The development of a vision based orientation sensor for the measurement of the orientation of a spherical body is presented. A specially designed grid pattern used to encode absolute orientation information onto a spherical surface in terms of latitude and longitude angles is discussed. A novel algorithm for the analysis of images of this grid pattern based on tracking along the grid lines is also developed that is designed to overcome limitations of other traditional image analysis techniques in terms of speed and accuracy. A procedure for recovering the absolute orientation of a spherical surface encoded with the special grid pattern using the line tracking algorithm, camera calibration, and Horn’s (1987) technique is also presented. Finally, experimental results of this procedure applied to a sequence of images are presented and discussed.