Open AccessThe use of homomorphic image processing to analyze coke grading
Author(s) -
V. N. Krouglov,
A. G. Gruh,
A. L. Gapak,
I. Khurelchuluun
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/966/1/012135
Subject(s) - coke , computer science , fast fourier transform , image processing , fourier transform , algorithm , computer vision , mathematics , image (mathematics) , engineering , mathematical analysis , waste management
The estimation of the geometrical sizes of particles of crushed solid fuel (coke), moving on the conveyor belt, is associated with a number of technical difficulties. One of the problems is the need for a non-invasive way of determining particle geometry. A promising way to solve it is to use devices based on machine vision systems. This paper describes the algorithmic part of the prototype of such a device. It is proposed to improve the quality of boundary detection between fragments of coke particles to perform homomorphic processing of the initial low-contrast video images. The algorithm for calculating the Fourier spectrum has been optimized based on the Fast Fourier Transform (FFT) with the mixed base. As a result, it becomes possible to reduce the computational cost for calculating two-dimensional Fourier spectra for complex multiplication operations by 1.33 times, and the number of complex addition operations by 1.67 times. The software of the prototype, built using the proposed methods, made it possible to obtain good convergence of the results for assessing the particle size distribution of samples of crushed coke with laboratory estimates. Thus, the maximum absolute average error of the machine vision system in assessing the size of crushed coke is only 3.37%, and the maximum error for all measurement classes do not exceed 6.9%.