The fractal estimation of wood color variation by the triangular prism surface area method

Color variations of the surfaces of fifteen wood species were characterized by fractal dimension of the triangular prism surface area method. Softwood and hardwood indicated apparently different in the mean fractal dimensions of red and green colors. Red color behaved steadier in softwood than in hardwood and green color varied comparatively stronger in hardwood than in softwood. No evident differences between softwood and hardwood were found in the variation of blue color of all the specimens. Following the low-to-high value order of the mean fractal dimensions, three types of combination of red (R), green (G) and blue (B) colors were found: RGB, RBG and BGR. There also existed six types of fractal dimension distribution; namely, plane, included plane, concave, convex, zigzag, and hilly distributions. Fractal dimensions across the grain changed greatly whereas those along the grain became relatively steady. The characteristic difference of color variation was defined for each species, which was inferred to characterize its own unique appearance of surface color. For color matching of wood parts, fractal dimension quantitatively furnishes essential information of color variation in local and overall features. Such evaluation can be efficiently carried out with few measurements along the grain and by detecting a single color (red, green or blue) only. Received 5 May 2000