A new way to detect the position and orientation of the wheeled mobile robot on the image plane

Abstract

In this paper, a new method to detect wheeled mobile robot's position and orientation is proposed. First, we paste a black rectangle board on the head of the robot as a mark, whose long side is parallel to the robot's orientation. And then on the image plane, we use CAMShift to track the robot after some pre-processing. We will get a search window which includes the target robot only by the use of the CAMShift algorithm. The search window's center is approximately identical to the robot's geometry center. Then we take appropriate approach (here we use the OpenCV function directly) to detect the angle between the black rectangle board's long side and the X axle of the image coordinate on the image plane, there we can get the robot's orientation.