3D Reconstruction of Multi-view Images Based on Modificated Histogram Equalization

Huan-xi XU, Wen-hua QIAN, Dan XU, Yuan-yuan PU, Guo-wu YUAN

Abstract


Aiming at the unsatisfactory effect of three-dimensional reconstruction of multi-scale image set and the fact that the reconstruction effect of large-scale scenes such as buildings is not realistic and the details are not prominent, this paper proposes a new image enhancement algorithm to preprocessed images, and uses the floating scale surface reconstruction algorithm to improve the surface reconstruction effect. Firstly, an image enhancement algorithm is proposed, which is based on histogram equalization. Secondly, the Structure from Motion algorithm is used to recover camera parameters, extract and match features, and generate sparse point clouds. And then,the MVS algorithm is used for dense reconstruction. Finally, the floating scale surface reconstruction algorithm is used for reconstruction. The experiment proved the proposed image enhancement algorithm improves the effect of feature point extraction. The surface reconstruction algorithm we use is efficiently for surface reconstruction and improves the 3D reconstruction effect.

Keywords


3D reconstruction, Image enhancement, Structure from motion, Point cloud reconstruction, Surface reconstruction


DOI
10.12783/dtcse/msam2020/34247

Full Text:

PDF

Refbacks

  • There are currently no refbacks.