Research On Homogenization Model Of SCFRW Based On Digital Image Processing

As the global forest resources declined, timber resources are becoming scarcer and scarcer, especially for such a lack of forest country. How to effectively improve the utilization of timber resources and value in use has become a hot spot concerned and researched by domestic and broad wood industry researchers. High-performance, high value-added functional wood composites provide a new direction for the research and application for it. This paper mainly studies features of Short Carbon Fiber Reinforced Wood (SCFRW). Compounding carbon fiber and wood fiber effectively gives the composite material strong mechanical properties, while bring it good conductivity and electromagnetic shielding properties. In the preparation process of composite materials, the mixed homogenization degree of discontinuous carbon fiber and the matrix material can seriously impact on the problem of macroscopic properties. So this paper discussed the homogenization model of SCFRW based on digital image processing technology.First, testing physical performances of samples obtained in the experiments, including electrical conductivity (surface resistivity), mechanical properties (density, internal bond strength, modulus of rupture, modulus of elasticity, thickness swelling rate of water absorption), electromagnetic shielding, and analyzing test data, we grasped the characteristic variation law of short carbon fiber reinforced wood in different preparation conditions, and studied the effecting characteristics of the micro-homogenization degree to the macro-performances of composite material; And then, the microscopic images were acquired from the testing samples, and 150 microscopic images were selected, which had the same magnification and typical shape features, to do image processing. On the basis of image preprocessed, this paper adopted the method combining maximum variance and mathematical morphology for image segmentation, so as to obtain binary images of good segmentation with carbon fiber as the target area. Select shape features (area ratio, average width and aspect ratio, etc.) which can effectively reflect the homogenization degree of samples as the characteristic quantity for extracting characteristic values; Finally, using the minimum width of single bar area, the average area ratio, the average aspect ratio, the total area ratio of bar area as inputs, using the surface resistivity, modulus of elasticity, modulus of rupture as outputs, homogenization BP neural network model of short carbon fiber reinforced wood functional composite material was established. The validity and reliability of the model was verified in the MATLAB environment.Through this research, we could organically combine the microstructure characteristics and macroscopic properties of composites, and achieved the purpose which we directly predicted the macro-level characterization performance according to the mixed homogenization degree of woodiness fibers and carbon fibers in SCFRW, so as to provide the favorable basis for the making-board technology design. Based on this, provide the necessary basic conditions and scientific guidance for the preparation and adaptability use of carbon fiber reinforced wood functional composite material.