영상 및 레이더 센서 융합 기반의 측후방 감지 시스템 개발

Abstract

This paper presents a blind-spot detection system using vision and radar sensor fusion. The proposed method associates detection results from rear-view image, side rectified image and radar sensor, then it fuses them using multi-measure Kalman filtering. Additionally, the detection zone is adaptively established based on the lane detection result to minimize false alarm caused by vehicles on the second side lane. Experimental results show that the proposed method exploits complementary merits of vision and radar sensors.