• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

Computer Engineering & Science

Previous Articles     Next Articles

A binocular stereo-view 3D reconstruction algorithm
based on contour extraction and depth screening

MA Jianshe,WEI Yunfeng,SU Ping   

  1. (Graduate School at Shenzhen,Tsinghua University,Shenzhen 518055,China)
  • Received:2016-05-18 Revised:2016-11-23 Online:2018-04-25 Published:2018-04-25

Abstract:

Aimming at the phenomeon that the 3D model, which is obtained by the 3D reconstruction algorithm based on depthmapfusion, is easily influenced by depth information error, a binocular stereoview 3D reconstruction algorithm based on contour extraction and depth screening is proposed. A standard checker board is adopted to calibrate the binocular 3D reconstruction system. Canny operator is used to detect the boundary of the target object, and morphological corrosion andexpansion methods are comprehensively used to extract the continuous boundary in a specified direction.On this basis,the depth information of the target object is filtered and interpolated in order to calculate the continuous depth value.The results show that, compared with the conventional 3D reconstruction algorithm, the target object of this algorithm has higher surface integrity, and the background hot pixels surrounding the target object are removed.

Key words: 3D-renconstruction, binocular stereoview, depthmap, contour extraction, interpolation