Classification of Lanzones Tree Leaf Diseases Using Image Processing Technology and a Convolutional Neural Network (CNN)
Adzmer S. Muhali, Noel B. Linsangan
Abstract
Nowadays, many artificial intelligence systems are developed to detect and classify plant leaf diseases via image classification, and those researches achieved high-performance results. However, those Artificial Intelligence Systems categorize various leaf diseases, but none precisely for Lanzones. Briefly, no existing or current research has been studied to detect and classify leaf diseases of Lanzones tree in the early stages using an Artificial Intelligence System. This paper presents a device for categorizing the healthy and with diseases (Algal Leaf Spot and Leaf Blight Spot) on the leaves of the Lanzones tree thru image classification using deep learning. It is limited to the device's development and implementation in the target locale. The methods applied include data pre-processing and augmentation techniques to the created CNN model. This will be integrated into a system using the Raspberry Pi 4 Model B and the Tensorflow Lite Interpreter. The performance analysis and evaluation of the system were based on the confusion matrix, which shows how the system successfully classified the healthy and disease of Lanzones leaves with an accuracy rate of 80%.