A Pose Estimation Method for Indoor UAV Based on Binocular Vision

As a means of fast transportation of goods, indoor unmanned aerial vehicles (UAVs) are limited in their widespread use in express logistics warehouses due to the complexity of their attitude estimation and navigation positioning, as well as the relatively expensive sensors they carry. This paper focuses on the attitude estimation of indoor UAVs and explores a solution based on binocular vision for indoor UAV attitude estimation. This solution utilizes the feature points obtained by the camera to calculate the change in the null space vector of the feature points during the attitude change process. Meanwhile, the binocular camera can obtain depth information to achieve accurate indoor UAV attitude estimation. The experimental results demonstrate that the proposed solution can achieve more accurate attitude estimation and provide effective attitude correction information for subsequent indoor drone navigation and positioning methods.