Experiment on Real-Time Image Processing in the Controlling of Mecanum Wheel Robotic Car

Abstract

The control method of the mecanum wheel robotic car is accomplished by computer vision. The camera attached to the computer is for the purpose of image processing with open source computer vision (OpenCV). It can capture the locations of the wheel robot and the target by the use of open source computer vision as pattern recognition scheme. By using the Oriented FAST and Rotated Brief (ORB) feature, the program finds the patterns. The locations of the target and the robot are obtained by K-Mean Clustering method. And then, the programming method is used to calculate their location difference to generate a movement command. The movement command is then sent to the microcontroller of wheel robot system by means of wireless communication using NRF radio modules. By doing so, the robotic car goes to target with the command of computer vision program. The experimental results confirmed that the performance of the control algorithm could be utilized for the real-time condition.