An RTK positioning and panoramic image-based accident investigation mapping system and method
By combining RTK positioning with panoramic imagery, traffic accident scene maps can be automatically identified and drawn, solving the problem of low efficiency in traditional surveying and achieving high-precision, standardized map generation and data reliability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SUZHOU WEISHITONG INTELLIGENT TECH CO LTD
- Filing Date
- 2026-02-28
- Publication Date
- 2026-05-29
AI Technical Summary
In current traffic accident scene investigations, traditional manual measurement and drawing methods result in low work efficiency, incomplete retention of on-site spatial information, inability to quickly generate standardized maps, and risks of human reading deviations and data entry errors.
The system uses an RTK high-precision positioning module to acquire the absolute spatial coordinates of key physical evidence at the accident scene in real time. Combined with a panoramic image acquisition module and an AI target attribute recognition unit, it automatically analyzes the objects in the images and defines them as category labels. Using a local national standard symbol resource library and a high-performance fusion drawing module, it generates a standardized vector scene map.
It enables the rapid and accurate generation of standardized traffic accident scene diagrams, ensuring data traceability and location accuracy, reducing manual operations, improving survey efficiency and map accuracy, and supporting instant export and printing.
Smart Images
Figure CN122116302A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of accident investigation technology, and in particular to an accident investigation mapping system and method based on RTK positioning and panoramic imagery. Background Technology
[0002] Traffic accident scene investigation is crucial for traffic accident identification and reconstruction analysis. Currently, most traffic accident scene investigations, both domestically and internationally, rely on traditional methods such as experience-based judgment, tape measure measurement, and manual drawing, which are prone to problems such as information omissions and judgment errors.
[0003] Currently, in solving traffic accident scene investigations, traditional manual measurement and drawing are used, resulting in low work efficiency, incomplete retention of on-site spatial information, and inability to quickly generate standardized maps. On-site lighting, traffic flow, and inclement weather often lead to deviations in manual readings, and manual paper records need to be entered into the computer later, which poses a risk of human input errors. Summary of the Invention
[0004] This invention discloses an accident investigation mapping system and method based on RTK positioning and panoramic imagery, aiming to solve the technical problems mentioned in the background art, such as low work efficiency, incomplete retention of on-site spatial information, and inability to quickly generate standardized maps, which are currently used in solving traffic accident scene investigations by traditional manual measurement and manual drawing.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: An accident investigation mapping system based on RTK positioning and panoramic imagery includes an RTK high-precision spatial positioning module, a panoramic image acquisition module, an AI target attribute recognition unit, a local national standard symbol resource library, a high-performance fusion drawing module, and a standardized vector scene map. The RTK high-precision spatial positioning module acquires the absolute spatial coordinates of key physical evidence at the accident scene in real time and forms the physical framework of the entire system. The panoramic image acquisition module captures panoramic images of the scene, serving as the raw material for AI recognition, ensuring that the mapping process has a basis and verifiable data. The AI target attribute recognition unit automatically analyzes objects in the images, defining them as specific category labels. Furthermore, the AI target attribute recognition unit no longer requires complex pose calculations; it only needs to confirm... This is a specific target. The local national standard symbol resource library stores various vector icons that conform to national safety industry standards. Each identified "object label" can find the corresponding "legally customized graphic symbol" here. The high-performance fusion drawing module receives the coordinate points of S1, the category label of T1, and the standard icon of T2. The high-performance fusion drawing module accurately places the "standard icon" on the "RTK coordinates" and automatically adjusts the size according to the preset global scale to complete the final rendering of the graphic. The standardized vector scene diagram automatically generates a traffic accident scene diagram containing the real scale, absolute coordinates, and standard symbols. The traffic accident scene diagram is in vector format and can be directly exported, wirelessly transmitted, or printed on-site as core physical evidence for determining liability in the accident.
[0006] By introducing RTK high-precision positioning assistance, a three-dimensional geographic coordinate network can be quickly established at the accident site. Investigators use sampling terminals to map key elements such as vehicle wheel tracks, damaged stress points, ground scratches, and debris boundaries. This process transforms complex physical trajectories into a coordinate skeleton with centimeter-level precision, ensuring the traceability and absolute positional accuracy of the collected spatial data.
[0007] In a preferred solution, the RTK high-precision spatial positioning module provides centimeter-level absolute coordinates, which is the only physical basis to replace traditional tape measure measurements. The field map in the local national standard symbol resource library must conform to the public security industry standards. The high-performance fusion drawing module is responsible for logically binding the scattered coordinate points with abstract national standard symbols. Without this engine, the collected data are just isolated points or maps, and it is impossible to synthesize a complete vector map.
[0008] An accident investigation mapping method based on RTK positioning and panoramic imagery specifically includes the following steps: Step 1: Synchronous Data Acquisition at the Site: Survey personnel use mobile data acquisition terminals to obtain the CM-level coordinates of key evidence points or center lines within seconds; Step 2: Panoramic Image Capture and Real-time Top-Down Reconstruction: The system utilizes edge computing capabilities to instantly convert panoramic images into a top-down view from a "higher-level perspective". Step 3: Intelligent Analysis and Preprocessing: The system calls the lightweight deep learning model built into the terminal to directly classify the targets in the top view and retrieve matching icon models from the built-in standard library; Step 4: Logical fitting of spatial coordinates and symbol labels: The identified object outline is superimposed on the measured RTK coordinate points for calculation. Step 5: Interactive one-click vector map generation: Surveyors can easily fix the generated vector symbols at the reference position with a simple click.
[0009] By combining RTK location trajectories with panoramic projection images, the system uses local algorithms for intelligent target identification, accurately categorizing vehicles, various types of non-motorized vehicles, and pedestrians. Subsequently, the system logically couples the AI-identified object attributes with their physical spatial coordinates. This "dynamic-static" association method ensures that each data collection point corresponds to an accurate physical evidence definition, providing a basis for automatic symbol placement.
[0010] In a preferred embodiment, the system also includes a data acquisition module, a mapping module, and a results output module. The data acquisition module realizes the original transformation from the physical world to the digital space. The data acquisition module includes an RTK centimeter-level positioning unit and a panoramic image evidence unit. RTK centimeter-level positioning: abandons traditional tape measure measurement and obtains the absolute coordinates of accident elements through differential positioning technology. Panoramic image evidence: automatically links the camera when collecting data points to capture a 360° environmental base map, ensuring that the "measurement data" and "visual evidence" are completely synchronized in the spatiotemporal dimension.
[0011] By integrating a panoramic module to simultaneously capture 360° omnidirectional digital images during coordinate acquisition, the collision details, traffic signs, surrounding facilities, and road environment at the accident scene are completely locked. This method changes the traditional local evidence collection model, effectively avoiding the omission of key evidence due to human error, and providing comprehensive base materials for subsequent semantic analysis.
[0012] In a preferred embodiment, the drawing module provides a highly intelligent drawing environment to assist manual drawing quickly. This module includes an AI recognition assistance unit, a standard icon library unit, and an interactive drawing terminal unit. The AI recognition assistance unit automatically identifies objects in images and provides attribute suggestions (labels). Instead of drawing directly, it acts as an "intelligent advisor," indicating the current coordinates to the user, reducing the time spent thinking and selecting. The standard icon library unit provides a complete set of vector icons conforming to national standards. Users do not need to draw them themselves; they can directly call standard symbols from the library. The interactive drawing terminal is the system's "operating console." Based on the coordinates provided by RTK and the attribute suggestions provided by AI, users can perform hand-drawing operations with a simple click and drag-and-drop interface, providing a good interactive experience. The output module standardizes the output of the completed drawings and includes vector field images. These vector field images save the manually drawn results in high-definition vector format, supporting arbitrary scaling and on-site wireless printing.
[0013] By relying on a pre-built national standard (GB) symbol resource library, the system supports directly "attaching" standard symbols to measured coordinate points, generating vector graphics with accurate scales and standardized annotations in a streamlined process. This step significantly reduces the workload of manual drawing in the later stages, enabling surveyors to quickly produce standardized site maps before leaving the site.
[0014] As shown above, an accident investigation mapping system based on RTK positioning and panoramic imagery includes an RTK high-precision spatial positioning module, a panoramic image acquisition module, an AI target attribute recognition unit, a local national standard symbol resource library, a high-performance fusion drawing module, and a standardized vector scene map. The RTK high-precision spatial positioning module acquires the absolute spatial coordinates of key physical evidence at the accident scene in real time, and it forms the physical framework of the entire system. The panoramic image acquisition module captures panoramic images of the scene, serving as the raw material for AI recognition, ensuring that the mapping process has a basis and verifiable data. The AI target attribute recognition unit automatically analyzes objects in the imagery, defining them as specific category labels. Furthermore, the AI target attribute recognition unit no longer requires complex pose calculations; it only... It is necessary to confirm that this is a specific target. The local national standard symbol resource library stores various vector icons that conform to national safety industry standards. Each identified "object label" can find the corresponding "legally customized graphic symbol" here. The high-performance fusion drawing module receives the coordinate point of S1, the category label of T1, and the standard icon of T2. The high-performance fusion drawing module accurately places the "standard icon" on the "RTK coordinate" and automatically adjusts the size according to the preset global scale to complete the final rendering of the graphic. The standardized vector scene map automatically generates a traffic accident scene map containing the real scale, absolute coordinates, and standard symbols. The traffic accident scene map is in vector format and can be directly exported, wirelessly transmitted, or printed on-site as core physical evidence for determining liability in the accident. The accident investigation mapping system and method based on RTK positioning and panoramic imagery provided by this invention has achieved a leap from "manual icon selection" to "machine target recognition". The recognition module can automatically tell the engine "this point is a vehicle", eliminating tedious manual annotation and achieving a true "second-level map generation" technical effect. It automatically formats the map according to official document standards (center title, top compass, clear legend) without the need for secondary post-processing, realizing a closed loop of "sampling is drafting, and drafting is archiving", which greatly saves the office time of grassroots police officers. Attached Figure Description
[0015] Figure 1 This is a schematic diagram of the overall structure of an accident investigation and mapping system based on RTK positioning and panoramic imagery proposed in this invention.
[0016] Figure 2 This is a flowchart illustrating the implementation of an accident investigation mapping method based on RTK positioning and panoramic imagery proposed in this invention.
[0017] Figure 3 This is a module structure diagram of an accident investigation mapping method based on RTK positioning and panoramic imagery proposed in this invention. Detailed Implementation
[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.
[0019] Reference Figure 1 An accident investigation mapping system based on RTK positioning and panoramic imagery includes an RTK high-precision spatial positioning module, a panoramic image acquisition module, an AI target attribute recognition unit, a local national standard symbol resource library, a high-performance fusion drawing module, and a standardized vector scene map. The RTK high-precision spatial positioning module acquires the absolute spatial coordinates of key physical evidence at the accident scene (such as brake marks, vehicle wheel hub center points, and curb stones) in real time, and serves as the physical framework of the entire system. The panoramic image acquisition module captures panoramic images of the scene, which serve as the raw material for AI recognition, ensuring that the mapping process has a basis and verifiable data. The AI target attribute recognition unit automatically analyzes objects in the images and defines them as specific category labels (e.g., motor vehicles, non-motor vehicles, human bodies, etc.). The unit no longer needs complex attitude calculations; it only needs to confirm that the location is a specific target. The local national standard symbol resource library stores various vector icons that conform to national safety industry standards. Each identified "object label" can find its corresponding "legally customized graphic symbol" here. The high-performance fusion drawing module receives the coordinate points of S1, the category label of T1, and the standard icon of T2. The high-performance fusion drawing module accurately places the "standard icon" on the "RTK coordinates" and automatically adjusts its size according to the preset global scale to complete the final rendering of the graphic. The standardized vector scene diagram automatically generates a traffic accident scene diagram containing the true scale, absolute coordinates, and standard symbols. The traffic accident scene diagram is in vector format and can be directly exported, wirelessly transmitted, or printed on-site, serving as core evidence for determining liability in the accident.
[0020] In a preferred embodiment, the RTK high-precision spatial positioning module provides centimeter-level absolute coordinates, which is the only physical basis to replace traditional tape measure measurements. The on-site map in the local national standard symbol resource library must conform to the public security industry standards. The high-performance fusion drawing module is responsible for logically binding the scattered coordinate points with abstract national standard symbols. Without this engine, the collected data are just isolated points or maps, and it is impossible to synthesize a complete vector map.
[0021] Reference Figure 2 and Figure 3 An accident investigation mapping method based on RTK positioning and panoramic imagery, specifically including the following steps: Step 1: Synchronous Data Acquisition at the Site: Survey personnel use mobile acquisition terminals to obtain the CM-level coordinates of key evidence points or center lines in seconds, establishing the unique physical benchmark at the site and solving the pain point of "inaccurate measurement". Step Two: Panoramic Image Capture and Real-Time Top-Down Reconstruction: Utilizing edge computing capabilities, the panoramic image is instantly converted into a top-down view from a higher perspective, resolving the "perspective cognitive bias." The base map seen by surveyors on the screen becomes the "blueprint" for subsequent drawing, eliminating the difficulty of reviewing multiple photos for mental modeling after taking pictures in the traditional method. Step 3: Intelligent Analysis and Preprocessing: The system calls the terminal's built-in lightweight deep learning model to directly classify targets in the top-down view and retrieves matching icon models from the built-in standard library, solving the problem of "rapid recognition and offline availability". The local library reduces network transmission latency, ensuring that even in weak network environments, accident elements can still be quickly identified and national standard symbols can be called. Step Four: Logical Fitting of Spatial Coordinates and Symbol Labels: The identified object outline is superimposed on the measured RTK coordinate points, solving the "virtual-real matching" problem. This is not simply placing an icon, but ensuring that the icon's size, vehicle direction, and spacing ratio are highly consistent with the actual physical quantities on site, demonstrating extremely high legal rigor. Step 5: Interactive one-click vector map generation: Surveyors can fix the generated vector symbols on the reference position with a simple click, which solves the "post-work pressure". The generated vector map can be directly exported, realizing "what you see is what you get", and completely eliminating the repetitive work of reviewing drafts and manually drawing CAD after returning to the unit.
[0022] In a preferred embodiment, it further includes a data acquisition module, a drawing module, and a result output module. The data acquisition module realizes the original transformation from the physical world to the digital space, and the data acquisition module includes an RTK centimeter-level positioning unit and a panoramic image verification unit.
[0023] In a preferred embodiment, RTK centimeter-level positioning: abandoning traditional tape measure measurement, the absolute coordinates of accident elements are obtained through differential positioning technology; panoramic image evidence: the camera is automatically linked when collecting data points to capture a 360° environmental base map, ensuring that the "measurement data" and "visual evidence" are completely synchronized in the spatiotemporal dimension.
[0024] In a preferred embodiment, the drawing module provides a highly intelligent drawing environment to assist manual drawing quickly. The drawing module includes an AI recognition assistance unit, a standard icon library unit, and an interactive drawing terminal unit. The AI recognition assistance unit automatically identifies objects in the image and provides attribute suggestions (labels). It does not draw directly but acts as an "intelligent advisor," prompting the user about the current coordinates and reducing the user's thinking and selection time.
[0025] In a preferred embodiment, the standard icon library unit provides a complete set of vector icons that conform to national standards. Users do not need to draw them themselves; they can directly call standard symbols from the library. The interactive drawing terminal is the system's "operation console." Relying on the coordinate points provided by RTK and the attribute suggestions provided by AI, users can perform hand-drawing operations with a simple click and drag-and-drop interface and a good interactive experience.
[0026] In a preferred embodiment, the output module standardizes the output of the completed drawings, and the output module includes a vector field map. The vector field map saves the results drawn manually in a high-definition vector format, supports arbitrary scaling, and can be wirelessly printed on site.
[0027] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. An accident investigation mapping system based on RTK positioning and panoramic imagery, comprising an RTK high-precision spatial positioning module, a panoramic image acquisition module, an AI target attribute recognition unit, a local national standard symbol resource library, a high-performance fusion drawing module, and a standardized vector scene map, characterized in that, The RTK high-precision spatial positioning module acquires the absolute spatial coordinates of key physical evidence at the accident scene in real time, and forms the physical framework of the entire system. The panoramic image acquisition module captures panoramic images of the scene, serving as the raw material for AI recognition, ensuring that the mapping process has a basis and data to rely on. The AI target attribute recognition unit automatically analyzes objects in the images and defines them as specific category labels. Furthermore, the AI target attribute recognition unit no longer requires complex pose calculations; it only needs to confirm that the location is a specific target. The local national standard symbol resource library stores various vector icons that conform to national safety industry standards. Each identified "object tag" can find its corresponding "legalized graphic symbol" here. The high-performance fusion drawing module receives the coordinate points of S1, the category label of T1, and the standard icon of T2. The high-performance fusion drawing module accurately places the "standard icon" on the "RTK coordinates" and automatically adjusts its size according to the preset global scale to complete the final rendering of the graphic. The standardized vector scene map automatically generates a traffic accident scene map containing the real scale, absolute coordinates, and standard symbols. The traffic accident scene map is in vector format and can be directly exported, wirelessly transmitted, or printed on-site, serving as the core physical evidence for determining liability in the accident.
2. The accident investigation mapping system based on RTK positioning and panoramic imagery according to claim 1, characterized in that, The RTK high-precision spatial positioning module provides centimeter-level absolute coordinates, which is the only physical basis to replace traditional tape measure measurements. The on-site map in the local national standard symbol resource library must conform to the public security industry standards. The high-performance fusion drawing module is responsible for logically binding scattered coordinate points with abstract national standard symbols. Without this engine, the collected data are just isolated points or images, and it is impossible to synthesize a complete vector map.
3. A method for accident investigation mapping based on RTK positioning and panoramic imagery, applied to any of the accident investigation mapping systems based on RTK positioning and panoramic imagery described in claim 1, characterized in that, Specifically, the following steps are included: Step 1: Synchronous Data Acquisition at the Site: Survey personnel use mobile data acquisition terminals to obtain the CM-level coordinates of key evidence points or center lines within seconds; Step 2: Panoramic Image Capture and Real-time Top-Down Reconstruction: The system utilizes edge computing capabilities to instantly convert panoramic images into a top-down view from a "higher-level perspective". Step 3: Intelligent Analysis and Preprocessing: The system calls the lightweight deep learning model built into the terminal to directly classify the targets in the top view and retrieve matching icon models from the built-in standard library; Step 4: Logical fitting of spatial coordinates and symbol labels: The identified object outline is superimposed on the measured RTK coordinate points for calculation. Step 5: Interactive one-click vector map generation: Surveyors can easily fix the generated vector symbols at the reference position with a simple click.
4. The accident investigation mapping method based on RTK positioning and panoramic imagery according to claim 3, characterized in that, It also includes a data acquisition module, a drawing module, and a result output module. The data acquisition module realizes the original transformation from the physical world to the digital space, and the data acquisition module includes an RTK centimeter-level positioning unit and a panoramic image verification unit.
5. The accident investigation mapping method based on RTK positioning and panoramic imagery according to claim 4, characterized in that, The RTK centimeter-level positioning abandons traditional tape measure measurement and obtains the absolute coordinates of accident elements through differential positioning technology; the panoramic image evidence consolidation automatically links the camera when collecting data points to capture a 360° environmental base map, ensuring that the "measurement data" and "visual evidence" are completely synchronized in the spatiotemporal dimension.
6. The accident investigation mapping method based on RTK positioning and panoramic imagery according to claim 5, characterized in that, The drawing module provides a highly intelligent drawing environment to assist manual drawing quickly. The drawing module includes an AI recognition assistance unit, a standard icon library unit, and an interactive drawing terminal unit. The AI recognition assistance unit automatically identifies objects in the image and provides attribute suggestions (labels). It does not draw directly but acts as an "intelligent advisor," prompting the user about the current coordinates and reducing the user's thinking and selection time.
7. The accident investigation mapping method based on RTK positioning and panoramic imagery according to claim 6, characterized in that, The standard icon library unit provides a complete set of vector icons that conform to national standards. Users do not need to draw them themselves; they can directly call standard symbols from the library. The interactive drawing terminal is the system's "operation console." Relying on the coordinate points provided by RTK and the attribute suggestions provided by AI, users can perform hand-drawing operations with a simple click and drag-and-drop interface and a good interactive experience.
8. The accident investigation mapping method based on RTK positioning and panoramic imagery according to claim 7, characterized in that, The output module standardizes the output of the completed drawings, and includes a vector field map. The vector field map saves the results drawn manually in a high-definition vector format, supports arbitrary scaling, and can be wirelessly printed on site.