| In the production process of particleboard, dosing and applying glue system played important roles. It effected product quality and production cost directly. The inappropriate ratio of glue not only effected mechanical properties, but also wasted the resource. The uniformity of sizing was an important index. In this article, image processing technology was used to quantitatively analyze the distribution of glue and evaluate sizing result by microstructure pictures of particleboard.Median filter was used to remove part of interference noise in the preprocessing of particleboard images. Gradient method was used to sharpen the images and enhance the edge of images. All those prepared for the next image processing.Traditional edge detection operators were used to divide the images of particleboard, the difference between various classic operators wad compared to analyze the advantages and disadvantages of the operators on division the images. Based on the shortage of the traditional edge detection operators, the morphological edge detection was introduced, which combined with OSTU to segment images effectively.In this article, the feature extraction and image analysis were studied. Based on shape features, feature extraction adopted geometric characteristics in the region and adopted chain code on the edge of the region. Four characteristic values were proposed to describe the adhesive distribution, including coverage, evenness, diameter of glue droplets and shape of glue droplets. Coverage and homogeneity were described by the geometric characteristics of binary image; diameter and shape of glue droplets were described by 8 chain code of the segmentation image. For the particleboard with sizing volume of 3%, 4% and 6%, the sizing efficiency was described by the four characteristic values respectively.At last, the stress analysis on different sizing of particleboard was done to analyze the relationship between coverage and mechanical properties of adhesive.
|
PDF Full Text Request