The Algorithm Of Drive Line Detection

With the development of China's automobile industry, as well as the popularity of the car into the home is becoming more the ensuing traffic problems are also becoming more serious, which is the most important traffic safety issue, this article focuses on today's cars the navigation of vehicle's supplementary The lane marking line recognition detection techniques and algorithms related to the corresponding study. In this paper, some shortcomings exist in traditional algorithms, made a new model features the structure of the road, combined with morphological image processing related to the theory and methods, obtained through a flexible simulation-based morphology detection algorithm for lane. The algorithm was first scanned image of the extract, according to its own ladder lane characteristics, creating a ladder of interest in the region, so that not only reduce the corresponding calculations, while the biggest benefit is reduced lane of the environment on lane line detection. And through the alternating sequence of filters related to the original image is filtered using a lane model features the structure of color images to deal with specific targets after the extraction, and then right outside lane marking lines to conduct operations to remove the double-edge, integration of the results, will be the edge line into the inside lane. Line Hough Transform and thus marked lane. This simulation involved in the handling of the images were obtained using light-sensitive color CCD image, is converted to 256 gray levels, in order to be more realistic response to the actual road conditions, the image acquisition were collected from different cities under different circumstances, the road The purpose is to simulation of a more complex traffic conditions, such as weather, the extent of road clean and so on, through the simulation of the algorithm can meet the expectations, to detect target identification line, high accuracy, the relative robustness of the traditional method has been markedly increase.