Pairwise registration in indoor environments using adaptive combination of 2D and 3D cues

作者：

Highlights：

• We propose a novel pairwise coarse registration framework for RGB-D data.

• It autonomously combines 2D and 3D features according to the scene.

• The combination is dictated by structure and texture in the scene.

• We evaluate our framework with realistic datasets.

• Results for indoor SLAM scenarios are quite promising.

摘要

•We propose a novel pairwise coarse registration framework for RGB-D data.•It autonomously combines 2D and 3D features according to the scene.•The combination is dictated by structure and texture in the scene.•We evaluate our framework with realistic datasets.•Results for indoor SLAM scenarios are quite promising.

论文关键词：Pairwise registration,RGB-D data,Local descriptors,Keypoint detectors

论文评审过程：Received 8 September 2016, Revised 13 June 2017, Accepted 28 August 2017, Available online 7 September 2017, Version of Record 6 February 2018.

论文官网地址：https://doi.org/10.1016/j.imavis.2017.08.008