Landscape mapping MAV using single image perspective cues

Abstract

We consider the problem of mapping with Miniature Aerial Vehicles (MAVs) in outdoor sites with distinguishable landscapes. The primary long range sensor in these MAVs is a miniature camera. While previous approaches first try to build a 3D model in order to do mapping, our method does require a 3D model. Instead, our method first classifies the type of site the MAV is in, and the uses vision algorithms based on perspective cues to estimate the landscape location and the do mapping. We tested our method on a number of sites with different landscapes. Our experiments show that our vision algorithms are reliable, and the enable the MAV to identify landscapes and map them.