Development of a Low Cost Crow Deterrent AI System

Abstract

Damage caused by crows such as damage to crops, scattering garbage on the streets, assaulting people walking beneath their nests, etc. has become a problem in various parts of the country. There are several damage abatement techniques such as protective nets, scarecrows, flashing lights, etc. However, existing methods have limited effects and are not always practical. Research using promising AI systems use expensive equipment. In this research, we propose an end-to-end real-time crow deterrent system based on the latest deep learning technology, using low-cost equipment. The proposed system detects the crow arrival and generates randomly frightening noises, including predator or bird distress calls. We conducted various experiments and confirmed the effectiveness of the proposed system.