A vision-based terrain morphology estimation model inspired by the avian hippocampus

Abstract

Homing pigeons are known for their ability to return home after being released from a location spanning up to hundreds of miles. They make use of detected visual features in the environment, the earth׳s magnetic field as well as using the hippocampus region of the brain to construct spatial maps of the environment. This is unlike present day UAVs that rely on GPS and radio/satellite communications with a ground station, both of which might not be available during a major disaster scenario such as a solar flare. In this paper, we take inspiration from the avian hippocampus and develop a preliminary model for estimating a terrain׳s morphology using visually detected features on the terrain. This could then be used to localise a portable micro-UAV during a demining task for humanitarian purposes in third world countries affected by buried land mines from previous wars. Our goal is that in future, the presented model and algorithm in this work would enable effective coverage of an affected area using the visual information obtained from the environment.

Keywords

• Micro-UAV
• Hippocampus
• Avian
• Neuroscience
• Robotics
• Terrain morphology