Estimation of position and orientation of objects from stereo images

Abstract

This paper presents a method combining shape, shading and camera models in order to obtain estimations of 3D position and orientation directly from image grey values. The problem is considered as an application of optimal parameter estimation theory, according to Liebelt [9]. This theory has been applied previously, where the emphasis was laid on time-delay, Burkhardt [2], and motion estimation, Diehl [3], Houkes [7]. It is applied here to provide an environment in which somewhat more complicated models can be designed with relative ease and to indicate how the behaviour of the parameters can be investigated. A shading model is added, offering explicit prediction of image grey values. We consider a stereo camera setup. In a similar way image sequences have been incorporated by Korsten [8]. The resulting non-linear estimation problem is linearized about a last parameter guess [9], so that a linear estimator can be applied to compute a new estimate. The various stages of the modelling process are separated by introducing several coordinate systems. Coordinate transformations will show the object from other points of view, and perform a perspective projection of the 3D scene into the 2D image plane. The explicit grey value prediction yields a template, having a definite extent in the image. This method requires no gradient images, as in the case of estimating shape from motion [7] or stereo [6]. To demonstrate the usefulness and the flexibility of our method, we consider a solid cube, illuminated by a point light source. The image is a reflection image originating from the interaction of light with the cube surface.