Research On Extraction Technology Of Wood Axial Parenchyma Based On Computer Vision

With the development of modem wood industry,wood as the basic form of wood industrialproduct,different species of wood is frequently used in people’s lives.However,they have a lot ofdifferences between the timber species,it is difficult to recognize them.Therefore,we should solvethe problem how to recognize wood accurately,by using auxiliary equipment in the preconditionthat do not destroy product.And the problem becomes the research direction at present.Axial parenchyma has important wood texture information,and it is one of the importantcharacteristics of wood.The major Purpose of this dissertation is to study hardwood microstructureimages.Taken for the image processing technology,we have distilled axial parenchyma.Firstly,denoising of axial parenchyma images can eliminate the influence of noise,and also facilitate theextraction of axial parenchyma.Then, mathematical morphology methods are applied toprocessing image.The form of axial parenchyma and pores is obtained from the cross-sectionimage of hardwood successfully.Finally,the axial parenchyma will be separated from pores bycalculating the enclosed area.In order to verify the feasibility and effect of the presented method,the algorithm wasimplemented using Matlab language.Several experiments were carried out.The simulationexperiment shows: median filter algorithm is used in image pretreatment.Using9*9as thetemplate has good effect.Block two value can help extraction of axial parenchyma.In themathematical morphology processing,we use crisscross structure element and and disk structureelement to process the image.The form of axial parenchyma and pores is obtained from thecross-section image of hardwood successfully.We use eight neighborhood and area histogram todetermine the area threshold.If so,it can separate the axial parenchyma from the catheter.At present,we use the mothed to do some experiments in10different tree species.Theexperimental results are very good.This has fully demonstrated the effectiveness of the proposedalgorithm.The realization of the method will pay a foundation for research of intelligent woodidentification based on axial parenchyma,provide a shortcut for measuring the content of axialparenchyma of different trees species,made the preparation for developing wood identificationsystem of axial parenchyma based on image.