Real-time automotive night-vision system for drivers to inhibit headlight glare of the oncoming vehicles and enhance road visibility

摘要

The vision problem of drivers during the night is mainly owing to the high-intensity headlight-beam of the oncoming vehicle from the reverse direction. It causes temporary blindness to the driver. To overcome this situation, people generally use color windshield glass, sun visor, and night-vision glass. But, these are not considered as the best solution since it decreases the light intensity of the entire view including the road. Many researchers used various image enhancement techniques to overcome this situation. But, the existing approaches are unable to dim the high-beam headlights of oncoming vehicles without affecting the road view. In this paper, a novel night-vision system is proposed to resolve the problem in real-time for manual-driving vehicles and autonomous vehicles. The proposed method includes region segmentation of frames, local enhancement techniques in different regions followed by adaptive Gaussian filtering. Pixels masking, gamma correction, and low-light pixel enhancement are applied to three distinct regions. Both autonomous vehicles and manual drivers can get a bright and a prominent view of the road with dim headlights of oncoming vehicles in real-time. Numerous heuristic real-time test reveals the performance superiority of the projected system compared to state-of-art methods in quantitative as well as qualitative point of view.