Multi-sensor identity tracking with event graphs

The ability to track moving objects is a key part of autonomous robot operation in real-world environments. Whilst for many tasks knowing the positions of objects may be sufficient, tracking the identity of targets may also be desirable. When objects are well separated preserving identities is trivial, however, the identities of objects that pass close to one another may become confused. This paper considers methods to maintain the identities of tracked objects using a combination of LIDAR and video data. When objects are well separated, they are tracked using location information from the LIDAR. When objects move together and their identities cannot be resolved, interactions are recorded and later resolved using appearance models. A vision based approach is adapted for use with LIDAR data and a new method for identity reasoning is proposed. The methods are validated on a dataset comprising a total of 37906 manually labelled point cloud segments.

Authors

• Peter Morton
• Bertrand Douillard
• James Patrick Underwood

Keywords

• Target tracking
• Laser radar
• Cameras
• Cognition
• Trajectory
• Image color analysis
• Identity Tracking
• Event Graph
• Point Cloud
• Video Data
• Real-world Environments
• Object Identification
• Manual Labeling
• Lidar Data
• Point Cloud Segmentation
• Training Set
• Support Vector Machine
• Centroid
• Tracking System
• 3D Space
• Root Node
• Nodes In The Graph
• Tracking Performance
• Image Patches
• Appearance Features
• Similar Color
• Track Section
• Point Cloud Data
• Objects In The Scene
• Number Of Scenarios
• Color Histogram
• Appearance Information
• Pedestrian Walking
• Complete Trajectory
• Camera Data
• Example Of Graph