Relative Pose Estimation Method for Unmanned Helicopter Landing Based on Vision

In the process of unmanned helicopter landing based on vision, a new pose estimation algorithm based on orthogonal iteration was proposed to solve the problem of relative pose estimation by selecting the angular feature of the landing image. Aiming at the requirements of accuracy, robustness and efficiency of pose estimation results, the objective space error was used as the error function. The initial value of iteration which affected the orthogonal iteration algorithm greatly was given based on the perspective projection model. Compared with the initial value which was random given, the number of iterations was reduced and the convergence speed was improved. The optimal value of spatial collinearity error was solved by absolute orientation method in the iterative process. The simulation results showed that the method has a good real-time performance, high precision and global convergence.