Publications
0000
1.
Morala-Argüello, Patricia; Fernández-Robles, Laura
SURFACE ROUGHNESS CLASSIFICATION IN METALLIC PARTS USING HARALICK DESCRIPTORS AND QUADRATIC DISCRIMINANT ANALYSIS Artículo de revista
En: 0000.
Resumen | Enlaces | BibTeX | Etiquetas: Co-ocurrence Matrix, discriminantAnalysis, Haralick, roughness, surface texture
@article{morala-arguello_surface_nodate,
title = {SURFACE ROUGHNESS CLASSIFICATION IN METALLIC PARTS USING HARALICK DESCRIPTORS AND QUADRATIC DISCRIMINANT ANALYSIS},
author = {Patricia Morala-Argüello and Laura Fernández-Robles},
url = {https://www.daaam.info/Downloads/Pdfs/proceedings/proceedings_2010/23374_Symp_1_head.pdf},
abstract = {An artificial vision system has been used to classify
metallic work-parts in base of their surface roughness.
Haralick features have been computed through the gray-level
co-occurrence matrix (GLCM) to analyze the texture of the
parts. Quadratic and Linear Discriminant Analysis (QDA and
LDA) algorithms have been worked out to classify the
descriptors. Results have proved the validity of this method to
classify metallic parts in two classes achieving hit rates of
97,4% using QDA.},
keywords = {Co-ocurrence Matrix, discriminantAnalysis, Haralick, roughness, surface texture},
pubstate = {published},
tppubtype = {article}
}
An artificial vision system has been used to classify
metallic work-parts in base of their surface roughness.
Haralick features have been computed through the gray-level
co-occurrence matrix (GLCM) to analyze the texture of the
parts. Quadratic and Linear Discriminant Analysis (QDA and
LDA) algorithms have been worked out to classify the
descriptors. Results have proved the validity of this method to
classify metallic parts in two classes achieving hit rates of
97,4% using QDA.
metallic work-parts in base of their surface roughness.
Haralick features have been computed through the gray-level
co-occurrence matrix (GLCM) to analyze the texture of the
parts. Quadratic and Linear Discriminant Analysis (QDA and
LDA) algorithms have been worked out to classify the
descriptors. Results have proved the validity of this method to
classify metallic parts in two classes achieving hit rates of
97,4% using QDA.