A comparison of extended object tracking with multi-modal sensors in indoor environment

TL;DR

Experimental results show that the object tracking method using a stereo camera achieves performance similar to that of a LiDAR sensor, with a cost difference of more than tenfold.

Abstract

This paper presents a preliminary study of an efficient object tracking approach, comparing the performance of two different 3D point cloud sensory sources: LiDAR and stereo cameras, which have significant price differences. In this preliminary work, we focus on single object tracking. We first developed a fast heuristic object detector that utilizes prior information about the environment and target. The resulting target points are subsequently fed into an extended object tracking framework, where the target shape is parameterized using a star-convex hypersurface model. Experimental results show that our object tracking method using a stereo camera achieves performance similar to that of a LiDAR sensor, with a cost difference of more than tenfold.

Figures (5)

• Figure 1: Point cloud from Cube
• Figure 2: TurtleBot 4
• Figure 3: Point cloud from D435i
• Figure 4: Detected robot points (blue) in point clouds of two sensors.
• Figure 5: Tracking results of two motion trajectories with Cube LiDAR and D435i camera. The coordinate in each image is on a Cartesian plane of the corresponding sensor's local coordinates. The light blue curve in each image represents the motion trajectory of the estimated centroid. Four representative timesteps' results are visualized in each image: the blue dashed line represents the estimated shape, the blue dot represents the centroid, and the black circles represent the sensor measurements.