The concept of depth from focus involves calculating distances to points in an observed scene, by modelling the effect that the camera's focal parameters have on images acquired with a small depth of field. This technique does not require special scene illumination, and needs only a single camera. This thesis provides a background understanding of the concept and theory of depth from focus, surveys the literature on current methods of obtaining depth from focus, analyzes the key problems, and presents a novel solution to the problem, complete with experimental results. Deconvolving and modelling the defocus operator, is the most difficult segment of calculating depth from focus. Isolating the defocus operator has conventionally been performed by taking local spatial regions, and inverse filtering in the spatial frequency domain. This thesis exposes some fundamental problems with this method: inaccuracies in finding the frequency domain representation and the presence of border effects. To solve the general depth from focus problem, a novel application of an iterative matrix based method is presented. This method uses two images of the same scene, obtained under different conditions of defocus. The defocus operator may be assumed using a parametric model, or experimentally measured. Trade-offs in implementation are resolved through regularization. The method is theoretically justified, and shown to eliminate the problems mentioned above. A constrained inverse filtering method and the author's iterative matrix based method are experimentally implemented on four scenes. The experiments show the iterative matrix solution consistently yielding more accurate results. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate