Study of Neural Network Training Algorithms in Detection of Wood Surface Defects

Accurate detection of defects through machine vision improves the economical growth of wood industry. In this paper six common defects on wood surface are considered for study. Quality of wood image is enhanced by Histogram Equalization method. Contrast enhanced images are subject to Thresholding se...

Full description

Saved in:
Bibliographic Details
Published inInternational journal of automation and smart technology Vol. 9; no. 3
Main Author Chandirasekaran, M.Thilagavathi
Format Journal Article
LanguageEnglish
Published 26.06.2025
Online AccessGet full text
ISSN2223-9766
2223-9766
DOI10.5875/j6m6ve94

Cover

More Information
Summary:Accurate detection of defects through machine vision improves the economical growth of wood industry. In this paper six common defects on wood surface are considered for study. Quality of wood image is enhanced by Histogram Equalization method. Contrast enhanced images are subject to Thresholding segmentation which examines the objects in the image and identifies the defect. The segmented images are cropped in to small blocks. Segmentationbased Fractal Texture Analysis (SFTA) feature extraction method is accomplished to extract 21 texture features from the wood images. The extracted features are fed in to the training algorithms such as Levenberg-Marquardt, Scaled Conjugate Gradient, Gradient Descent with Adaptive Learning Rate, Bayesian Regularization and Resilent Backpropagation. The Performance of training algorithms are analyzed with several performance metrics. The result obtained shows a considerable improvement in accuracy of 98.2 % by Bayesian Regularization tool. 
ISSN:2223-9766
2223-9766
DOI:10.5875/j6m6ve94