Classification of compression wood using digital image analysis
1995
Andersson, C. | Walter, F.
Compression wood causes warping in sawn wood products. In order to compare the deformations in boards sawn from logs that commonly have compression wood, a distinct classification method for compression wood must be used. Visual human detection in combination with a planimeter might be used to determine the amount of compression wood in a log cross section, but this is a subjective and time-consuming method. This paper describes a new method that uses digital image analysis techniques. Color images of wood disks that were 2 to 3 mm thick were digitized. The computer program developed uses a supervised maximum-likelihood classification to discriminate between three kinds of wood-types in the images: normal wood, mild, and severe compression wood. The most important descriptors computed by the program are the proportions of the different wood-types and the angle from the pith to the center of gravity of the two kinds of compression wood. The advantages of this method are that it is more objective, faster, and easier to implement than other methods. The estimations of the amount of total and mild compression wood are quite good, while those of severe compression wood are not so good. The estimations of the angle are satisfactory. The use of computer tomography and ultraviolet light might be other approaches to classify compression wood.
显示更多 [+] 显示较少 [-]