Vehicle detection based on wheel part detection
- Resource Type
- Conference
- Authors
- Ruan, Yu-Sheng; Chang, I-Cheng; Yeh, Hung-Yu
- Source
- 2017 IEEE International Conference on Consumer Electronics - Taiwan (ICCE-TW) Consumer Electronics - Taiwan (ICCE-TW), 2017 IEEE International Conference on. :187-188 Jun, 2017
- Subject
- Bioengineering
Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Engineered Materials, Dielectrics and Plasmas
Engineering Profession
General Topics for Engineers
Geoscience
Nuclear Engineering
Photonics and Electrooptics
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Transportation
Wheels
Vehicle detection
Detectors
Feature extraction
Matched filters
Optical filters
Conferences
vehicle part
wheel detector
ROI segmentation
HOG
MB-LBP
- Language
This paper proposed an effective vehicle detection system based on wheel parts detection. The system is composed of two principal modules: feature detector construction and vehicle detection. Feature detector construction is to train a vehicle part model based on Adaboost using HOG and MB-LBP. Vehicle detection is formed by three sub-modules. ROI segmentation segments the searching region where wheels appear frequently, and this region is further divided into three sub-ROIs corresponding to three different aspect ratios. Wheel determination filters the outliers from the detected results in each sub-ROIs, and find the relationship between front wheels, back wheels and tail light parts. Vehicle localization focuses on localizing vehicles using those matched wheels. The experiments show that the proposed approach can offer good detection results under different environments.