Android-Based Patrol Robot Featuring Automatic Vehicle Patrolling and Automatic Plate Recognition

Chian C. Ho, Shih-Jui Yang, Jian-Yuan Chen, Chang-Yun Chiang, Hsin-Fu Chen
2017 Journal of Robotics, Networking and Artificial Life (JRNAL)  
This work develops an Android-based patrol robot featuring Automatic Vehicle Patrolling (AVP) and Automatic Plate Recognition (APR). The AVP feature integrates 3 novel methods, wheel-wheelcover-based AdaBoost wheel detection, contour-wheel-oriented vehicle approaching, and Ad-Hoc-based remote motion control. The APR feature integrates 4 novel methods, Wiener-deconvolution vertical edge enhancement, AdaBoost plus vertical-edge plate detection, vertical-edge horizontal-projection
more » ... ation stain removal, and customized optical character recognition. Implementation results show the vehicle detection rate and plate recognition rate of the Android-based patrol robot are over 96% and over 94%, respectively, under various scene conditions. On the other hand, the average execution time of AVP and APR of the Android-based patrol robot takes at most 8 second per round and at most 0.8 second per frame, respectively.
doi:10.2991/jrnal.2017.3.4.13 fatcat:5o7wtbhttnhd5m3mrvzcje3adi