Algorithm Design of Lane Departure Warning System Based on Image Processing
Peijiang Chen, Junhao Jiang · 2018
Since the advent of lane departure warning system, its content has been expanding and its performance has been improving. The mainstream lane departure warning system is mainly based on machine vision theory. In this paper, the algorithm design of lane departure warning system is carried out from the angle of image processing. The algorithm of image processing mainly uses the matrix properties of the digital image, so the speed of the image depends greatly on the size of the image. In this paper, an adaptive image cutting method based on threshold is added to the image preprocessing, which effectively removes the unrelated parts and improves the speed of the algorithm. Based on the clipped images, the feature points of the lane line are obtained by the median filtering, edge enhancement, binarization, edge extraction and other preprocessing methods. Finally, the Hough transform and analytic geometry method are used to get the lane line and lane center line, and analyze the role of lane center line in lane departure warning.