Research On The Algorithms Of Point Cloud Boundary Extraction And Corner Detection

Scale measurement is one of the basic means which are used to understand and reform the world by mankind. Advantage of the measuring method based on Reverse Engineering is that it acquires all external imformation through conversing surfacial imformation of object to discrete corordinate points. Product development cycle and precision can be greatly shortened by this method. According to practice of Modern Shipbuilding Huge Size 3-D Numeral Measurement System, such RE key technologies to hull section model are researched.1. Preprocess the images collected, such as image filter and image segmetation. A new centre-extracted method that combines genetic algorithm with extremum method is proposed based on analyzing and summarizing the current algorithms.2. Denoising is implemented to the gained point cloud adopting a method based on adaptive threshold.3. Geographic scanning and method based on homogeneity data points and their surrounding points distributed are implemented to spacial flat point cloud. A new characteristic boundary extraction method based on mesh generating method combines with geographic scanning is proposed and verified.4. Method based on curvature boundary is impletmented. A new adaptive corner detection method based on boundary is proposed and its feasibility is verified according to features of hull section point cloud.