Automatic landing site detection for UAV using supervised classification

Abstract

UAVs plays an important role in day today life. They are used in a variety of applications like search and rescue, power line management etc. During operation of the UAV, sometimes it gets in operated and it has to be landed in the immediate neighborhood. The landing should occur in a proper location because improper landing or landing on the undesirable platform will cause UAV loss. Here a supervised classification is implemented for finding the landing site. In a populated area to land on rooftops a rule set is formed, such as length width, area etc. to locate the rooftop. Similarly train set of buildings, roads and lakes are composed to segregate the area, and an appropriate landing site for a UAV is determined. The results are compared with K nearest neighbor and Support Vector Machine algorithm.