A Digital Image Processing System Based On Wood Tissue

The paper applied digital image processing in wood tissue and structure so as to promote automation and intelligence in measuring technique and wood recognition based on analyzing developing situation of the application of digital image processing in wood structure, the development of wood recognition, in China and other countries.It mainly analyzed the morphological characteristics of pores and resin canal on transverse section and wood ray on tangential section on the basis of wood structure. The variation of pores in ring porous wood, semi-ring porous wood(semi-diffuse porous wood) and diffuse porous was discussed, the distinction between pore, wood ray with their likes were made. The shape and size of wood ray is different form that of longitudinal parenchyma; Compared with pores, resin canal doesn't have smooth edge and it's density is far lower than that of pores.For the first time, the tissue image segmentation was put forward based on mathematic morphology so as to classify the wood type. And it's also the first time for wood ray on tangential section be cut apart by combining mathematic morphology and algebraic operation. With a structural element of "disk" and open operation, pores can be well segmented through open operation. It's more complicate to segment wood ray for you need to adopt different parameters and methods in different situation.Some factors of pores such as the morphological factor, tissue proportion factor, dimension factor, and density of distribution were extracted based on segmented image. it digitized the wood structure and made the foundation of wood image recognition.It showed that statistics classification is effective in classifying wood type based on above extracted feature. In statistics classification, analysis of variance were applied to analyze the coefficient of variation of area of pores. The result showed that the coefficient of variation of areas have prominent difference between ring porous wood and diffuse porous wood for it's higher than 0.1 in ring porous wood while it's lower than 0.1 in diffuse wood, and the coefficient in semi-ring porous wood (or semi-diffuse porous) is in the middle of that of ring porous and diffuse wood.After study and training on characteristic parameter, it showed that the coefficient of variation of area, density of distribution of pores and tissue proportion have high relation with wood type recognition, and the algorithm is effective for classifying softwood and hardwood, ring porous wood, semi-ring porous (or semi-diffuse porous), diffuse porous wood.A image processing system for wood tissues and wood structure...