Research And Application Of Key Techniques For 3D Panoramic Driving Assist System

Driving assist system has become an important auxiliary equipment for traffic safety.Traditional driving assist system can only provide 2D panoramic image around the car,with the problem of single perspective and poor user experience,and are not helpful to discover some potential security risks.With the rapid development of 3D technology,panoramic driving assist system with 3D effect has become a new research field.However,the related research and application of 3D-based panoramic driving assist system is still relatively insufficient,and there are many optimization spaces left.The paper compared and analyzed the advantages and disadvantages of the existing 3D panoramic driving assist systems in technical viewpoint.Based on the requirements of the system,we designed the scheme to implement the 3D panoramic driving assisted system,and researched the key technologies used in the system,and developed a demo system based on Android platform.The main work of the paper is as follows:(1)designed a scheme to implement the 3D panoramic driving assist system,and researched the key technologies needed in the preprocessing stage of the system,such as camera calibration and overlooking transformation.(2)designed a two-dimensional panoramic top view mosaic algorithm,which improves the quality of the panorama image by introducing the least squares methods in the original algorithm to improve the correctness of the matching,and using the weighted average method to eliminate the splicing seam.(3)researched the implementation of 3D model and texture mapping algorithm on Android platform,and developed a demo system.The experimental results showed that the demo system can obtain panorama image 360° around the car with good 3D visual effect,and had the function of changing the viewpoint freely.