earticle

영어

The cell slices of 5 species of common coniferous wood, including Cupressaceae, Abies nephrolepis, Picea jezoensis var. microsperma, Pinus koraiensis and Pinus massoniana and so on, are analyzed for the establishment of a hexagonal shape to simulate the wood cell structure. By using the image processing technology, the relevant parameters are extracted, including quasi-circular degree, the surface area, perimeter, width and thickness of cells, wall thickness, ratio of wall to cavity, ratio of cavity to diameter and cell density. In the experiment, mathematical analysis was made on 5,000 cells of the 5 species, and the cell sampling section was determined. After sampling, with the parameter values of cells obtained compared to the standard values, the error is within 10%. Afterwards, the value range of the cell parameters of each species is determined, which provides data basis for the recognition of wood species. Experiments show that the algorithm performs with better robustness in the recognition of wood species.