Research And Implement Of Data Segmentation Technology In Reverse Engineering

Reverse Engineering is one of the most active research points in CAD/CAM fields,butdata segmentation is a key point in reverse engineering and a difficulty problem.Datasegmentation as a basic process of surface reconstruction divides the data region accordingto the distributing feature of point cloud shape.It is meaningful to quicken the process ofreverse engineering research and application.This thesis gives an improved surface-base method of data segmentation base onexistent research achievements,first neighbor relationship of point cloud is built,thenanalyse the normal and curvature of the point cloud.Basing on the two steps above edgefeature points are picked up and surface type is analysed,at last the author realizes the datasegmentation through improved region-growing process.All the research an achievement sums up as follows:1. A improved k-neighbor search algorithm is designed,which improved theefficiency obviously.The author adopt "micro-tangent-plane" to calculate thenormal,and design a "local lamp-house" method to adjust the direction ofnormal,improved the efficiency obviously.2. Improved the surface-base method of data segmentation, using the edge pointwhich picked up by the edge-based method in region-growing process based onsurface classify, which improved the veracity and dependability.Besides, a prototypical system based on these algorithms is implemented in this thesis,and several typical data segmentation examples are given as well. The results of theseexamples indicate the correctness of the algorithms above.