Efficient and Robust Classification Method英文文献资料.pdf

528 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 15, NO. 4, APRIL 2005 Efficient and Robust Classification Method Using Combined Feature Vector for Lane Detection Pangyu Jeong and Sergiu Nedevschi, Member, IEEE Abstract— The aim of this paper is to develop a method for low- 1) An adaptive/ predefined gray level image splitting into cost and accurate classification of highways and rural ways image subimages pixels for lane detection. The method uses three main components: 2) Two classes K-Mean classification of the subimages adaptive/predefined image splitting, subimage level classification and class merging based on homogeneity checking conditions. In using combined and partially tuned gray and texture the first step, a preclassification in road and nonroad pixels is car- based feature vector. ried out, on the resized input image, using the decision tree method. 3) Class merging based on homogeneity checking using As a result of this first step we obtain the road reference feature the reference road feature value. value, and the lane-markings positions in case of highways. For the rural ways image splitting we use a predefined division method, Two different lane detection procedures are used in accor- and for the highways we use an adaptive division method based on dance to input images: highways and rural ways. The analysis of the detected lane-markings. The proposed classification is carried the decision window is used for automatic selection of the lane out on the subimages using the K-mean classifier on a composed detection procedure that is performed in accordance to input gray and texture based feature ve