Particle-Size Analysis of Wood Fiber and Powder Based on Image Processing and Recognition

Particle-Size Analysis of Wood Fiber and Powder Based on Image Processing and Recognition

Honge Ren (Information and Computer Engineering College, Northeast Forestry University, Harbin, China), Jian Zhang (Information and Computer Engineering College, Northeast Forestry University, Harbin, China), Meng Zhu (Information and Computer Engineering College, Northeast Forestry University, Harbin, China) and Mian Liu (Information and Computer Engineering College, Northeast Forestry University, Harbin, China)
Copyright: © 2018 |Pages: 14
DOI: 10.4018/JITR.2018070108

Abstract

This article describes how to realize a wood fiber and powder automatic production and detection, using a mesh detection method based on image processing and particle-size analysis. With this method, the image will be transformed into an HSV color space, segmented based on S component with OSTU; calculated and analyzed with shape factor F, then rectangle fit, and finally the rectangle's length is converted into mesh. During wood flour mesh recognition, this article proposes the use of a rectangle fitting algorithm. In accordance with the actual demand for mesh recognition, a micro-nano wood flour mesh recognition system based on a modular approach has been designed and developed. This is as well as a mesh recognition design program based on wood flour features. The experiments and particle-size analysis results demonstrate that compared with traditional methods, the proposed approach is of higher accuracy and appropriate for extension.
Article Preview

2. Wood Powder Image Processing And Recognition

2.1. Particle Feature Extraction

Target recognition accuracy directly depends on image segmentation whose algorithms usually aim at gray images that may lose much color information. This paper adopts HSV color space that fits eyes sensory and has the independence and uniformity of color image processing. After analyzing HSV’s component images, S component shows enough saturation information, has the biggest contrast between the target and background and better highlights complete particles. So segment the image by OSTU and area threshold filtering based on S component.

According to the inherent characteristics, the images can be divided into two categories: the image characteristics of texture feature description object shape description of the object surfaceshape feature and gray change. Wood powder were mainly based on the mesh shape feature recognition of wood powder, according to the research needs of this paper, the extraction of the wood particles of the perimeter, area.

Perimeter is represented by the symbol P, the calculation formula is as follows.

(1) where Nl is a number of wood flour on the boundary, Nh is a number of wood powder boundary.

The area is represented by the symbol A, the calculation formula is as follows.

(2) where f (i, j) is image, MN Is the side length of image.

In consideration of morphological feature, shape factor F has been used.

(3) where P denotes the perimeter of the particle projection region; A denotes the area of the particle projection region.

With a comparison and analysis about various data, it’s found that F has a close contact with a particle’ projection shape:

  • 1.

    When F is some values such as 1.275, 1.927, 2.160 and so on, its projection shape is close to ellipse-like; after a detailed observation and analysis on wood powder samples, the bigger the gap between the long diameter and the short diameter of the ellipse is, the closer the ellipse is to a oblong, namely a rounded rectangle;

  • 2.

    The greater F is, the closer wood powder projection shape is to a needle-like shape, which appears to be rectangle-like under the microscope;

  • 3.

    As for a regular rectangle, the greater the value of F is, the more the shape of a particle is out of roundness and the bigger its surface’s acute angle is.

Complete Article List

Search this Journal:
Reset
Open Access Articles
Volume 13: 4 Issues (2020): Forthcoming, Available for Pre-Order
Volume 12: 4 Issues (2019): 3 Released, 1 Forthcoming
Volume 11: 4 Issues (2018)
Volume 10: 4 Issues (2017)
Volume 9: 4 Issues (2016)
Volume 8: 4 Issues (2015)
Volume 7: 4 Issues (2014)
Volume 6: 4 Issues (2013)
Volume 5: 4 Issues (2012)
Volume 4: 4 Issues (2011)
Volume 3: 4 Issues (2010)
Volume 2: 4 Issues (2009)
Volume 1: 4 Issues (2008)
View Complete Journal Contents Listing