This paper presents a novel, efficient, vision-based technique to estimate the attitude of Powered Two-Wheeled Vehicles (P2WV) with respect to the road surface. This method is based only on a monocular camera fixed on the vehicle. The estimation is done in two steps. The first step relies on a novel technique that utilizes perspective distortion and Inverse Perspective Mapping to measure the sub-pixel width difference of the lane markers and estimate the roll relative angle. The second step uses the vanishing point technique from the lane markers to estimate the yaw and the pitch. These last are then converted to the same Euler convention, and the rotation matrix of the camera is recovered. This permits estimating the vehicle attitude. The presented method was tested on several scenarios with different high speeds and road shapes using the BikeSim framework. The results show a very low error of estimation and robustness against camera pose change. The main advantage of this method is that the estimation is not accumulative, thus eliminating the estimation drift error over time.