Efficient and robust unsupervised inverse intensity compensation for stereo image registration under radiometric changes

作者：

Highlights：

• This paper proposes a new 5-D vector cost metric structure, which is more robust to radiometric changes than the traditional methods.

• A novel unsupervised inverse intensity compensation method based on 5-D vector for stereo images is proposed in this paper.

• Additional stereo pairs with intensity compensation can be obtained, which can guide the subsequent image processing.

• Experiments performed on stereo images of the Middlebury datasets and ours demonstrate this methodology in calculation accuracy and time all have good performance.

摘要

•This paper proposes a new 5-D vector cost metric structure, which is more robust to radiometric changes than the traditional methods.•A novel unsupervised inverse intensity compensation method based on 5-D vector for stereo images is proposed in this paper.•Additional stereo pairs with intensity compensation can be obtained, which can guide the subsequent image processing.•Experiments performed on stereo images of the Middlebury datasets and ours demonstrate this methodology in calculation accuracy and time all have good performance.

论文关键词：Image registration,5-D vector,Photometric changes,Intensity compensation,RGB-Gradient tree

论文评审过程：Received 29 January 2020, Revised 21 August 2020, Accepted 29 October 2020, Available online 5 November 2020, Version of Record 19 November 2020.

论文官网地址：https://doi.org/10.1016/j.image.2020.116054