A morphology-based approach for three-dimensional reconstruction and notion estimation

Computer vision, or the application of computer-based image analysis and interpretation, is an emerging technology. Computer vision systems have been used in automated processing, analysis and understanding of images in many industrial applications. 3D object reconstruction and motion estimation present many interesting challenges in such applications. Over the last few decades, many good schemes have been developed to deal with these problems. In this thesis, a new scheme for estimating 3D profile and motion parameters of objects in an unstructured environment is presented.; The scheme takes advantage of the morphology formulation at both the low-level image processing (image preprocessing) as well as high-level image processing (e.g., 3D object reconstruction and motion estimation). A new morphology-based transform is introduced to estimate motion parameters for both rigid and flexible objects. Furthermore, the thesis will discuss how the proposed scheme is extended so that it can deal with the 3D object surface reconstruction problem with limited surface measurements.; A number of experiments are conducted to examine the proposed scheme and compare its performance to a commonly used technique for motion estimation. Based on the experimental results, the advantages and disadvantages of this scheme are discussed.