Road hidden danger automatic acquisition and positioning system based on multi-device cooperation

Abstract

The invention is suitable for the technical field of road traffic safety, and provides a road hidden danger automatic acquisition and positioning system based on multi-device collaboration, which comprises a sensing layer, a data processing layer, a hidden danger identification and positioning layer and an application layer, and is characterized in that the sensing layer is used for acquiring related multi-source data of a road; according to the automatic road hidden danger collecting and positioning system based on multi-device collaboration, through multi-device collaboration collection of the ultra-high-definition vision module, the Beidou positioning terminal and the like, parking or manual intervention is not needed, all-directional road hidden danger data obtaining is achieved, the problem that the collection dimension of a single device is insufficient is solved, and the collection coverage range and convenience are improved. An edge computing and centralized processing combined mode is adopted, the defects of traditional cloud computing are avoided, and the data processing efficiency and the system stability are improved. A linear regression model and a deep learning technology capable of being updated online are fused, the core scene recognition accuracy and the complex scene adaptability are guaranteed, the hidden danger recognition requirement is comprehensively covered, and the traffic accident risk is reduced from the source.

Description

Technical Field

[0001] This invention belongs to the field of road traffic safety technology, and in particular relates to an automated road hazard collection and positioning system based on multi-device collaboration. Background Technology

[0002] As the core carrier of transportation, the safety of roads directly affects the safety of people's lives and property, transportation efficiency, and the stable development of the social economy. Road hazards are a major cause of traffic accidents, encompassing various types such as road surface defects, abnormal road alignment, road connection problems, missing or damaged safety facilities, and special environmental risks. Timely, comprehensive, and accurate investigation of such hazards is a key prerequisite for preventing traffic accidents and improving road safety, and is also one of the core tasks of traffic management departments.

[0003] Currently, road hazard investigation relies heavily on manual surveys, which suffers from low efficiency, high costs, and strong subjectivity. Furthermore, manual investigations are difficult to adapt to complex road conditions and large-scale investigation needs, and are prone to overlooking hazards. In existing technologies, some single-device data acquisition systems can achieve some data acquisition functions, but they have shortcomings such as insufficient positioning accuracy, limited hazard identification dimensions, and poor real-time data processing. Summary of the Invention

[0004] This invention provides an automated road hazard collection and location system based on multi-device collaboration, aiming to solve the problems of current road hazard investigation work relying heavily on manual surveys, which suffers from low efficiency, high cost, strong subjectivity, and easy omissions.

[0005] This invention is implemented as follows: an automated road hazard collection and location system based on multi-device collaboration, comprising a perception layer, a data processing layer, a hazard identification and location layer, and an application layer;

[0006] The perception layer is used to collect multi-source road-related data, the data processing layer preprocesses and fuses the collected data, the hazard identification and location layer realizes hazard classification and accurate location, and the application layer is used for data display, management and early warning.

[0007] Preferably, the perception layer includes an ultra-high-definition vision module, a high-precision positioning module, an attitude perception module, and a communication module. The ultra-high-definition vision module uses a high-speed camera to capture images of the road surface, road facilities, and surrounding environment. The high-precision positioning module is a Beidou positioning terminal used to provide positioning information. The attitude perception module includes a gyroscope and a level to acquire device attitude parameters and road elevation-related information. The communication module is used to realize the real-time transmission of the collected data.

[0008] Preferably, the data processing layer includes a positioning data processing unit, an image preprocessing unit, and a multi-source data fusion unit. The positioning data processing unit performs coordinate transformation and data cleaning. The image preprocessing unit performs image enhancement, segmentation, and feature extraction operations. The multi-source data fusion unit fuses image data, positioning data, attitude data, and elevation-related data to form a unified hazard analysis dataset.

[0009] Preferably, the hazard identification and location layer includes a scene identification unit and a location fusion unit. The scene identification unit uses a linear regression model to identify scenes such as steep slopes and sharp bends, and uses a deep learning model to identify scenes such as road openings, waterside or cliffside locations, and various road facilities. The location fusion unit associates the identified hazard types with geographic coordinate information to achieve accurate hazard location.

[0010] Preferably, the application layer includes a visual integrated dashboard, a data management module, and an early warning module. The visual integrated dashboard is used to update and summarize location information, hazard information, and risk analysis data in real time. The data management module is used to build a basic database of road assets, store various types of collected data, processing results, and identification and location information, and the early warning module is used to provide immediate early warnings for major hazards.

[0011] Preferably, after preprocessing, the images acquired by the ultra-high-definition vision module can extract information on road surface defects, surrounding environmental features, and road facility outlines, providing multi-dimensional image data support for hazard identification.

[0012] Preferably, the data processing layer adopts a combination of edge computing and centralized processing. The edge computing unit is deployed locally on the acquisition device to realize real-time data preprocessing, and the centralized processing server is used for deep fusion and global analysis of multi-source data, reducing the degree of human intervention.

[0013] Preferably, the deep learning model supports online updates and optimizations, and can continuously improve its recognition adaptability to various road scenarios and facilities through iteration. The linear regression model achieves accurate recognition of the corresponding scenario by analyzing road-related geometric parameters.

[0014] Beneficial effects

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention provides an automated road hazard collection and positioning system based on multi-device collaboration. This system achieves comprehensive acquisition of road hazard-related data through the collaborative collection of multiple devices, including ultra-high-definition vision modules, Beidou positioning terminals, gyroscopes, and levels. This solves the problem of insufficient data collection dimensions from a single device, and the collection process requires no stopping or manual intervention, significantly improving the collection coverage and convenience. Simultaneously, it adopts a combination of edge computing and centralized processing to achieve real-time data preprocessing and deep fusion of multi-source data, reducing manual intervention and effectively solving the problems of transmission latency and poor adaptability to edge scenarios in traditional cloud computing. This improves data processing efficiency and system stability. Furthermore, it combines a linear regression model with online-updable deep learning technology for hazard identification. The linear regression model ensures the accuracy of identification for core scenarios such as steep slopes and sharp bends by analyzing road geometric parameters, while the deep learning model continuously iterates to improve its adaptability to complex scenarios such as road openings and facilities. This comprehensively covers the identification needs of different types of road hazard, significantly improving the comprehensiveness of identification and reducing the risk of traffic accidents from the source. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall architecture of the present invention;

[0017] Figure 2 This is a schematic diagram of the internal architecture of the perception layer in this invention;

[0018] Figure 3 This is a schematic diagram of the internal architecture of the data processing layer in this invention;

[0019] Figure 4 This is a schematic diagram of the internal architecture of the hazard identification and location layer in this invention;

[0020] Figure 5 This is a schematic diagram of the internal architecture of the application layer in this invention.

[0021] In the diagram: 1-Perception layer, 11-Ultra-high-definition vision module, 12-High-precision positioning module, 13-Attitude perception module, 14-Communication module, 2-Data processing layer, 21-Positioning data processing unit, 22-Image preprocessing unit, 23-Multi-source data fusion unit, 3-Hazard identification and positioning layer, 31-Scene recognition unit, 32-Positioning fusion unit, 4-Application layer, 41-Visual integrated dashboard, 42-Data management module, 43-Early warning module. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0023] Please see Figure 1-4 The present invention provides a technical solution: an automated road hazard collection and location system based on multi-device collaboration, comprising a perception layer 1, a data processing layer 2, a hazard identification and location layer 3, and an application layer 4;

[0024] The perception layer 1 is used to collect multi-source data related to roads; the data processing layer 2 preprocesses and fuses the collected data; the hazard identification and location layer 3 realizes the classification and accurate location of hazards; and the application layer 4 is used for data display, management and early warning.

[0025] The perception layer 1 includes an ultra-high-definition vision module 11, a high-precision positioning module 12, an attitude perception module 13, and a communication module 14. The ultra-high-definition vision module 11 uses a high-speed camera to capture images of the road surface, road facilities, and surrounding environment. The high-precision positioning module 12 is a Beidou positioning terminal used to provide positioning information. The attitude perception module 13 includes a gyroscope and a level to acquire equipment attitude parameters and road elevation-related information. The communication module 14 is used to realize the real-time transmission of the collected data.

[0026] The data processing layer 2 includes a positioning data processing unit 21, an image preprocessing unit 22, and a multi-source data fusion unit 23. The positioning data processing unit 21 performs coordinate transformation and data cleaning, the image preprocessing unit 22 performs image enhancement, segmentation, and feature extraction operations, and the multi-source data fusion unit 23 fuses image data, positioning data, attitude data, and elevation-related data to form a unified hazard analysis dataset.

[0027] The hazard identification and location layer 3 includes a scene identification unit 31 and a location fusion unit 32. The scene identification unit 31 uses a linear regression model to identify scenes such as steep slopes and sharp bends, and uses a deep learning model to identify scenes such as road openings, water and cliffs, and various road facilities. The location fusion unit 32 associates the identified hazard types with geographic coordinate information to achieve accurate hazard location.

[0028] Application layer 4 includes a visual integrated dashboard 41, a data management module 42, and an early warning module 43. The visual integrated dashboard 41 is used to update and summarize location information, hazard information, and risk analysis data in real time. The data management module 42 is used to build a basic database of road assets, store various types of collected data, processing results, and identification and location information. The early warning module 43 is used to provide immediate early warnings for major hazards.

[0029] After preprocessing, the images acquired by the ultra-high-definition vision module 11 can extract information on road surface defects, surrounding environmental features, and road facility outlines, providing multi-dimensional image data support for hazard identification.

[0030] Data processing layer 2 adopts a combination of edge computing and centralized processing. The edge computing unit is deployed locally on the acquisition device to realize real-time data preprocessing, while the centralized processing server is used for deep fusion and global analysis of multi-source data, reducing the degree of human intervention.

[0031] Deep learning models support online updates and optimizations, and can continuously improve their ability to identify and adapt to various road scenarios and facilities through iteration. Linear regression models achieve accurate identification of corresponding scenarios by analyzing road-related geometric parameters.

[0032] The working principle and usage process of this invention: The system can be mounted on a vehicle. In practice, the patrol vehicle is equipped with the system and is in motion. During this process, the ultra-high-definition vision module 11, Beidou positioning terminal, gyroscope and level of the perception layer 1 are activated simultaneously to automatically collect road images, positioning data, attitude parameters and elevation-related information. This data is transmitted in real time to the local edge computing unit by the communication module 14. The data processing layer 2 then preprocesses the transmitted data, generating a unified hazard analysis dataset through image optimization, coordinate transformation, data cleaning and multi-source data fusion. The hazard identification and positioning layer 3 calls the linear regression model and the online-updable deep learning model to identify core scenarios such as steep slopes and sharp bends, as well as complex scenarios such as road openings and facilities. It also associates geographic coordinates to determine the precise location of the hazard. Finally, the application layer 4 displays hazard information and risk data through the visualization dashboard 41. The data management module 42 completes data storage traceability, and the early warning module 43 provides immediate reminders for major hazards, assisting traffic management departments in carrying out rectification work and forming a closed-loop management process.

[0033] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An automated road hazard acquisition and location system based on multi-device collaboration, characterized in that: It includes a perception layer (1), a data processing layer (2), a hazard identification and location layer (3), and an application layer (4); The perception layer (1) is used to collect multi-source data related to roads. The data processing layer (2) preprocesses and fuses the collected data. The hazard identification and positioning layer (3) realizes hazard classification and precise positioning. The application layer (4) is used for data display, management and early warning.

2. The automated road hazard collection and location system based on multi-device collaboration as described in claim 1, characterized in that: The perception layer (1) includes an ultra-high-definition vision module (11), a high-precision positioning module (12), an attitude perception module (13), and a communication module (14). The ultra-high-definition vision module (11) uses a high-speed camera to capture images of the road surface, road facilities, and surrounding environment. The high-precision positioning module (12) is a Beidou positioning terminal used to provide positioning information. The attitude perception module (13) includes a gyroscope and a level, used to acquire equipment attitude parameters and road elevation information. The communication module (14) is used to realize the real-time transmission of the collected data.

3. The automated road hazard acquisition and location system based on multi-device collaboration as described in claim 1, characterized in that: The data processing layer (2) includes a positioning data processing unit (21), an image preprocessing unit (22), and a multi-source data fusion unit (23). The positioning data processing unit (21) performs coordinate transformation and data cleaning. The image preprocessing unit (22) performs enhancement, segmentation, and feature extraction operations on the image. The multi-source data fusion unit (23) fuses image data, positioning data, attitude data, and elevation-related data to form a unified hazard analysis dataset.

4. The automated road hazard collection and location system based on multi-device collaboration as described in claim 1, characterized in that: The hazard identification and location layer (3) includes a scene identification unit (31) and a location fusion unit (32). The scene identification unit (31) uses a linear regression model to identify steep slopes, sharp bends and other scenes, and uses a deep learning model to identify road openings, water and cliffs and other scenes and various road facilities. The location fusion unit (32) associates the identified hazard type with the geographic coordinate information to achieve accurate hazard location.

5. The automated road hazard collection and location system based on multi-device collaboration as described in claim 1, characterized in that: The application layer (4) includes a visual integrated dashboard (41), a data management module (42), and an early warning module (43). The visual integrated dashboard (41) is used to update and summarize location information, hidden danger information, and risk analysis data in real time. The data management module (42) is used to build a basic database of road assets and store various types of collected data, processing results, and identification and location information. The early warning module (43) is used to provide immediate early warning for major hidden dangers.

6. The automated road hazard acquisition and location system based on multi-device collaboration as described in claim 2, characterized in that: After preprocessing, the images collected by the ultra-high-definition vision module (11) can extract information on road surface defects, surrounding environmental features and road facility outlines, providing multi-dimensional image data support for hazard identification.

7. The automated road hazard acquisition and location system based on multi-device collaboration as described in claim 3, characterized in that: The data processing layer (2) adopts a combination of edge computing and centralized processing. The edge computing unit is deployed locally on the acquisition device to realize real-time data preprocessing, and the centralized processing server is used for deep fusion and global analysis of multi-source data to reduce human intervention.

8. The automated road hazard acquisition and location system based on multi-device collaboration as described in claim 4, characterized in that: The deep learning model supports online updates and optimizations, and can continuously improve its ability to identify and adapt to various road scenarios and facilities through iteration. The linear regression model achieves accurate identification of corresponding scenarios by analyzing road-related geometric parameters.

Images