Statistical Property of Point Cloud Helps to Its Reduction and Denoising

In the realm of reverse engineering, a new approach for analyzing and processing the scattered point cloud was proposed. With the high frequency of laser scanner in data sampling contributing to a raw point cloud of high density, statistical method is applicable in that the local expectation of points could be estimated from its mathematical distribution. In calculating the expectation, as different from earlier research, a confidence region was employed to discard noisy points, by which eliminate errors possibly brought in other means. A case study was conducted in justifying the validity of this method, and software such as SPSS and Solidworks are adopted in the statistical analyzing and CAD modeling of a mouse. Results show the precision and efficiency of this application, which generally could serve as a promise idea in the processing of point could.