tailieunhanh - A study on real time lane change assistant method with side mirror cameras
According to the achieved data, lane detection is developed by analyzing angles of the line segments, and area between two lanes on the same side of vehicle is defined. In the vehicle detection, based on the brightness and darkness between vehicle and road, vehicle is detected in real-time using the simple algorithm. Finally, kalman filter is used in vehicle tracking for vehicle information such as distance or speed. | Journal of Automation and Control Engineering Vol. 4, No. 6, December 2016 A Study on Real Time Lane Change Assistant Method with Side-Mirror Cameras Van Quang Nguyen Department of Mechanical and Automotive Engineering, Inje Univeristy, Gimhae, Gyeongnam, Korea Email: vanquang741987@ Duksun Yun Intelligent System Research Lab., Korea Automotive Technology Institute (KATECH), Cheonan, Korea Email: dsyun@ Woosung Che Department of Mechatronics Engineering, Kyungsung University, Busan, Korea Email: wsche@ Heung Seob Kim and Kwangsuck Boo High Safety Vehicle Core Technology Research Center, Inje Univeristy, Gimhae, Gyeongnam, Korea Email: mechhsk@, mechboo@ In the lane detection, we can approach on two features. The first based on boundary detection consists of three approaches: Edge detection [1], [2], vanishing point detection [3], [4], and Hough transform [5]. Edge feature and Hough transform are relied on the difference of brightness between road and road boundary, or between lane markings and road. And lane markings are detected by clear edges or by straight lines. Various articles assume that the lane markings converge at the vanishing point in image. Therefore, they firstly detect the vanishing point, then lane or road based on the vanishing point. The second one relies on the region features, which include texture features [6], color features [7], [8], and intensity invariant features [9]. Texture-based road detection methods focus on identification of textural differences between road and on road regions to segregate. In addition, color based features approach on the difference of color between the road surface and the outside area of road. Road detection is studied on the shadows and illuminant invariant image. Vehicle detection and tracking have been widely researched in recent years. Two common knowledgebased and region-based methods are proposed. The knowledge-based methods employ a priori knowledge .
đang nạp các trang xem trước
