Algorithm design for a position tracking sensor based on pattern recognition

In this paper, a pattern recognition based algorithm is developed to solve position tracking problems. In particular, the algorithm is applied to finding two dimensional position deviations along X and Y axes and to finding the rotating angle of objects through comparing the shape numbers between the saved template and the pictures of static objects captured by a ´vision´ element such as a CCD camera, while the shape numbers are obtained from critical points which are determined by using the so-called directional-flow-change (DFC) method. It is noted that the size of the captured images can be reduced in our design in order to accommodate real-time application with reasonable accuracy and reliability. The algorithm is implemented and tested in a CCD camera based position tracking sensor for a real sheet of foam barrier. The experiment results show that the sensor performs very well in obtaining the pattern and finding the deviations and hence could be potentially applied to the real assembly manufacturing process.