Klasifikasi Tanaman Beringin (Ficus Bernjamina) berdasarkan Citra Daun Menggunakan Algoritma K-Nearest Neighbors

Abstract

One of the problems faced when choosing a banyan, whether to be used as a shade plant, bonsai, or medicinal plant, is to identify the appropriate type of banyan. So research must be done to find out the desired type of banyan. One way that can be used to classify is with digital image processing technology, namely by extracting features or characteristics from digital images or images. The challenge is how to classify banyan plants based on leaf images using digital image processing. This study aims to design or design and compile a digital image processing program and the K-Nearest Neighbors (KNN) algorithm for the classification of the banyan species which can be used as a model for an automatic classification system using computer equipment. The results of the research on the process of testing the classification of ficus plants based on texture and shape characteristics on leaf images using the K-Nearest Neighbors algorithm can be concluded that the application has been successfully designed and built and can be used for the texture and shape feature extraction process and can be used for the classification process. From feature extraction, seven GLCM texture features are obtained, namely energy, entropy, contrast, homogeneity, IDM, variance, and dissimilarity, and 2 shape features, namely roundness, and compactness. The test results show a relatively low accuracy value of 56.25% with data on the number of images recognized according to the type of ficus as many as 18 and not recognized as many as 5 images