Vision Based Algorithms for MAV Navigation

This paper focus on the development of vision based navigation algorithms for Micro Aerial Vehicles (MAV's). The proposed algorithms provides a framework for two potential navigation applications such as terrain detection and real time target tracking. Image processing algorithms are formulated to detect or track the target based on its color feature. The color based target extraction is performed considering several color models such as RGB, Normalized RGB, HSI, YCbCr, YUV, YIQ, CIELAB and CIELUV. It is observed that Y based color models are having superior performance in terms of thresholding time and accuracy. Utilizing optimal color model, navigation algorithms are developed. Simulation and experimentation results confirm that, the proposed algorithms can be an effective choice for MAV applications.