An Indoor Vision-Based Markov Localization Technique of a Quadrotor

This work develops a vision-based quadrotor localization and orientation determination approach for indoor environment. The quadrotor is equipped with a downward looking camera. Two algorithms are implemented to acquire and analyze the images received from the camera. Firstly, an image processing technique is introduced for detecting visual targets and features to calculate the relative position and orientation of the quadrotor in addition to it's altitude. Then Markov localization technique is used to increase the robustness of location recognition. The proposed algorithm provided absolute position, orientation and altitude at various altitudes and quadrotor's inclinations by fusing the two algorithms. Simulation results demonstrate that the proposed algorithm provided accurate results and is viable in an indoor environment and GPS denied areas.