Abstract:

With the textural features of wood images as key bases, it aimed to realize the retrieval and identification of wood species. It adopted such contents as color, grey and texture of wood images, and measured with nine parameters: hue, saturation, illuminance, contrast, angular second moment, sum of variances, long run emphasis, fractal dimension, and wavelet horizontal energy proportion, then performed retrieval of wood species according to maximal similarity theory. The results showed that the retrieval and identification of wood species could be realized on bases of image textural features and relevant algorithms. The retrieval accuracy and uniqueness by limen discriminance of synthetical features were better than those by minimal differences method. But when the texture of samples was weak or textureless, the uniqueness of retrieval result would not be very satisfying. Even then, it is a worthwhile developing retrieval and identification method of wood species.

Key words: wood species, retrieval, image, textural features