Microscopic Image Feature Extraction Method Of Wood Powder Particles And Mathematical Models Research Of Particle Size

As a new type of energy-saving and environmental protection material, wood powder is being used widely. The wood powder is made into the field of composite materials and polymer materials and then the new environmental protection materials --wood plastic composite materials been appearing on the timber industry. Wood plastic composite materials both have the performance of the log materials and polymer materials, their production and use is rapid increase year by year at the same time the performance requirements is becoming more and more high. The size and shape of filling particle are the most important characters of wood plastic composite materials, they influence their mechanical properties directly. This paper doing the research from the the particle size and the aspect ratio of wood powder,then to the establishment of the mathematical model, then to finding the relationship and the reason of this relationship between particle size and aspect ratio, all of this research is significant to the improve of mechanical properties of wood plastic composite materials.The subject of this paper is supported by "Harbin Research Special Funds of Science and Technology Innovative Talents" and is part of "wood ultrafine crushing air classifier modeling and computer mesh recognition". The task is measuring the particle size of wood power and rhe length and width of the tracheid through of microscope; getting the date of the parameter of geometric features of wood power through digital image processing technology; establishing the mathematical model of the particle size; designing "The Wood Particle Real-time Testing Software". The main work of this paper is as follows:(1) The traditional way of put wood particles is manual which is time-consuming and poor effect, according to this shortcomings this paper proposed the use of solution system which water and glycerol is 1:1 as the solvent of wood power. To achieve the purpose of t wood power dispersed uniform and not easy to be influenced by environment. Completed the task of measuring the particle size of wood power and rhe length and width of the tracheid.(2) According to the characters of wood whose edge is complex, size is small, quality is light and the characters of microscopic image that is easily influenced by lighting source, this paper proposed the image processing algorithm of wood power microscopic image based on the Lab space. And finally achieved the segmentation and feature extraction of images. The possess of the algorithm:convert the color space from RGB to Lab of the original microscopic image and extract the b component; median filter on the b component; the denoised image is clustering for Class 2 with the K-means algorithm and then get binary image of flour; binary image were first opened after closing operations of mathematical morphology with 5×5 structure elements; mark each single flour particle by eight connected region labeling method; calculated the geometric area of wood flour by the method of the number of target pixels in statistics and the length, width, area of the minimum bounding rectangle of wood flour by Spindle Law on the labeled image; obtain the data of area, perimeter, equivalent diameter, density, aspect ratio and squareness of each single particle flour.(3) Fitting curve between particle size and aspect ratio is created and optimized with the least squares method, Gaussian mathematical model of and aspect ratio is obtained by analyzing and evaluating the fit function of polynomial, and Fourier and Gaussian, then the second derivative calculation is obtained according to its fitting curve equation. Analysis and discuss the Gaussian model combined with the second derivative and wood tracheid data and the cause of the trend of model fundamentally.(4) To achieve the purpose of monitoring of grinding process of wood power real-timely, this paper designed "The Wood Particle Real-time Testing Software". Completed the function of acquisiting microscopic image of wood power, processing the image, extracting geometric parameters of wood power, calculatting the particle size of wood flour.