Research on indoor multipoint data acquisition with a micro UAV

Abstract

Unmanned Aerial Vehicles (UAVs) have been widely used in many fields. But inside the room, it is difficult for the drone to perform the appointed task without the signal of satellite navigation system. Besides, wireless sensor networks (WSNs) have been widely used in various fields. But it cannot be centralized to process and maintain. Keeping in view the above shortcomings, a new method based on target recognition and locating technology is proposed in this paper, which acquires measured data of different known points by a micro UAV equipped with a set of sensors in a closed space. The key issue for this topic is path planning in a closed space, which will be executed with target recognition and visual localization. The principles of target recognition and monocular visual localization are explained. The experimental setup is designed for monocular visual localization and the multi-data acquisition with a drone in a room. In the experiment, the drone can recognize the appointed oriented roundel and estimate the distance between the camera and the object. In the test of distance estimation, the accuracy and the recognition rate decreases as the distance increases. And for robustness at the distance of 50.00 cm, the mean is 50.54 cm and the standard deviation is 0.23 cm. The result shows that the new method proposed in this paper is feasible.