Abstract: To extract sharp features from scattered point cloud sampled from piecewise smooth surfaces,a robust feature detection method using local reconstruction is proposed in this paper.First,for each point,a weight which measures the feature likelihood of a point is assigned according to a covariance analysis on its local neighborhood.By threshold filtering,the initial feature points are detected.Then,in the local neighborhood of each initial feature point,a triangle set is constructed,which effectively reflects the local feature structure.Subsequently,by applying the shared nearest neighbor clustering algorithm on the normal of triangles,we can obtain the clusters of points in the local neighborhood.Finally,for points of each cluster,one plane is fitted.Based on the fitted plane,the initial feature point is further identified as a true feature point,if it is nearly locating on the intersection of multiple fitting planes.Experimental results show that our method is simple,stable and insensitive to the size of selected neighborhood and robust to noise.