Implementation of Autonomous Unmanned Aerial Vehicle with Moving-Object Detection and Face Recognition

Various personal robots have developed for performing useful tasks autonomously without the human interaction. The development of human navigation and tracking in the real time environment will lead to the implementation of more advanced tasks that can performed by the autonomous robots. This project proposes a system that can track objects using Speeded Up Robust Feature (SURF) detection and human face recognition using Eigenface recognizer. In this paper, combinations of algorithms such as modified Eigenface, Haar-Cascade classifier and SURF resulted in a more robust system for object detection and face recognition. The proposed system was implemented on AR drone 2.0 using the Microsoft Visual Studio 2010 platform together with OpenCV. The testing of the proposed system carried out in an indoor environment in order to evaluate its performance in terms of detection distance, angle of detection, and accuracy of detection. 132 images of different people were used for face recognition at four different detection distances. The best average result of 92.22% was obtained at a detection distance of 50cm to 75cm. It noted that the detection accuracy decreases as the detection distance is increased. Furthermore, improve various limitations, strengths of the AR drone and the proposed algorithm.