The Study Of A Few Problems Related To Advanced Driver Assistance Systems

Advanced driver assistance systems can effectively reduce accidents due to driver inattention. Advanced driver assistance systems’ key technology is object detection which has attracted the attention of many domestic and foreign departments, enterprises and schools. Now advanced Driver Assitance System based on vision is a popular research direction, since the visual sensors are cheap and easy to install, and the influence of the driver’s line of sight is small. In this thesis the driver assitance systems’ research and analysis is researched, including lane line detection and tracking, the car detection and pedestrian detection. The main contents of this thesis can be divided into three parts.1. This thesis realizes an algorithm for lane line detection based on combining Hough transform and Catmull-Rom spline. The proximal lane line can be approximated straight line, and remote lane line can be approximated curve. For different lane line uses different fitting algorithm. Finally, kalman filter algorithm is realized to improve the tracking efficiency.2. This thesis realizes an algorithm for vehicle detection method based on Haar-like features and Adaboost algorithm. In addition, this thesis develops a single feature to represent the shadow at the bottom of vehicles. This algorithm is proposed to improve the capability of front vehicle detection of the driver assistance system.3. The thesis proposes a pedestrian detection method based on the Histograms of Oriented Gradients(HOG) and improved the feature extraction algorithm. The method extractses HOG features over an entire image and collect them in H, S, V three channels, then a classifier using Support Vector Machine(SVM) is tarined to realize the Pedestiran detecion.