PERANCANGAN SISTEM KLASIFIKASI DAUN SAWI HIJAU BERDASARKAN SERANGAN HAMA MENGGUNAKAN ALGORITMA CONVOLUTIONAL NEURAL NETWORK (CNN)

Green mustard (Brassica juncea) is a horticultural crop that is susceptible to pest attacks, making rapid and accurate identification necessary. This study aims to design a green mustard leaf image classification system based on pest attacks using a Convolutional Neural Network (CNN) with the EfficientNetB3 architecture. The dataset consists of 1,000 images divided into five classes: healthy, grasshopper, leaf-damaging caterpillar, mealybug, and mole cricket. The research process includes preprocessing, dataset splitting, transfer learning, fine-tuning, and evaluation using a confusion matrix. The dataset was divided into 80% training data and 20% validation data. The results showed a validation accuracy of 87.50%. The model was then implemented in a Flask-based system to classify images and display confidence scores. The developed system can assist in the early identification of pest attacks on green mustard leaves.

Keywords: Convolutional Neural Network (CNN), Green Mustard, Image Classification, EfficientNetB3.