Line Extraction Method Based on Distance Weighted Least Squares and Self-Adaption Segmentation

A line extraction method based on distance weighted least squares under split-and-merge framework is proposed. With a self-adaption region segmentation algorithm, the laser radar data divided into different regions, and the lines in each region is segmented by Iterative-End-Point-Fit algorithm, on this basis, the relationship between distances of measurements and variances are analyzed, and lines are fitted based on distance weighted least squares method. Experimental results demonstrate that the robustness of segmentation, the precision of line extraction and the efficiency of algorithm are improved.