Circular feature and non-coplanar point based monocular pose measurement

This paper proposed a novel monocular pose measurement method based on circular feature and non-coplanar point feature. The proposed method extracts the circular feature in the first frame using the Orientation Transform based circle detection algorithm (OTCircles). For the subsequent frames, a global optimal searching mechanism is adopted to track the circular feature. The circular feature tracking enables the proposed method to measure the pose efficiently and robustly for the image sequence. In order to eliminate the uncertainty of the recovered parameters of the circular feature, a non-coplanar point is introduced in the proposed method. The point feature is detected via template matching. Based on the detected circular feature and non-coplanar point feature, the proposed method measures the six-degree-of-freedom (6DoF) pose. Experiments on simulated data and real images indicate that the proposed method measures the 6DoF pose parameters efficiently and accurately.