A detection of straight and slant fiber woods is needed to sort the wood quality. This detection has been discussed by using a fiber slop angle method. In this paper, we use the same sample and increase the accuracy of this detection. We use a number of the small distances between the fibers. This method is applied to generate the feature. Then the feature is inputted on a thresholding. The method increases 5.2% accuracy over the previous ones. A new accuracy is reached on 84.4%.

detection, number of the small distances, wood, feature extraction.