Recent Advances In Intelligent Inspection Of WoodBoards

Abstract

Automated Visual Inspection (AVI) is being proposed for the task of inspecting wood sheets for quality control. The motivation for this is that currently, the human graders employed for the task can only perform inspection with an accuracy of around 60%. For wood inspection, AVI operates by employing a camera to acquire an image of the sheet and then utilising appropriate image processing hardware and software routines to find and classify surface defects. A typical AVI framework includes the stages of image acquisition, image enhancement, segmentation, feature extraction, classification and grading. The image acquisition stage obtains an image of the sheet. Image enhancement improves the quality of the acquired image to facilitate segmentation. Frequently, this stage is not used because it is considered more important to obtain high quality images. Image segmentation divides the image into clear wood and defect regions. The result of this stage is called the segmented image which contains objects that represent the defects. Then, feature extraction is employed to calculate numerical values to represent each object. The classification stage determines the type of each object based on its features. Finally, the board is given a grade based on the number of defects found and the size and severity of each one. Grading is a task which can be implemented with the aid of a grading table and a simple expert system. Therefore, it offers few opportunities for research. This paper gives details of advances in the areas of image acquisition, image enhancement, segmentation, feature extraction and classification for Automated Visual Inspection of wood boards. Artificial intelligence techniques, such as neural networks, genetic algorithms and fuzzy logic, are concentrated upon. Transactions on Information and Communications Technologies vol 20, © 1998 WIT Press, www.witpress.com, ISSN 1743-3517