APPLYING AI TECHONLOGY IN PROTECTION OF SWIFTLETS FROM SOME PREDATORS

Abstract

In recent years, swiftlet farming is developing rapidly with the ability to get high profits and provide many jobs for farmers. However, some popular predators such as snakes, hawks and owls, are a big problem that cannot be ignored. This paper proposes a model of repelling and alarming when the predatory animals enter the swiftlets’ house. A deep learning model-based software is developed to identify some popular predators from images captured by camera. An IoT-based hardware is applied to send alarm messages to mobile phone when the software detects snake attacked or turn on a light to repel hawks and owls. First stage testing results show that the deep learning model can classify some popular predators with over 90% accuracy while the hardware system can text to right phone numbers when detecting snakes, and turn on the light when recognizing hawks or owls.