Orientation codes-based template matching method on workpiece detection

A new method for the workpiece image detection was proposed according to the characteristics of online work. The algorithm extracted orientation codes as a feature and consisted of two stages: rough matching and precision matching. Firstly, some points with the matching rate in a certain range were selected using histograms of orientation codes and rotation angles were computed roughly. In the second stage, the template was rotated in terms of the angle computed, matching based the orientation code between template and subimage was performed only for the points selected. Experiments result demonstrates the proposed method can work well even when workpiece was rotated or light fluctuates.