Encode–decode network with fully connected CRF for dynamic objects detection and static maps reconstruction

作者：

Highlights：

• An encode–decode network is proposed to calculate the dynamic probability.

• A 3D fully connected CRF structure is present for the soft segmentation of network.

• An EDN-CRF octomap is structured for static maps.

摘要

•An encode–decode network is proposed to calculate the dynamic probability.•A 3D fully connected CRF structure is present for the soft segmentation of network.•An EDN-CRF octomap is structured for static maps.

论文关键词：Robot vision,Encode–decode network,CRF,Dynamic objects detection,Static maps reconstruction

论文评审过程：Received 15 August 2018, Revised 5 June 2020, Accepted 4 March 2021, Available online 16 March 2021, Version of Record 9 April 2021.

论文官网地址：https://doi.org/10.1016/j.image.2021.116237