A new approach for camera supported machine learning algorithms based dynamic headlight model's design

Traffic accidents continue to be a significant issue in modern society. Accidents usually happen on dark, mountainous, narrow, steep and curved roadways. One of the primary causes of such accidents is the drivers’ weak sight brought on by the headlights of moving vehicles. In this study, a dynamic headlight model was designed using camera supported machine learning algorithms to improve the drivers’ vision during night drive. In this design, the issues of enabling a lighting field supported by image processing programmed with machine learning, dynamic adjustment of the high beam headlights’ LED cells in response to the vehicle approaching from the opposite direction, traffic-sign recognition system, lane-keeping system, and automatic adjustment of headlight angles were addressed. In this direction, a novel dynamic headlight model that will reduce the risk of accidents caused by lighting was presented, and its analyses were performed.