Research on 3D Reconstruction Based on a Single Image

Abstract:

Aimed at the fact that most of the objects was symmetrical, we present a method reconstructing 3D model from single 2D image. At first, the angle, inclination angle and pivot angle should be established in the ternary perspective transformed matrix T, then the calibration which indicated the outline of the object can pick up the characteristic line. It resolved the host extinguishes information in accordance with the similar features of the parallel group of lines projection angle, then ensured the viewpoint situation and the object plane of symmetry, and made use of the object imagination plane of symmetry in accordance with symmetry feature through the interactive alignment specifying three pairs of the known symmetry image coordinates and the space coordinates of the corresponding points confirm the perspective transformative matrix T, then reverse prove the other space coordinates of the surface of the objects.

• Periodical: Advanced Materials Research (Volumes 108-111)
• Pages: 3-10
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.108-111.3
• Online since: May 2010
• Authors: Yong Yan Yu, Zhi Jian Wang
• Keywords: Perspective Inverse Transform, Reverse 3D Coordinate, Symmetrical Object, Viewpoint Position

References

[1] JIN Xi zi: Computer Image Processing(Science Publishers, China, 2004).

[2] HE Dong jian. Digital Image Processing(Xidian University Publishers, China, 2003).

[3] SUN Jia guang. Computer Graphics(TSingHua University Publishers, China, 1998).

[4] SUN Han, REN Ming wu, YANG Jing yu. A Fast and Practical Algorithm for Line Detection. Computer Application Research, 2006, 3(2): 256-260.

[5] SHU Yuan, Tan Zheng. Newautomatic vanishing point detection algorithm. Infrared and Laser Engineering, 2003, 32(5): 479-483.

[6] CHEN Ningfan, Cai Lidong. An vanishing points detection algorithm based on clustering. Chinese Journal of Serology and Image Analysis, 2006, 10(1): 205-210.