Development of a Robotic System for Agricultural Pest Detection: A Case Study on Chili Plants

Abstract

Chili peppers, a key agricultural commodity in Indonesia, are highly susceptible to pest infestations and diseases, leading to significant economic losses and challenges in sustainable farming. This study presents the design and implementation of a real-time pest detection system that integrates robotics, computer vision, and deep learning to enhance agricultural productivity. The system is built on a Raspberry Pi 5 and Arduino Mega Pro Mini, utilizing a camera for image capture and ultrasonic sensors for navigation. A ResNet-based model was trained on a dataset of 2,703 chili leaf images, categorized into healthy and diseased classes, achieving a detection accuracy of 91%. The system provides early warnings to farmers through a web-based interface, allowing timely intervention and reducing reliance on chemical pesticides. While promising, the system faced challenges such as environmental variability, which influenced image recognition accuracy. By automating pest detection and promoting precision farming, this innovation addresses the need for sustainable agricultural practices, contributing to global food security and reducing environmental impact.