Application of moment invariants to visual servoing

Abstract

Abstract-In this paper, we present how moment invariants can be used to design a decoupled 2D visual servoing scheme and to minimize the nonlinearity of the interaction matrix related to the selected visual features. Experimental results using a 6 dof eye-in-hand system to position a camera parallel to planar objects of complex shape are presented to demonstrate the efficiency of the proposed method.

Point-based and region-based image moments for visual servoing of planar objects

“…For dense objects, image moments are efficiently computed from the contour points using Green's theorem [26]. In all the experiments, and as in [28], we have used the following model of the interaction matrix in the control law (29):…”

Image moments: generic descriptors for decoupled image-based visual servo

“…we obtain after tedious developments the interaction matrix L µij and we can check it is slightly different to that obtained in the continuous case (see [9]):…”

“…Recent developments are concerned with the selection or the transformation of visual data to design visual features that improve the behavior of the system [2], [6], [7], [9]. The objective of these works is to avoid the potential problems that may appear using basic techniques (typically reaching local minima or a task singularity) [1].…”

“…In our previous works [9], a decoupled control has been proposed using the image moments computed from a continuous objet. Six combinaitions of moments were proposed to obtain an adequate behavior.…”

“…Furthermore, up to now, the decoupling properties have been obtained only when the object is parallel to the image plane [9]. In this paper, we propose an original method to deal with the more general case where the desired object position may have any orientation with respect to the camera frame.…”