Constrained spaces like doorways make efficient navigation using a wheelchair difficult and hazardous, especially in the cases where the user suffers from motor impairments. This paper introduces a monocular vision based autonomous framework for the fundamental task of doorway passing in an electric wheelchair. We propose a novel Lyapunov-based visual control process which is used to generate a smooth trajectory about a single doorpost. The scheme relies on line features which represent the doorposts in an image to calculate the wheelchair motion. This framework ensures that the wheelchair is able to pass the doorway regardless of its starting position. Results of experiments on a robotic wheelchair show that the system is able to perform robustly in different corridors with a variety of door representations.