The paper presents a novel method for estimating the 3D orientation of a camera in urban environments, leveraging the Manhattan world (MW) assumption. The main contributions include the definition of equivalence classes of equiprojective orientations, a new small rotation (SR) model to account for smooth camera motion, and decoupling the estimation of elevation and twist angles from the compass angle. The method uses a probabilistic sequential estimation framework based on an MW likelihood model, significantly reducing the search space for each orientation estimate. Experimental results demonstrate the effectiveness of the method in handling low-quality video sequences, even with many spurious edges. The approach avoids traditional steps like feature detection and correspondence, making it efficient and robust.The paper presents a novel method for estimating the 3D orientation of a camera in urban environments, leveraging the Manhattan world (MW) assumption. The main contributions include the definition of equivalence classes of equiprojective orientations, a new small rotation (SR) model to account for smooth camera motion, and decoupling the estimation of elevation and twist angles from the compass angle. The method uses a probabilistic sequential estimation framework based on an MW likelihood model, significantly reducing the search space for each orientation estimate. Experimental results demonstrate the effectiveness of the method in handling low-quality video sequences, even with many spurious edges. The approach avoids traditional steps like feature detection and correspondence, making it efficient and robust.