Design Of Position System Of The Road-lineation Vehicle Based On Machine Vision With DSP

In recent years, China’s road construction develops rapidly, and road lineation is an important part of the road, which is closely related to driving safety, but China’s road construction and equipments develop lag. Among them, road lineation vehicle demand obviously. Most of the road lineation machine in our country is manual operation, low efficiency. Therefore, the intelligent autoi ation needs obviously.In order to solve the problem that road lineation machine marks line automaticly. This paper based on TI C6000 - DSP, using the camera to collect road image, using the DSP development board for processing and analysis of the road image, thus completing the midline and other lines.The hardware part introduces the composition and function of the video capture circuit board, include capture and display circuit module, and memory mapping space. The program design includes system and application, the main function of the system program is initialization, prepare for the loading of the application; the various functions of the application are responsible for implementation of lineation control system.The image processing algorithm includes three parts:1. Initializing the road images, including image-gradation, thresholding, edge detection;2. Using thresholding and distance transform to realize the middle lineation control. Select the road after thresholding processing as the target, then select both sides of the road as the background. The distance matrix can be got after the distance transform. The peak point of the road distance matrix shows the road center, and the vehicle is controled to walk along the midline of the road ultimately.3. Using Hough Transform to realize other lineation control. The Hough Transform can detect the edge of the road line, then use this line as reference, control the vehicle to run parallel with it.The development board test results show that: the recognition of road centerline has been completed based on distance transform, the maximum distance between road pixels and the sides display the brightest in the grayed image. So that these pixels are the center of the road; using the Hough Transform algorithm can also realize the recognition of other lines; processing and recognition process are stable, even the DSP runs stable and there’s no lag and crash.