Moiré Pattern-Based Stripe Density Measurement Using Low-Frequency Component Detection

This article presents the measurement of stripe density using image processing. The two essential techniques are a digital grid overlay on the target image to create a moiré pattern and a moving average filter to remove the high-frequency components of the moiré pattern image. The moving average filter produces a low-frequency signal whose frequency is equal to the difference between the frequency (density) of the target stripes image and the frequency of the overlaid grid stripes. We use a method where grid lines are drawn with a lower density on the left side and progressively increase in density towards the right, surpassing the target stripe density. When superimposed on the target image, if the frequency of the grid stripes matches the frequency of the target stripes, the moving average filter will produce an output signal with a frequency equal to zero. An automatic stripe density measurement system is achieved by detecting situations where the frequency value of the output signal from the moving average filter equals zero. The results of measuring the density of simulated stripes created with the program for densities ranging from 30-180 per unit length (L−1) and the number of threads of the fabric on six images showed that the system could measure the actual values consistently, with an error of less than ±1 L−1.