Image-Based Analysis for Pests and Diseases Classification of Katokkon Chili Pepper Using Deep Learning

• Published in: International Review on Modeling and Simulations Vol. 18; no. 2; p. 136
• Main Authors: Pineng, Martina, Palantei, Elyas, Areni, Intan Sari, Wardi, Wardi
• Format: Journal Article
• Language: English
• Published: Naples Praise Worthy Prize 30.04.2025
• Subjects: Artificial neural networks; Classification; Deep learning; Machine learning; Peppers; Pests; Taste
• ISSN: 1974-9821; 2533-1701; 1974-9821
• DOI: 10.15866/iremos.v18i2.26409

This Katokkon chili pepper comes from the region of Tana Toraja, South Sulawesi, Indonesia. Its long, slender, cone-like shape and bright red color when ripe are the characteristics of this variety. Katokkon chili has a spicy taste and a bit of sweetness. The level of spiciness is moderate. Its unique shape and taste distinguish it from other types of chilies. This study presents the development of a deep learning-based classification model that classifies pests and diseases in Katokkon chili using images. The research method involves collecting image data from Katokkon chili infected with fruit fly pests, Anthracnose diseases, and pictures of healthy chili. The datasets used are 568 images, 454 of which have been training and 114 have been testing. The deep learning model used is the Convolutional Neural Network (CNN) and Adam has been used as the optimizer, with batch size 32, and 50. According to the data, the CNN model has obtained a level of accuracy rate of 0.95, demonstrating superiority over conventional methods. This study has obtained high-performance results by using the CNN model to classify pests and diseases of the Katokkon chili pepper.