Research On Flow Feature Extraction And Complicated Contour Matching Of Ect System

Currently, 2D image reconstruction based on electric field sensitivity theory could only reflect the material distributions in one sectional plane in some special time, but incapable of reflecting the flow pattern change details. The function is very limited. In this condition, we brought forward the 3D reconstruction method.Flow pattern features extraction and contour matching which will be discussed are very important in 3D surface rendering method. The accuracy of reconstructed image is influenced by the algorithm quality. Currently, shape feature extraction is widely used. It could only deal with the simple branch problem. For adaptation in the multi-branch problem, we brought forward the feature extraction in both shape and spatial relationship. Aiming at the low accuracy rate of traditional matching algorithms, a fuzzy pattern recognition method based on average distance between contours is proposed.To improve the efficiency of surface rendering method, this paper analyzed the space-time concurrency between all the nodes and then proposed the parallel image processing method. This method is originated from the pipelining theory.At last, for comprehensively compare the performance of Volume Rendering and Surface Rendering, this paper also gave a research to Volume Rendering method, and then gave a evaluation to the result.The main research contents and achievements are involved as follows:1. Optimized the 2D image pre-processing method, parallel processing is involved in some links.2. Simulated Annealing Algorithm model based on electric field sensitivity principle is proposed.3. A new feature extraction method based on both shape features and spatial relationship features is proposed.4. Use fuzzy set to resolve the contour matching problem.5. Use Matlab and Ansys to mesh the 3D pipes into finite elements.