基于视觉感知模型的智能车辆道路检测技术-控制科学与工程专业论文.docx

基于视觉感知模型的智能车辆道路检测技术 基于视觉感知模型的智能车辆道路检测技术 RESEARSH ON INTELLIGENT VEHICLE LANE DETECTION BASED ON VISUAL PERCEPTION MODEL ABSTRACT Intelligent vehicle integrated multi-sensor data fusion, visual information processing, environment modeling, planning decision-making, multi-level driver assistance functions in one system, which has been rapid development in recent years. Lane detection technology based on vision is a hot area of intelligent vehicle research direction, and lane detection and recognition is the main research contents to achieve this technology. Research paper should include three dimension about visual perception model structure, lane image processing and lane fitting and lane departure warning. From the perspective of understanding the lane environment, application based on visual perception model of visual attention mechanisms used to detect lane in front of moving vehicle prospects to make an warning when the vehicle deviated from the place through acquisition of vehicle current position information and direction information. The first we use vision-based perceptual model of visual attention mechanisms for the initial image perception and recognition, then extract the image feature information, arise the map of character salience from different image feature information through center-surround differences and normalization, generate saliency map using the strategy of multi-character merging, test high significance of area in saliency map through the mechanism of competition and should mark it. After inhibition of the most significant regions in order to shift the next highest significant region, we will extract the focus of attention regions until the completion of perceptual identification. In order to validate the effectiveness of the algorithm, we do related experiments under different lane environment, obtain experimental results, compare with other algorithm at same time. KEY WORDS: visual perception model, lane image processing, lane detection ii