Wood Species Recognition And Wood Defect Detection Based On Multispectral

Wood is the raw material of wood products,and also a very important renewable resource in our daily life.Recently,People's demand for wood products has soared.The detection of wood and wood products has become a hot research project for the blazing hand.Wood is the most demanding variety in the process of making wooden products.It is not only to identify the type of plank but also to identify the defects on wood surface in the daily production using.With the development of economy and the progress of society.The traditional reliance on manual experience and knowledge has been unable to meet the requirements of the high tempo and modern wood industry.In order to adapt to the requirement of high efficiency,higher recognition rate and short training time,computer aided recognition emerges.This paper presents a method for identifying tree species by spectral differentiation.In the experiment,five types of trees were selected in the north,such as birch,poplar,white pine,Pinus sylvestris,Larch.The experimental instrument selected is the Fieldspec Profr produced by ASD(measuring wavelength range is 350)to collect the spectral reflectance curve of five kinds of plank surface,using 10 nm to collect spectral reflection curve.The spectral differential method is used to select the feature of the band,and the effect of five species is identified by using Euclidean distance to test the spectral differential method.The boards to be processed will often have a portion of the surface with knots,and it is difficult for the naked eye to determine whether it is a normal plate when the knots of the surfaces of the boards are covered with dyes(such as paints)of the same or similar color.In this paper,a new method for surface defect detection of wood board is proposed by image fusion technique.The near infrared image and the visible image are collected by using the weighted average method,PCA algorithm,wavelet transform,Laplacian pyramid transformation,and so on.Then,the fusion of different algorithms After the image is carefully observed and compared and calculated by its information entropy.