Study On Road Line Detection And Tracking Algorithm Based On Illuminant Invariance

Due to the rapid development of domestic traffic,the negative impact caused by the rapid increase is traffic accidents,many of which due to lane deviation,so real-time,high reliable lane line detection and tracking has become the main content of vehicle navigation performance requirements.In recent years,there has been some progress in this area because of the efforts of many researchers,for example,the road identification method used in freeway scenes is very mature.In this paper,the detection and tracking of lane line are studied,among which lane detection is widely used in automatic driving and collision prevention alarm system.The lane line detection system is to extract the effective lane line information,through the digital image processing technology and matching.In this paper,the lane line detection includes the pretreatment,lane line detection algorithm,lane line tracking:the kalman filter algorithm and CSK algorithm and the modified CSK algorithm.(1)Lane detection algorithm.Through preprocessing algorithm,inverse perspective transformation,Gauss filtering and quantile method,different illumination images using lane detection method is prepared.Then,Gray scale processing for different illumination images is adopted.Lane identification is mainly based on the improved linear random sample consistency.(2)The lane line tracking algorithm.Based on the study of the Kalman filter algorithm and CSK(Exploiting the Circulant Structure of Tracking-by-Detection with Kernels)tracking algorithm,meanwhile,found that when target is obscured,tracking is failed.Therefore,the modified CSK tracking algorithm is necessarily studied.(3)Algorithm test.In the real scene,lane detection and tracking algorithm is tested.The detection results show that the algorithm can achieve accurate and fast lane detection;tracking results show that,afer which the Kalman filter and CSK tracking algorithm were comparied and analysis for experimental data,the efficiency of CSK is more high than Kalman tracking’s,and the modified CSK algorithm can successfully achieve tracking when target is obscured.