Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace

Research Article

Location of feature points in 3D reconstruction of multi vision color image based on principal component analysis

Download370 downloads
  • @INPROCEEDINGS{10.4108/eai.27-8-2020.2295947,
        author={Huan-ping  FENG and Li-wei  ZHANG},
        title={Location of feature points in 3D reconstruction of multi vision color image based on principal component analysis},
        proceedings={Proceedings of the 13th EAI International Conference on Mobile Multimedia Communications, Mobimedia 2020, 27-28 August 2020, Cyberspace},
        publisher={EAI},
        proceedings_a={MOBIMEDIA},
        year={2020},
        month={11},
        keywords={keywords principal component analysis; multi vision color image 3d reconstruction; feature point location;},
        doi={10.4108/eai.27-8-2020.2295947}
    }
    
  • Huan-ping FENG
    Li-wei ZHANG
    Year: 2020
    Location of feature points in 3D reconstruction of multi vision color image based on principal component analysis
    MOBIMEDIA
    EAI
    DOI: 10.4108/eai.27-8-2020.2295947
Huan-ping FENG,*, Li-wei ZHANG1
  • 1: Heibei Institute Of Communication College
*Contact email: aini2125@tom.com

Abstract

Traditional image feature point location methods, due to the existence of calculation errors, lead to the accuracy of the location of feature points decreased, so based on principal component analysis, a new multi vision color image 3D reconstruction feature point location method is proposed.In this method, the gray level of color image is transformed by principal component analysis, and the color features of the image are obtained according to the gray level difference of local color areas;According to the changing characteristics of environment light coefficient and diffuse reflection light coefficient, the control parameters affecting the three-dimensional light shadow effect of multi vision color image are set;The least square method is used to eliminate the position error of feature points, and according to the confidence degree of association rules between nodes, more accurate feature point positioning results are obtained.Experimental results show that the accuracy of the proposed method is 11.27% higher than that of the traditional method. Therefore, the proposed method is more suitable for feature point location.