<![CDATA[Green mustard is a leading agricultural commodity in Indonesia but often faces pest attacks such as Spodoptera litura, which can reduce productivity by up to 85%. This study aims to develop an automated pesticide sprayer prototype using Convolutional Neural Network (CNN) technology with the VGG19 architecture. The system utilizes Raspberry Pi, Arduino, ESP8266, and a camera to detect pests in real-time and accurately spray pesticides. The dataset used consists of 1,380 images, divided into 10% for testing, 25% for validation, and 75% for training. The model evaluation values for the ‘mustard with pests’ class achieved precision, recall, and F1-Score of 96% each, while for the ‘mustard without pests’ class they were 95% each. In addition, the MAPE (Mean Absolute Percentage Error) value of 4.61% shows that the percentage error of the model prediction is very small. The developed VGG19 model achieved an accuracy of 95% and high efficiency after conversion to the TFLite format, reducing model size by 75.57%. This tool is highly recommended to enhance farmers' work efficiency, reduce excessive pesticide use, and support sustainable agriculture. Its ability to operate autonomously and precisely makes it an ideal solution to assist farmers regarding pest problems.]]>
Copyrights © 2025
