IDENTIFIKASI SERAT BAMBU MENGGUNAKAN EKSTRAKSI CIRI STATISTIK ORDE 2 (GLCM) DAN PENGUKURAN JARAK K-NN

Abstract

Indonesia is a large bamboo producer. Many benefits can be taken from bamboo trees, among others, as an alternative material for environmentally friendly construction, handicraft, and even become a safe material for use. Based on the property of its mechanical strength, bamboo has high tensile strength and fiber content, including fiber length, inter-fiber adhesive, namely lignin and the higher diameter of bamboo fiber, causing bamboo stems to become stronger and stiffer so that bamboo quality is getting better. One objective is to use a texture analysis of statistical features extraction of digital image processing. Feature extraction is a process to get the characteristics of visual perception. Texture information can be used to distinguish the surface properties of objects in images that are related to coarse and fine. This research uses a second-order statistical calculation of Gray Level Co-occurrence Matrices (GLCM) by measuring contrast, energy, homogeneity, and correlation levels to determine roughness from bamboo image textures that have irregular patterns. The second method is to use similarity measurements with the K-NN method in which in this study K = 3 with testing images of 28 images obtained an accuracy of 0.8, precission of 0, 8 and f-measurement of 0.9.