DOT: Dynamic Object Tracking for Visual SLAM

In this paper we present DOT (Dynamic Object Tracking), a front-end that added to existing SLAM systems can significantly improve their robustness and accuracy in highly dynamic environments. DOT combines instance segmentation and multi-view geometry to generate masks for dynamic objects in order to allow SLAM systems based on rigid scene models to avoid such image areas in their optimizations. To determine which objects are actually moving, DOT segments first instances of potentially dynamic objects and then, with the estimated camera motion, tracks such objects by minimizing the photometric reprojection error. This short-term tracking improves the accuracy of the segmentation with respect to other approaches. In the end, only actually dynamicmasks are generated. We have evaluated DOT with ORB-SLAM 2 [1] in three public datasets. Our results show that our approach improves significantly the accuracy and robustness of ORB-SLAM2, especially in highly dynamic scenes.

Authors

• Irene Ballester
• Alejandro Fontán
• Javier Civera 0001
• Klaus H. Strobl
• Rudolph Triebel

Keywords

• Geometry
• Image segmentation
• Visualization
• Simultaneous localization and mapping
• Motion segmentation
• Heuristic algorithms
• Dynamics
• Object Tracking
• Dynamic Objects
• Dynamic Tracking
• Visual Simultaneous Localization And Mapping
• Dynamic Object Tracking
• Instance Segmentation
• Rigid Model
• Dynamic Scenes
• Camera Motion
• Neural Network
• Visual System
• Semantic Segmentation
• Tracking Error
• Object Motion
• Motion State
• Posterior Mode
• Objects In The Scene
• Segmentation Errors
• Differential Entropy
• Projection Point
• Camera Pose
• Visual Odometry
• Pose Tracking
• Track Quality
• Subset Of Pixels
• Mask R-CNN