Depth Estimation from a Sequence of Images

5.1 INTRODUCTION This chapter addresses the problem of depth estimation using a sequence of images. In the previous chapter, we have eliminated the correspondence problem by taking two specific orthogonal views. In this chapter, this limitation is sought to be removed by employing point correspondence. The equation of motion derived from spherical projection will be reformulated so as to apply Kalman filtering for its solution. Most of the literature dealing with structure determination from one or two views is discussed in the previous chapter. Here, we highlight some contributions on structure determination from a sequence of monocular images. The problem of analyzing sequence of monocular images (intensity or range) or stereo images to extract three dimensional motion and structure is an active area of research in computer vision. While analyzing monocular images reliable tokens (features) such as curves, comers are detected from the spatial variation of image intensities, assuming that they correspond to markings on 3D objects. Second, these tokens are tracked over time to recover depth and 3D velocities of the corresponding 3D tokens. One possibility is to consider two-dimensional tracking which gives us matches between different frames and to use these matches for estimating three dimensional motions. For some of the work published for recovering motion and structure from n point matches, p line matches, between q views where typically n is 5, pis 6 and q is 2 or 3 refer to [Huang et al. Three dimensional tracking is the other possibility [Zhang et al., 1988]. A review on the computation of motion from a sequence of images is 87