A Riemannian approach for free-space extraction and path planning using catadioptric omnidirectional vision

作者：

Highlights：

• We propose different Riemannian metrics to handle free space extraction and path planning problems using catadioptric vision

• The unified Riemannian framework allows to adapt the construction of the metric to the considered task for image processing

• It operates directly on the catadioptric image plane without need to unwrap or project the images to other representations

• We prove the efficiency of the proposed framework over classical Euclidean tools for processing catadioptric images

摘要

•We propose different Riemannian metrics to handle free space extraction and path planning problems using catadioptric vision•The unified Riemannian framework allows to adapt the construction of the metric to the considered task for image processing•It operates directly on the catadioptric image plane without need to unwrap or project the images to other representations•We prove the efficiency of the proposed framework over classical Euclidean tools for processing catadioptric images

论文关键词：Catadioptric vision,Riemannian metric,Geodesic distance,Free-space segmentation,Path planning

论文评审过程：Received 1 August 2019, Revised 20 December 2019, Accepted 29 December 2019, Available online 19 January 2020, Version of Record 1 February 2020.

论文官网地址：https://doi.org/10.1016/j.imavis.2020.103872