Design and Experimental Investigation of DMD Lamp to Improve Driver's Visual Perception

Shuo Zhang,Jianwei Huang,Ping Su,Jianshe Ma

DOI: https://doi.org/10.1109/icosr57188.2022.00011

2022-01-01

Abstract:The core of driving is vision. Drivers need to obtain information based on vision perception to make decisions. Previous studies on smart headlamps mostly focused on several specific issues, such as the anti-glare function and the projection function. These functions indirectly have a certain effect on the driver's visual perception, but they do not fully consider the surrounding environment, which is easy to affect the driver's perception of other elements in the scene, resulting in potential safety hazards. Besides, most of the existing research focuses on the optical design, there is no design and test at the system level. This research creatively presents the concept of visual perception optimization for the whole scene ahead, uses an ordinary vehicle monocular camera to capture the scene ahead, and uses a YOLO deep learning algorithm to identify vehicles and obtain location information. Since eliminating glare interference, for other elements of the scene, we write an algorithm based on the minimum perceptual difference theory of human eyes to get the optimized light type to improve visual perception in real time. The pixel structure DMD (Digital Micromirror Device) is used for variable depth projection to realize the real-time perception optimization of the front scene. Experimental results show that the system can improve the image quality, natural degree and human eye acceptance of the captured scene, improve the visual perception of human eyes and ensure traffic safety.