Research On Image Segmentation And Visualization Of Cell Wall Structure Of Vascular Bundles In Phyllostachys Pubescens Culm

Usually,the physical or chemic methods are used to study bamboo cell wall crystal, but these methods are not intuitionistic, we can't directly acquired cell wall crystal distribution. As the development of computer image processing and artificial intelligence technique, we take advantage of realizing this aim.In this paper, the acquiring methods on vascular bundle crystal image in the transverse section of Phyllostachys Pubescens culms are introduced. Applying digital image processing technique, we realize good segmentation of the vascular bundle crystal image, Based on the segmented image, we study vascular bundle visualization structure by applying 3D reconstruction technique.Phyllostachys Pubescens cell wall includes vascular bundles cells and parenchyma cells. Vascular bundle cells arrange closely, and that parenchyma cells arrange loose and it looks like net. According to their characteristics, we present two image segmentation algorithms based on fuzzy inference and shape feature respectively, and combine above two algorithms to realize vascular bundles crystal image segmentation. Firstly, we make primary segmentation by using above first algorithm. After first segmented, the most of parenchyma cells image is eliminated, vascular bundle crystal image becomes a single integrated unit. Secondly, we get whole vascular bundle crystal image by using above second algorithm based on primary segmentation. Algorithm software is programmed with VC++ 6.0 in Windows XP Operation System. The testing result shows that combining above two algorithms we can segment vascular bundle crystal image accurately, no matter what vascular bundle crystal image is one or not. It lays a good foundation for the later measurement of bamboo crystallinity and vascular bundle shape parameters.Besides, a primary research on 3D visualization of vascular bundle structure is done. Firstly, two-dimensional images of Phyllostachys Pubescens cell wall are gained with Optical Coherence Tomography(OCT) system of Tsinghua University. Secondly, above two image segmentation algorithms are used to segment all two-dimensional images. At last, a 3D reconstruction is done by 3D-Doctor software.