Pengembangan Sistem Monitoring Berbasis Ground untuk Tanaman Kubis (Brassica Oleracea L) Menggunakan Basis Data dalam Optimalisasi Penyemprotan Pestisida Otomatis

Abstract

Cabbage (Brassica oleracea L) is a vegetable commodity in Argosari, Lumajang, which is frequently threatened by the cabbage head caterpillar (Crocidolomia pavonana). This infestation often leads to crop failure and excessive pesticide waste due to inefficient conventional spraying methods. This research aims to design and evaluate an intelligent monitoring and automated spraying system for early detection and precision treatment of cabbage plants. The methodology involves field image acquisition using a portable webcam, utilizing the standard planting distance (50 cm × 60 cm × 60 cm × 60 cm) as a mobilization path. Captured images are processed using a Convolutional Neural Network (CNN) architecture trained over 100 epochs to extract leaf damage features. The CNN classification output is converted into binary data (Healthy = 1, Diseased = 0) and synchronized in real-time with a local database to trigger field actuators. Experimental results demonstrate that the CNN model achieves optimal performance (Good Fit) with a training accuracy of 87.94%, validation accuracy of 78.90%, and a validation loss of 0.3895. The system effectively automates spraying only on detected diseased plants. This implementation proves to increase chemical usage efficiency and enables precision automated spraying on cabbage plants identified with infestation