The Open Automation and Control Systems Journal

2015, 7 : 290-295
Published online 2015 April 17. DOI: 10.2174/1874444301507010290
Publisher ID: TOAUTOCJ-7-290

Study on Wood Board Defect Detection Based on Artificial Neural Network

Lin Wenshu , Shao Lijun and Wu Jinzhuo
College of Engineering and Technology, Northeast Forestry University, Harbin, 150040, P.R. China.

ABSTRACT

Due to the increasing contradiction between supply and demand of timber resources, it is necessary to realize the reasonable utilization of the resources. Fast and accurate identification of the location and size of the defects on a piece of wood board is a premise of efficient utilization. In this paper, the identification and positioning of board defects by computer vision and artificial neural network technology was discussed. Through the acquisition of the board image by CCD camera and image processing by using MATLAB, the location and size of the defects on the board was obtained. Finally, the artificial neural network was constructed to identify the defects, and the results showed that board defect identification rate can reach 86.67%. The study has provided a new idea and method to improve lumber recovery, which also provided a theoretical and technical foundation for advanced automation and intelligent wood processing industry.

Keywords:

Artificial neural network, Computer vision technology, Defect testing, Image processing.