<![CDATA[Technological advancements play a crucial role in enhancing the efficiency of modern agriculture, particularly in addressing pest management challenges. This study focuses on the development of an automatic pest detection system for shallot crops using a combination of Arduino Uno microcontroller, ESP32-CAM camera module, and YOLOv3 object detection model. The system is designed to detect pests in real-time through images captured by ESP32-CAM and analyzed using YOLOv3, then provide an automatic response by spraying pesticides only in areas where pests are detected. The study began with the development of hardware and software for the automatic pest detection system. Arduino Uno is used as the main microcontroller to control the entire system, while ESP32-CAM is responsible for capturing images and detecting pests. The YOLOv3 model is trained using the COCO dataset, supplemented with sample images of pests on shallot crops to improve detection accuracy. The training process is conducted using a GPU to speed up model learning. Field tests on shallot crops infested with various types of pests show that this system has a high accuracy rate in detecting pests and effectively provides automatic pesticide spraying responses. The spraying system's effectiveness reaches 93%, ensuring pesticides are sprayed only in areas where pests are detected, thus optimizing pesticide use and reducing negative environmental impacts. This system offers an efficient and environmentally friendly solution for pest control and has significant potential for application in various agricultural scenarios. This research contributes to the improvement of agricultural productivity and the welfare of farmers in Indonesia.]]>
