Abstract

For three-dimensional (3D) measurement of a large object, stereo matching needs to be done under a relatively long baseline in order to reduce the accumulated errors caused by the splicing process, but this brings new challenges to matching efficiency. The images under a short baseline have an advantage avoiding a very large disparity level that is usually presented in a large baseline to improve efficiency, whether the image pairs to be matched are of translation or rotation configuration. This paper mainly focuses on the general case existing for both, and the purpose is to efficiently execute a stereo matching in this scene based on short baseline estimation. A rectification method of three images used for baseline estimation is first studied to make the image planes become coplanar, which removes the vertical disparity. The three rectified images can form a stereo pair with a short baseline and the other one has a long baseline. Then the relationship that exists between disparities of a pixel on rectified stereo pairs with different baselines is deduced by considering the quantization error introduced in the projection and rectification. In this way, the disparity search range of a pixel on the rectified stereo pair with the long baseline is reduced from the usual hundreds of pixels to a few pixels predicted by its disparity with the short baseline, which greatly improves the calculation efficiency and makes it especially suitable for the 3D measurement of large objects. Experimental results are presented to demonstrate the efficiency and accuracy of the method.

© 2021 Optical Society of America

Collaborative semi-global stereo matching

Penghui Bu, Hong Zhao, Jiaxing Yan, and Yusheng Jin
Appl. Opt. 60(31) 9757-9768 (2021)

Multi-brightness layers with a genetic optimization algorithm for stereo matching under dramatic illumination changes

Tao Yang, Rui Yang, and Yuehong Qiu
Appl. Opt. 60(24) 7371-7380 (2021)

Four-direction global matching with cost volume update for stereovision

Yusheng Jin, Hong Zhao, and Penghui Bu
Appl. Opt. 60(19) 5471-5479 (2021)

Occlusion disparity refinement for stereo matching through the geometric prior-based adaptive label search

Junwei Wang, Wei Zhou, Ziheng Qian, Jiaqi Shen, and Hanming Guo
J. Opt. Soc. Am. A 38(11) 1619-1630 (2021)

Binocular stereo matching of real scenes based on a convolutional neural network and computer graphics

Liaoyu Kou, Kai Yang, Lin Luo, Yu Zhang, Jinlong Li, Yong Wang, and Liming Xie
Opt. Express 29(17) 26876-26893 (2021)