Smart Driving Vision Assistive Technology Research

Both the track and detection of moving vehicles and lane detection are important parts of intelligent traffic system(ITS).In this paper, we apply the computer vision technology into vehicle detection and lane boundary detection. It is also an important sub-system of intelligent transportation system and has broad application prospect and valuable direction in the following subjects. This article is divided into two modules. One is the moving vehicle detection and tracking in foreground. The other is the road lane detection in background. The first module is mainly on the static context of moving target detection and tracking, The objective of which is to obtain motion parameters of moving objects and to achieve real-time tracking of moving objects. Optical flow can distinguish the foreground and the retreated background. It has high precision and can access the moving target motion parameters directly.In terms of moving vehicle detection and tracking. it focuses on detection of moving vehicle based on the regional optical flow and local feature calculation. We use the clustering and segmentation of optical flow to these features to obtain the area. we remove the false target according to a priori knowledge to accurate the real tracking targets and real-time update the background and tracking objects.In terms of lane detection, it uses probability Hough transform to extract the straight or curved road With the least squares fitting the straight line. For the lane keeping problem of intelligent vehicle in visual navigation, we take monocular vision to detect the road boundaries and get the current path information the current road situation.Experiment results show that this method can detect and track the moving targets accurately and achieve a good a lane tracking and prediction.