Research on Multi-view 3D Image Local Region Restoration Algorithm based on Virtual Reality 

Hui Hu, Song Hu, Sibo Huang, Yingxue Cai, Jia Chen, Zhaoquan Cai* 

Huizhou University, Huizhou, 516007, China 

Abstract: This paper studies the local region restoration algorithm of multi-view 3D images, which lays a foundation for the analysis of image detail features. A multi-view 3D image local region restoration algorithm based on virtual reality is proposed. The texture information conduction model of multi-view 3D image is constructed. The corner matching detection of multi-view 3D image is realized by using Harris corner detec-tion algorithm. The adaptive texture segmentation method is used to segment the local region of the image, the information enhancement method is used to repair the local region of the image, and the virtual reality technology is used to realize the local region restoration of the multi-view 3D image. The simulation results show that the proposed method has strong information compensation ability and high peak signal to noise ratio (PSNR) for local region restoration of multi-view 3D images, and it improves the recognition ability of images.

Keywords: Virtual reality; Multi-view 3D Image; Repair; Enhancement