Model for Identification and Prediction of Leaf Patterns: Preliminary Study for Improvement

Abstract

Purpose: Many studies have conducted studies related to automation for image-based plant species identification recently. Types of plants, in general, can be identified by looking at the shape of the leaves, colors, stems, flowers, and others. Not everyone can immediately recognize the types of plants scattered around the environment. In Indonesia, herbal plants thrive and are abundantly found and used as a concoction of traditional medicine known for its medicinal properties from generation to generation. In the current Z-generation era, children lack an understanding of the types of plants that benefit life. This study identifies and predicts the pattern of the leaf shape of herbal plants. Methods: The dataset used in this study used 15 types of herbal plants with 30 leaf data for each plant to obtain 450 data used. The extraction process uses the GLCM algorithm, and classification uses the K-NN algorithm. Result: The results carried out through the testing process in this study showed that the accuracy rate of the leaf pattern prediction process was 74% of the total 15 types of plants used. Value: Process of identifying and predicting leaf patterns of herbal plants can be applied using the K-NN classification algorithm combined with GLCM with the level of accuracy obtained.