Real-Time Agricultural Monitoring with Agrobot: A Raspberry Pi and YOLO Based Solution

Abstract

Plant diseases and pests pose a serious threat to agriculture, causing substantial damage to crops and reducing productivity. Early detection and treatment of these issues are critical to improve crop quality and yield. This paper presents a robotic framework based on artificial intelligence for real-time detection of plant diseases and pests, with a specific focus on tomatoes, a major crop in Sudan. The system “Agrobot” is equipped with a Raspberry Pi controller and object detection model YOLO for fast and accurate detection of the most severe diseases and common pests that damage tomato crops. The robot navigates the field while a Pi camera mounted on it scans the crops for diseases and pests using a live video feed. Servo motors move the camera in all directions for comprehensive scanning. In addition, this paper provides a comparison between two YOLO models, YOLOv4 and Tiny YOLOv4, to determine their accuracy and efficiency in detecting plant diseases and pests. Our results show that the proposed system can effectively detect tomato diseases and pests, paving the way for more efficient and precise crop management.