PROTECTING CROPS FROM WILDLIFE ANIMALS IN SMART AGRICULTURE WITH REAL-TIME OBJECT DETECTION USING YOLOV5 ALGORITHM

Abstract

Agriculture is essential for human survival, as it provides food, employment, economic growth, livelihood, and rural development, while also maintaining environmental balance and food security. However, due to the damage caused by wild animals, many farmers are abandoning cultivation. Existing techniques for deterring animals from agricultural fields are limited to their detection and the use of ultrasonic sounds. The proposed approach utilizes the YOLO machine learning algorithm to identify animals in the fields, generate ultrasonic sounds based on the detected species, activate Light-Emitting Diodes (LEDs) to simulate fire, and send an alert message to an authorized individual upon detection. The results obtained from this method surpass current approaches in reliability, precision, recall, and F1-score, achieving values ranging from 94 to 96 %.