Identifikasi Penyakit Daun Cendana Berbasis Deep Learning Convolutional Neural Network

Abstract

Leaf diseases are the primary pathological constraint reducing the productivity and economic value of sandalwood trees (Santalum album). Conventional symptom identification is often hampered by observer subjectivity and low time efficiency. This study implements computer vision technology using a Deep Learning algorithm based on a Convolutional Neural Network (CNN) to automate the identification of sandalwood leaf diseases. By utilising convolutional layers, the model is able to autonomously extract morphological and chromatic features to recognise complex leaf damage patterns without the need for manual feature extraction. The results of the study show that the CNN model delivers superior performance, with accuracy, precision, recall, and F1-score reaching 100%, as well as a computational time efficiency of 1 minute and 18 seconds. It is hoped that the implementation of this technology will serve as a precise early-diagnosis tool for forestry practitioners in supporting conservation efforts and safeguarding the sustainability of the sandalwood population, particularly in the Timor region.