Monocular vision for pose estimation in space based on cone projection

Abstract

This paper presents a three-dimensional (3-D) pose estimation algorithm based on monocular vision. The algorithm relies on the circle target whose radius is known, with the scale condition given, the depth information of the circle can be recovered incompletely, and finally the pose of the target can be estimated by single projection only. First, the circle target was mapped to be an upright elliptic cone in the pinhole imaging model. Second, radius constraint was applied to recover partial depth of the circle target based on the upright elliptic cone. Experimental work concerning the validity and accuracy of this method is presented; furthermore, an application case for robot-aided positioning is introduced.