Comparison of UAV based person detection and tracking techniques

Motion detection, based on UAV surveillance, is one of the challenging task due to the continuous motion of the camera. In this paper, image alignment based on feature detection and feature matching has been presented using FAST algorithm which provides the important features whereas both RANSAC and MSAC algorithm is used to estimate the affine transformation utilizing a filter algorithm to remove the particular noise and make system easier. To detect the motion of a person, optical flow and background subtraction techniques are applied. Moreover, both techniques are compared for each ego-motion compensation algorithm used and subsequent results have been shown.