Search papers, labs, and topics across Lattice.
This paper introduces a novel iterative method for relative pose estimation between two calibrated cameras, leveraging Powell's Dog Leg algorithm to achieve significant speed improvements. The new approach retains the precision of Nister's state-of-the-art five-point method while being approximately twice as fast, making it particularly effective for real-time applications. Additionally, it can be extended to handle more than five points, enhancing its utility as a refinement step in Structure and Motion tasks.
Achieving double the speed of the leading five-point pose estimation method without sacrificing precision could revolutionize real-time camera tracking applications.
Robust estimation of the relative pose between two cameras is a fundamental part of Structure and Motion methods. For calibrated cameras, the five point method together with a robust estimator such as RANSAC gives the best result in most cases. The current state-of-the-art method for solving the relative pose problem from five points is due to Niste虂r [9], because it is faster than other methods and in the RANSAC scheme one can improve precision by increasing the number of iterations. In this paper, we propose a new iterative method, which is based on Powell's Dog Leg algorithm. The new method has the same precision and is approximately twice as fast as Nister's algorithm. The proposed method is easily extended to more than five points while retaining a efficient error metrics. This makes it also very suitable as an refinement step. The proposed algorithm is systematically evaluated on three types of datasets with known ground truth.
