DETECTION OF ANIMALS IN AGRICULTURAL LAND USING CNN

Abstract

The agricultural sector faces significant losses due to animal assaults, which damage crops and property, particularly in regions where farmed land extends into former animal habitats. Traditional methods to prevent such damage are often ineffective and labor-intensive, making it difficult for farmers to monitor vast areas continuously. This problem exacerbates human-animal conflict, affecting both human and animal safety. To address this, we propose an automated system that uses deep learning, specifically convolutional neural networks (CNNs), to detect animals entering agricultural lands. The system employs real-time surveillance through cameras that monitor the farm throughout the day. When an animal is detected, the system categorizes it using the YOLO algorithm and triggers sound-based scare tactics to deter the animal without causing harm. Additionally, geo-location data and images are sent to farmers and forest officials to help them take further action if needed. This approach not only reduces crop damage but also minimizes human-animal conflicts, providing a humane and efficient solution. By automating the monitoring and response process, the system reduces the need for human intervention, ensuring the safety of both crops and animals while enhancing agricultural productivity. Key Words : Convolutional neural network, real-time surveillance, YOLO