Extracting Characteristic Points of the Surfaces without a Support Threshold

It is a challenging problem to find the characteristic points without a support threshold in computer aided design. The previous works are focusing on the extracting characteristic points with a threshold, but how to initialize a reasonable threshold depends on the experiment other than scientific theory. So a novel approach is presented to extract the characteristic points of the combined surfaces without a threshold based on outlier detection algorithm. First of all, the surface domain and its depth are defined according to the combined surfaces´ characteristic and then R-tree is constituted. The outlier detection algorithm was employed to extract the characteristic points according to the normal unit. This approach has been applied successfully to the reverse design for the parts of a motorcycle. The experiment indicates that the characteristic points can be extracted accurately to prove the approach effective, and the same time meet the requirement of real-time