A Probabilistic Approach to Surface Extraction from Range Data

In this research, we proposed a new method to model complex range data using a parametric model. We first compute local surface curvatures by quadratic fitting of local surfaces, and then extract the curved surfaces and plane regions from the range data by examining the probability distribution of local surface curvatures. After extracting these regions, one shape may be divided into multiple regions by the characteristics of the feature space. Then, after the parameters of each region are estimated, the fitting error between regions of the integration candidate is calculated. Finally, by applying statistical testing techniques on the fitting error, we establish integration criteria and integrate the regions that satisfy this criteria. From this process, it is possible to perform a robust surface extraction. Experimental results on real range data show the effectiveness of the proposed method.