Machine learning-based Acoustic Repellent System for Protecting Crops against Wild Animal Attacks
Devsmit Ranparia, Gunjeet Singh, Anmol Rattan, Harpreet Singh, Nitin Auluck
Abstract
In this paper, we present some insights on the issue of crop destruction by wild animals. This is a serious concern for the affected farmers throughout the world and leads to significant social and financial distress among them. In order to understand the background of this problem, a survey of Katli village, Rupnagar, (India) was conducted. The main aim of the current work is to develop a device to protect crops from damage by wild animals by diverting them from the farms, without harming them physically. In this context, an Acoustic Repellent System has been designed which uses a convolutional neural network (CNN) based machine learning model and an IR camera to identify target animals, such as wild boar, nilgai, and deer. A Raspberry Pi (Rpi) module has been integrated with a camera and a frequency generator to recognise different animals and produce corresponding frequencies that keep them away from the farms of interest. Moreover, the architectural aspects of the proposed solution have also been detailed. Lastly, the potential impact of the proposed solution has been discussed.