Evaluating the Impact of Semantic Segmentation and Pose Estimation on Dense Semantic SLAM

Abstract

Recent Semantic SLAM methods combine classical geometry-based estimation with deep learning-based object detection or semantic segmentation. In this paper we evaluate the quality of semantic maps generated by state-of-the-art class- and instance-aware dense semantic SLAM algorithms whose codes are publicly available and explore the impacts both semantic segmentation and pose estimation have on the quality of semantic maps. We obtain these results by providing algorithms with ground-truth pose and/or semantic segmentation data available from simulated environments. We establish that semantic segmentation is the largest source of error through our experiments, dropping mAP and OMQ performance by up to 74.3% and 71.3% respectively.

arXiv commentPaper accepted to IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2021

References in corpus (9)

• ORB-SLAM2: an Open-Source SLAM System for Monocular, Stereo and RGB-D Cameras
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Voxblox: Incremental 3D Euclidean Signed Distance Fields for On-Board MAV Planning
• CubeSLAM: Monocular 3D Object SLAM
• AI2-THOR: An Interactive 3D Environment for Visual AI
• Volumetric Instance-Aware Semantic Mapping and 3D Object Discovery
• 3DSSD: Point-based 3D Single Stage Object Detector
• BenchBot: Evaluating Robotics Research in Photorealistic 3D Simulation and on Real Robots
• The Robotic Vision Scene Understanding Challenge