Three Dimensional Image Reconstruction Based on Normalized Convolution Algorithm of SIFT

With the rapid development of the information world, the diversification of information and diversity has become an inevitable trend, we perceive the mostly information from visual system, but three dimensional image is described by the whole vision system. The image data will be influenced by other factors such as noise or light in transmission process or conversion, it will inevitably result in image degradation and could not achieve our expected for 3D image, we make full use of the characteristics of the local stability characteristics for two-dimensional image through the reconstruction process of the 3D image, make normalized convolution of repairing algorithm, then re-establish of three-dimensional image.