부산대학교 도서관

Currently, one of the fastest-growing industries in the world is the automotive industry. More and more vehicles are encountered on the road that have Advanced Driver-Assistance Systems (ADAS) installed to perform complex operations while driving. They constantly monitor the environment around the vehicle and reduce the risk of traffic accidents. In this paper, the ADAS solution for estimating the distance of detected pedestrians and drawing a 2D top-view map of their positions in the vicinity of the vehicle is proposed. The solution processes frames simultaneously captured by four cameras located on different sides of the vehicle. It is developed using the C programming language and finally implemented on a real ADAS embedded platform. The proposed solution achieved high performance when tested on video sequences captured by real automotive cameras. Furthermore, it is capable of processing 15 frames per second when operating with state-of-the-art automotive camera resolutions.