Alternative Vision Approach to Ground Vehicle Detection System Utilizing Single Board Computer (SBC) for Motor Control

Abstract

- This paper presented an alternative approach for ground vehicle identification for vehicle following. The vision system is tested for its feasibility in image processing on a limited resources platform. The limited resources platform consists of off-the-shelf webcam as vision sensor, single board computer (SBC) as its main hardware and a microcontroller as a sensor board. The image processing library used for image processing is OpenCV, an open source library. For optimization purposes Design of Experiment (DOE) is used to determine the factors that contribute to the accuracy of the vehicle identification. This system is then attached serially to ultrasonic sensor in order to demonstrate the safety system for the follower vehicle. Once integrated SBC will decide the action of the motor of the RC car. The experimental results reveal that the approach under limited resources platform is able to identify the needed features to indicate presence of a ground vehicle to some extent. The paper also highlights the limitations of the system which may be addressed in future works.