A novel method for point and segment identification of indoor environment based on laser radar

As a solution of indoor-environment feature extraction for mobile robot, a method employing laser sensor was proposed to extract corner points and line segments in the indoor environment. After establishing multi-dimensional data space in which the extreme points were searched, the feature of corner points was determined by testing the potential extreme points with hypothesis testing theory. In the polar coordinate system, the features of segments were first estimated via the probability model and finally obtained by curve-fitting with the weighted constrained least square method. The result of experiments indicates that the proposed method can effectively extract the features of points and segments in indoor environment.