Research Of3D Image Reconstruction Based On Clustering Preprocessing

Three-dimension (3D) image reconstruction is important in the research field of computer visualization technology. The technology of3D reconstruction is widely used in the field of military affairs, spaceflight, industry, medical science and so on. Visualization technology plays an indispensable role in various fields driving the improvement and perfection of the3D image reconstruction.This paper firstly discussed and summarized the3D reconstruction algorithm. The classical Marching Cubes algorithm in surface rendering is introduced, as well as voxel and iso-surface. The clustering method of ladder edge surface is also particularly described.When processing large data with traditional3D reconstruction algorithm, the slow operating speed will be obvious. Present researchers have done little to increase the reconstruction speed. This paper put forward the clustering algorithm to preprocess3D image so as to increase the3D reconstruction speed. The two preprocessing clustering algorithms were brought into3D image, K mean value algorithm and C mean value algorithm. Clustering is to classify research targets according to specific principles and requirements. The basis of clustering is a similarity measurement method of research targets without any other categories or experiences. So it can be called no-supervision classification. Clustering algorithm can get new clustering centre point after iteration. The size of target image can be calculated according to the new clustering centre point, then separate the3D data field of target image. Reconstructing3D image in smaller3D data field can largely increase the speed of3D image reconstruction.Experiment results show that the reconstruction time decreased largely after preprocessed by clustering compared with those without preprocessing. The proposed algorithm is proved correct and efficient and is also extended to some other3D biology medical image research field.