An Intelligent Animal Repellent System for Crop Protection: A Deep Learning Approach

Abstract

Human Wildlife Conflicts (HWC) refers to the negative interactions between human and wild animals, with undesirable consequences both for people and their resources, on the one hand, and wildlife and their habitats on the other (IUCN 2020).HWC, caused by competition for natural resources between humans and wildlife, influences human food security and the well-being of both humans and animals. As a result of human population growth and the transformation of land use in many regions, the number of these conflicts has increased in recent decades. HWC is a significant global threat to reliable development, food security, and conservation in urban and rural landscapes alike. In general, the consequences of HWC include crop destruction, reduced agricultural productivity, competition for grazing lands and water supply, livestock predation, injury and death to humans, damage to infrastructure, etc. From farmers’ perspectives, their main concern is to protect their crops from animal intrusion and crop destruction. It is necessary to have an affordable better technique for crop protection from animal attacks. Human-wildlife conflict is on the verge of its extremity. Because of the increased deforestation and human penetration into their area, the wild animals have to roam around aimlessly and intrude into the human habitats. While considering the safety of both humans and animals, it is necessary to think and act differently than the traditional solutions (like shooting, trapping, electric fences, etc). This work presents an intelligent animal repellent system, particularly for crop protection, with the advent of some deep learning techniques. This automatic intrusion and repellent system uses the MobileNet SSD model for better performance. On detection of the hazardous animal, the system produces an alarm sound and makes an alert notification to the responsible authorities to make an awareness of the detection. This rapid way of detection makes it more human-friendly and the harmless way of repulsion makes it more animal friendly at the same time.