferably, the visualization interface supports path heat map display and AR / VR interaction mode;

[0055] Path heat map display: Generate heat distribution of high-frequency travel areas based on historical path data;

[0056] AR / VR interaction mode: Immersive path preview and editing are achieved through virtual reality devices.

[0057] The beneficial effects of the present invention are as follows:

[0058] 1. The present invention uses a dynamic weight allocation module to adjust the path weight matrix based on the obstacle density and traffic area priority in the 3D model. Based on real-time information such as obstacle density and traffic area priority in the 3D model, the path weight matrix is dynamically adjusted. In indoor environments, by monitoring the distribution of obstacles in real time, the path weights corresponding to areas with high obstacle density are increased, thereby guiding the path planning algorithm to avoid these areas. At the same time, the path weights corresponding to traffic areas with higher priorities are reduced, allowing these areas to be preferentially selected during path planning. This improves the adaptability and flexibility of path planning, ensuring that the planned paths are feasible and optimal in different indoor environments.

[0059] 2. The present invention uses a multi-objective optimization algorithm and supports multi-condition path planning with the shortest distance, the least turning points, and avoiding high-risk areas. In indoor path planning, comprehensive consideration of multiple objectives can more comprehensively meet user needs. When planning the path from the mall entrance to a specific store, not only the shortest distance should be considered, but also the turning points should be minimized to improve walking convenience. At the same time, high-risk areas under construction should be avoided to ensure the safety of pedestrians. Through the multi-objective optimization algorithm, trade-offs and optimizations can be made between multiple objectives to find the optimal path that best meets user needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] The present invention will be further described below with reference to the accompanying drawings.

[0061] Figure 1 This is a flow chart of the optimal path planning method of the present invention;

[0062] Figure 2 It is a flow chart of the implementation steps of the Dijkstra algorithm in the present invention;

[0063] Figure 3is a flow chart of the coordinate matching conversion steps in the present invention;

[0064] Figure 4 is a flow chart of the three-dimensional reconstruction visualization steps in the present invention;

[0065] Figure 5 This is a framework diagram of the optimal path planning system in the present invention;

[0066] Figure 6 This is a framework diagram for laser point cloud production in the present invention. DETAILED DESCRIPTION

[0067] In order to make the technical means, creative features, objectives and effects achieved by the present invention easier to understand, the present invention is further described below in conjunction with specific implementation methods.

[0068] like Figures 1 to 4 As shown, the optimal path planning method for an indoor simulation three-dimensional model according to the present invention includes the following steps:

[0069] A1. Parse DWG / DXF format drawings, extract vector data containing engineering coordinates and attribute information, and generate 2D / 3D vector files in SHP / SVG format;

[0070] By deeply parsing and processing drawing format files such as dwg / dxf, users can obtain the specific location and attribute information of each layer and component they need, that is, layer the entire structure. Once the parsing is completed, users can manually select the specific points, lines, and surface vectors they need for export according to their specific needs.

[0071] It should be noted that the exported vectors not only contain the engineering coordinates, but also the attribute information of these coordinates, which provides an important basis for subsequent work. After processing, this information will be restored or converted into a two-dimensional vector file or information that can construct a three-dimensional image. This file or information is usually stored and presented in the shp or svg format.

[0072] A2. Perform threshold segmentation on the grid plane image, extract the edge contours and image coordinates of the non-connected areas, and establish the materialized plane partitions.

[0073] This process is to identify and process the exported drawing plane raster images, that is, images in png or jpg format. The specific processing method includes segmenting or detecting the image, and then using corresponding image processing techniques such as threshold segmentation to extract non-connected areas. These areas are usually independent of each other and have different shapes, sizes and characteristics.

[0074] During the extraction process, it is also necessary to extract the edge contours of the area and record their corresponding image coordinates for subsequent processing operations.

[0075] After the extraction and recording is completed, the system will output the corresponding position and edge information of each entity or area. In this way, each entity or area in the image will have a corresponding position mark and edge description, which greatly facilitates the understanding and analysis of the image content.

[0076] By implementing the above steps, the zoning and classification of the image plane is achieved. In other words, this process is like measuring and marking a piece of unclassified land in detail, and finally dividing the land into regions, each of which has a clear boundary and coordinate position.

[0077] A3. Based on the engineering survey coordinate system and scale, vector, entity and 3D data are unified into the same coordinate system, and coordinate conversion is completed through individual building reference points.

[0078] In the process of modeling and rendering, it is very important to maintain the coordinate consistency between vectors, entities, and three-dimensional data generated under different paths. To achieve this goal, the coordinate transformation relationship between various parameters must be constructed.

[0079] First, obtain the original data and scale information of each part by obtaining the engineering measurement coordinate system of the drawing itself. This is the basic work because all transformations must be based on this coordinate system.

[0080] After obtaining the coordinate system information, the relevant scale information is retained. The scale information is an important reference in the conversion process. It can ensure that different vectors and entities still maintain their original proportional relationship after conversion, avoiding the generation of errors.

[0081] Combining this basic information with image icons, which are actually visual representations, allows people to more intuitively understand the relative positions and proportional relationships of various parts.

[0082] Finally, by referencing the reference point coordinates of the individual buildings, we can convert the engineering coordinates of each entity. This step is the key to the entire conversion process, ensuring that all entities still maintain the correct position and proportion after conversion.

[0083] Through the above steps, the coordinate consistency between the vectors, entities, and three-dimensional data generated under different paths can be effectively maintained, thereby ensuring the accuracy of modeling and rendering.

[0084] A4. Use the PyVista library to combine the two-dimensional vector skeleton line with the set height to construct a three-dimensional white membrane structure and add texture to generate a three-dimensional indoor model.

[0085] The vectors obtained by analysis are used for three-dimensional reconstruction in order to generate a white membrane structure that can represent the skeleton of the wall. The white membrane structure can provide a three-dimensional sense for the model and also facilitate subsequent operations.

[0086] Next, the previously extracted two-dimensional entity is combined with the white membrane structure to generate a three-dimensional surface. In this process, corresponding attributes or textures are added to the model in three-dimensional space as needed. For example, architectural elements such as windows and doors can be added according to actual conditions, or corresponding colors and textures can be added to the walls, realizing the transition from two-dimensional to three-dimensional and completely restoring the three-dimensional entity.

[0087] Finally, with the help of the PyVista library, based on the two-dimensional vector skeleton line and combined with the set height, a three-dimensional wall is formed in the Z direction. At the same time, combined with the two-dimensional solid surface, a complete three-dimensional surface is constructed. In this process, the combination of the two-dimensional vector skeleton line and the set height forms a three-dimensional wall in the Z direction, which undoubtedly makes the three-dimensional model look more realistic and more in line with the actual situation.

[0088] A5. Based on the starting and ending points input by the user, the optimal path is calculated in the planar grid map using the Dijkstra algorithm and converted into 2D coordinates for visualization.

[0089] Based on step A3, a planar grid map is calculated and generated. Mouse interaction events are monitored to obtain the user-entered starting point P1 and end point P2, and coordinate conversion is performed. The converted results are passed to the scipy.sparse.csgraph.dijkstra library to calculate the optimal path.

[0090] The implementation of Dijkstra's algorithm includes:

[0091] B1. Configure the parameters csgraph, directed, and indices through the Dijkstra function of the Scipy library to dynamically adjust the path weight;

[0092] B2. Update the coordinates of the starting and ending points in real time based on user interaction events, and optimize the vertical projection accuracy of the path by combining the 3D model height information.

[0093] In the coordinate matching transformation step:

[0094] C1. Generate a multi-level coordinate mapping relationship by superimposing scale information and image icons;

[0095] C2. Use the reference point coordinates to dynamically correct the path planning results to ensure the topological consistency of the path and the 3D model.

[0096] The 3D reconstruction visualization steps specifically include:

[0097] D1. Associate the path planning results with the door and window entities in the 3D model to dynamically avoid obstacle areas;

[0098] D2. Adjust the model height parameters based on user interaction and update the Z-axis projection of the path in real time.

[0099] like Figure 5 As shown, a system for optimal path planning of an indoor simulated three-dimensional model is provided. The system adopts the above-mentioned optimal path planning method of the indoor simulated three-dimensional model, and includes:

[0100] Vector parsing module: used to parse drawings and output standardized vector files;

[0101] Entity segmentation module: extracts plane entity areas and coordinates through image segmentation technology;

[0102] Coordinate transformation engine: Integrates engineering coordinate systems, scales, and reference points to unify coordinates of multi-source data.

[0103] The coordinate transformation engine achieves high-precision matching in the following ways:

[0104] Introducing laser point cloud data as an auxiliary benchmark to calibrate the coordinate deviation between vector data and 3D models;

[0105] During the path planning process, coordinate consistency is verified in real time to trigger the automatic correction mechanism.

[0106] like Figure 6 As shown, high-precision point clouds and panoramic data of the ground and basement spaces of buildings (such as subway stations) are obtained through laser scanning, and the spatial locations of buildings (such as subway stations), entrances, passages to stairs, ticket halls, security checks, and waiting rooms are obtained through point clouds.

[0107] The location, size, and texture of shops, ticket machines, security equipment, security booths, waste bins, fire hydrants (above and below ground), cameras, equipment cabinets, police stations (inside subway stations), explosion-proof tanks, ATMs, and other facilities and equipment are obtained through point clouds, panoramic images, and field photos to build a 3D model.

[0108] 3D modeling unit: Generates textured 3D models based on the PyVista library and dynamically binds them to path planning results;

[0109] Path Optimizer: Uses Dijkstra algorithm to calculate the optimal path and supports real-time interactive adjustment.

[0110] Real-time perception of changes in obstacle density in three-dimensional scenes and dynamic adjustment of path weights avoids the failure of traditional static weights in complex environments.

[0111] Supports priority division of traffic areas (such as emergency channels) to improve task execution efficiency;

[0112] The path optimizer includes: a dynamic weight allocation module and a multi-objective optimization algorithm;

[0113] Dynamic weight allocation module: adjusts the path weight matrix based on the obstacle density and traffic area priority in the 3D model;

[0114] The calculation method of the dynamic weight allocation module is:

[0115] Assume that there are n nodes in the three-dimensional model, then:

[0116] Path weight matrix W = [w ij ], where w ij represents the initial weight from node i to node j;

[0117] Obstacle density matrix O = [o ij ], where o ij represents the obstacle density on the path from node i to node j;

[0118] Passage area priority matrix P = [p ij ], where p ij Indicates the priority of the passage area from node i to node j;

[0119] Define the adjustment coefficient function f(o) of obstacle density to weight ij );

[0120] Using the linear function f(o ij )=1+k2·o ij , where k1 is a constant and k1>0, which is used to control the influence of obstacle density on the weight;

[0121] Define the adjustment coefficient function g(p ij );

[0122] Using inverse proportional function Wherein, k2 is a constant, and k2>0, which is used to control the influence of the priority of the pass area on the weight;

[0123] Dynamic weight matrix update:

[0124] The updated dynamic weight matrix W'=[w'' ij ], where w' ij =w ij ·f(o ij )·g(p ij ).

[0125] Preferably, the code is:

[0126]

[0127] Multi-objective optimization algorithm: supports multi-condition path planning such as shortest distance, least turning points, and avoiding high-risk areas.

[0128] Multi-objective optimization algorithm calculation method:

[0129] Let the starting point be s, the end point be t, and the path r be a node sequence {r0=s,r1,…,r m =t};

[0130] The shortest distance objective function d(r): The sum of the dynamic weights between all adjacent nodes on path r is

[0131] Minimum inflection point objective function c(r): The number of direction changes on path r. The inflection point can be determined by calculating the vector angle between three adjacent nodes.

[0132] Avoid high-risk area objective function h(r): If the path r passes through a high-risk area node, a larger penalty value is assigned; if not, the value is 0;

[0133] The multi-objective problem is transformed into a single-objective problem by weighting. Let the weight of d(r) be α, the weight of c(r) be β, the weight of h(r) be γ, and α+β+γ=1, then the comprehensive objective function is:

[0134] F(r)=α·d(r)+β·c(r)+γ·h(r).

[0135] In this embodiment, A * The algorithm performs path search. During the search process, the cost estimation function of the node is:

[0136] f(n)=g(n)+h(n), where g(n) is the actual cost from the starting point to node n, and h(n) is the estimated cost from node n to the end point, calculated using the comprehensive objective function F(r).

[0137] Preferably, simplify A * The algorithm code is:

[0138]

[0139]

[0140]

[0141] Visual interface: Overlays the path and 3D model, and provides multi-view rendering and parameter configuration functions.

[0142] The visualization interface supports path heat map display and AR / VR interaction mode;

[0143] Path heat map display: Generate heat distribution of high-frequency travel areas based on historical path data;

[0144] AR / VR interaction mode: Immersive path preview and editing are achieved through virtual reality devices.

[0145] Integrated satellite modeling: Integrate OSGB / OBJ format 3D models generated by satellite imagery with indoor models to expand the global coverage of path planning and optimize path continuity in indoor and outdoor connection areas using satellite point cloud data.

[0146] Aerial triangulation was performed using multi-view, multi-angle satellite imagery from the same orbit and at a resolution of 0.5 meters. The resulting point cloud was exported for tilt modeling.

[0147] According to the actual modeling effect, ground image data is re-shot for locations where there are model holes and missing parts. The model is modified and manually reconstructed based on the re-shot images, and finally the 3D result data in OSGB and OBJ formats is exported.

[0148] The above-mentioned front, back, left, right, up and down are all based on the Figure 1 As a benchmark, according to the person's observation perspective, the side of the device facing the observer is defined as the front, the left side of the observer is defined as the left, and so on.

[0149] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the scope of protection of the present invention.

[0150] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the foregoing embodiments. The foregoing embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for optimal path planning for an indoor simulated three-dimensional model, characterized by: The following steps are involved: A1. Parse DWG / DXF format drawings, extract vector data containing engineering coordinates and attribute information, and generate 2D / 3D vector files in SHP / SVG format; A2. Threshold segmentation is performed on the grid plane image to extract the edge contours and image coordinates of the non-connected areas and establish the substantive plane partitions; A3. Based on the engineering survey coordinate system and scale, vector, solid, and 3D data are unified into the same coordinate system, and coordinate conversion is completed through individual building reference points; A4. Use the PyVista library to combine the 2D vector skeleton line with the set height to construct a 3D white membrane structure and add texture to generate a 3D indoor model. A5. Based on the starting and ending points input by the user, the optimal path is calculated in the planar grid map using the Dijkstra algorithm and converted into 2D coordinates for visualization.

2. The method for optimal path planning for an indoor three-dimensional simulation model according to claim 1, characterized in that: The implementation of the Dijkstra algorithm includes: B1. Configure the parameters csgraph, directed, and indices through the Dijkstra function of the Scipy library to dynamically adjust the path weight; B2. Update the coordinates of the starting and ending points in real time based on user interaction events, and optimize the vertical projection accuracy of the path by combining the 3D model height information.

3. The method for optimal path planning for an indoor three-dimensional simulation model according to claim 2, characterized in that: In the coordinate matching conversion step: C1. Generate a multi-level coordinate mapping relationship by superimposing scale information and image icons; C2. Use the reference point coordinates to dynamically correct the path planning results to ensure the topological consistency of the path and the 3D model.

4. The method for optimal path planning for an indoor three-dimensional simulation model according to claim 3, characterized in that: The three-dimensional reconstruction and visualization step specifically includes: D1. Associate the path planning results with the door and window entities in the 3D model to dynamically avoid obstacle areas; D2. Adjust the model height parameters based on user interaction and update the Z-axis projection of the path in real time.

5. A system for optimal path planning of an indoor simulated three-dimensional model, the system adopting the optimal path planning method of an indoor simulated three-dimensional model according to claim 4, characterized in that: include: Vector parsing module: used to parse drawings and output standardized vector files; Entity segmentation module: extracts plane entity areas and coordinates through image segmentation technology; Coordinate transformation engine: Integrates engineering coordinate systems, scales, and reference points to unify coordinates of multi-source data; 3D modeling unit: Generates textured 3D models based on the PyVista library and dynamically binds them to path planning results; Path optimizer: uses Dijkstra algorithm to calculate the optimal path and supports real-time interactive adjustment; Visual interface: Overlays the path and 3D model, and provides multi-view rendering and parameter configuration functions; Integrated satellite modeling: Integrate OSGB / OBJ format 3D models generated by satellite imagery with indoor models to expand the global coverage of path planning and optimize path continuity in indoor and outdoor connection areas using satellite point cloud data.

6. The optimal path planning system for indoor simulation three-dimensional models according to claim 5, characterized in that: The path optimizer includes: a dynamic weight allocation module and a multi-objective optimization algorithm; Dynamic weight allocation module: adjusts the path weight matrix based on the obstacle density and traffic area priority in the 3D model; Multi-objective optimization algorithm: supports multi-condition path planning such as shortest distance, least turning points, and avoiding high-risk areas.

7. The indoor simulation three-dimensional model optimal path planning system according to claim 6, characterized in that: The calculation method of the dynamic weight distribution module is: Assume there are n nodes in the three-dimensional model, then: Path weight matrix W = [w ij ], where w ij represents the initial weight from node i to node j; Obstacle density matrix O = [o ij ], where o ij represents the obstacle density on the path from node i to node j; Passage area priority matrix P = [p ij ], where p ij Indicates the priority of the passage area from node i to node k; Define the adjustment coefficient function f(o) of obstacle density to weight ij ); Using the linear function f(o ij )=1+k2·o ij , where k1 is a constant and k1>0, which is used to control the influence of obstacle density on the weight; Define the adjustment coefficient function g(p ij ); Using inverse proportional function Wherein, k2 is a constant, and k2>0, which is used to control the influence of the priority of the pass area on the weight; Dynamic weight matrix update: The updated dynamic weight matrix W'=[w' ij ], where w' ij =w ij ·f(o ij )·g(p ij ).

8. The indoor simulation three-dimensional model optimal path planning system according to claim 6, characterized in that: Multi-objective optimization algorithm calculation method: Let the starting point be s, the end point be t, and the path r be a node sequence {r0=s,r1,…,r m =t}; The shortest distance objective function d(r): The sum of the dynamic weights between all adjacent nodes on path r is Minimum inflection point objective function c(r): the number of direction changes on path r; Avoid high-risk area objective function h(r): If the path r passes through a high-risk area node, a larger penalty value is assigned; if not, the value is 0; The multi-objective problem is transformed into a single-objective problem by weighting. Let the weight of d(r) be α, the weight of c(r) be β, the weight of h(r) be γ, and α+β+γ=1, then the comprehensive objective function is: F(r)=α·d(r)+β·c(r)+γ·h(r).

9. The indoor simulation three-dimensional model optimal path planning system according to claim 6, characterized in that: The coordinate conversion engine achieves high-precision matching in the following way: Introducing laser point cloud data as an auxiliary benchmark to calibrate the coordinate deviation between vector data and 3D models; During the path planning process, coordinate consistency is verified in real time to trigger the automatic correction mechanism.

10. The indoor simulation three-dimensional model optimal path planning system according to claim 7, characterized in that: The visualization interface supports path heat map display and AR / VR interaction mode; Path heat map display: Generate heat distribution of high-frequency travel areas based on historical path data; AR / VR interaction mode: Immersive path preview and editing are achieved through virtual reality devices.

Images