The initial development and assessment of an active computer vision system is described, which is designed to meet the growing demand for 3 dimensional models of real-world objects. Details are provided of the hardware platform employed, which uses a modified gantry robot to manoeuvre the system camera and a purpose-built computer controlled turntable on which the object to be modelled is placed. The system software and its computer control system are also described along with the occluding contour technique developed to automatically produce initial models of objects. Examples of models constructed by the system are presented and experimental results are discussed, including results which indicate that the occluding contour technique can be used in an original manner to identify regions of the object surface which require further modelling and also to determine subsequent viewpoints for the camera.