Intelligent IoT-Enabled Crop Defense System for Preventing Animal and Bird Intrusion

Abstract

Animal and bird intrusions in modern agricultural fields are persistent problems that may cause serious damage to crops and financial loss. AI-powered intelligent surveillance systems are imbued with the power of machine learning for effective and reliable solutions. Based on this, this paper proposes a real-time wildlife detection and monitoring system that will help farmers in effectively and efficiently detecting and handling intruding animals and birds. The YOLOv8 algorithm, which is a high-end deep learning framework for fast and accurate object detection, is used to implement the proposed system. A camera captures continuous images of the farm environment, then pre-processing of the images using OpenCV could be done, including noise reduction, resizing, and normalization, for increased accuracy in object detection. After detection, the images are sent to the remote server and deleted automatically after processing to save storage. Other steps necessary to provide real-time efficient performances are dimensionality reduction, feature extraction, and image compression. After detecting the intrusion, multiple automated responses from the system include sending an email to the farmer with a detected species and timestamp, switching the buzzer on for immediate notification, and showing the detection details on the LCD display. When nighttime falls, LED floodlights automatically turn on to improve visibility and keep nighttime wildlife away. Continuous improvement of the YOLOv8 model will enable it to recognize a wide range of species, and with changing environmental conditions, update its model accordingly.