Wood Texture Features Extraction by Using GLCM Combined With Various Edge Detection Methods

2016 International Congress on Theoretical and Applied Mathematics, Physics and Chemistry

数学;物理学;化学

Fahrurozi, A.^1 ; Madenda, S.^1 ; Ernastuti^1 ; Kerami, D.^2

Computer Science Department, Gunadarma University, Indonesia^1

Mathematics Department, University of Indonesia, Indonesia^2

关键词: Edge detection methods;    Edge operator;    Glcm parameters;    Gray-level co-occurrence matrix;    Image forming;    Laplacian of Gaussian;    Natural materials;    Texture features;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/725/1/012005/pdf
DOI  :  10.1088/1742-6596/725/1/012005

来源: IOP

【 摘 要 】

An image forming specific texture can be distinguished manually through the eye. However, sometimes it is difficult to do if the texture owned quite similar. Wood is a natural material that forms a unique texture. Experts can distinguish the quality of wood based texture observed in certain parts of the wood. In this study, it has been extracted texture features of the wood image that can be used to identify the characteristics of wood digitally by computer. Feature extraction carried out using Gray Level Co-occurrence Matrices (GLCM) built on an image from several edge detection methods applied to wood image. Edge detection methods used include Roberts, Sobel, Prewitt, Canny and Laplacian of Gaussian. The image of wood taken in LE2i laboratory, Universite de Bourgogne from the wood sample in France that grouped by their quality by experts and divided into four types of quality. Obtained a statistic that illustrates the distribution of texture features values of each wood type which compared according to the edge operator that is used and selection of specified GLCM parameters.