A point cloud data reduction method based on curvature

As a non-contact-type device could sample part surface data with high speed and accuracy, it becomes the most popular instrument for capturing the surface data of a part. However, it creates a large amount of point data which must be reduced to decrease computational time and to lower the storage requirement. Aiming at the limitations of point cloud data reduction methods developed in the past, a new reduction method based on curvature is proposed in this paper. It includes searching k-nearest neighbors for constructing data topology, calculating and adjusting tangent plane normal, estimating the curvature by using paraboloid fitting method, and setting the principles of data reduction. The experimental results show that the new method reduces the number of points significantly while preserving the geometry characteristics perfectly.