QuEst: A Quaternion-Based Approach for Camera Motion Estimation From Minimal Feature Points

In this letter, we consider the problem of recovering the rotation and translation changes of a moving camera from captured images. This problem is traditionally solved using the epipolar constraint, where the rotation and translation changes are recovered from the essential matrix. We propose a new formulation based on quaternion representation of rotation. Using this formulation, we recover the rotation and translation changes separately without resorting to an essential matrix. We compare the estimation accuracy and execution time of the proposed method with several state-of-the-art algorithms using both synthetic and actual image sequences. On average, our approach shows better accuracy in the presence of noise, but it has a larger execution time. For the benefit of community, we have made the implementation of our algorithm available online and free.