Multi-class Obstacle Identification using Shape Descriptors in Video for Autonomous Vehicles
Shumin Liu, Xujuan Xu, Longyun Duan
Available Online June 2017.
- https://doi.org/10.2991/icmia-17.2017.135How to use a DOI?
- Obstacle identification, Shape descriptor, Shape context, Axis of Least Inertia
- Obstacle identification has been widely studied as part of the broader obstacle detection research area for Autonomous Vehicles (AV). Existing in-vehicle sensing systems are concentrated on obstacle detection for pedestrian or vehicle, and limited work has been conducted on multi-class obstacle classification. In the process of obstacle identification, the selection of classification features is particularly critical. As a set of features to describe a given shape or contour, shape descriptor has attracted much attention in recent years and play an important role in pattern recognition. This paper proposed a shape descriptor based multi-class obstacle identification method where the traffic obstacles (in the front of self-vehicle) be classified into one of four classes: vehicle, lateral moving pedestrian, longitudinal moving pedestrian, and unknown (such as trees, road lamp, barricade etc.). Here a variety of shape descriptors extracted from the contour curve are involved, such as Rectangularity, Compactness, Elongation, Circularity, Shape Context, and Axis of Least Inertia. Finally, the identification results using these descriptors are contrastive analyzed. Though a single shape descriptor does not achieve ideal identification results for traffic obstacle, but this will provide a new idea for multi-class obstacle identification using shape descriptor in video for AV.
- Open Access
- This is an open access article distributed under the CC BY-NC license.
Cite this article
TY - CONF AU - Shumin Liu AU - Xujuan Xu AU - Longyun Duan PY - 2017/06 DA - 2017/06 TI - Multi-class Obstacle Identification using Shape Descriptors in Video for Autonomous Vehicles BT - 2017 6th International Conference on Measurement, Instrumentation and Automation (ICMIA 2017) PB - Atlantis Press SN - 1951-6851 UR - https://doi.org/10.2991/icmia-17.2017.135 DO - https://doi.org/10.2991/icmia-17.2017.135 ID - Liu2017/06 ER -