Enhancing Fault Detection in Image Analysis: A Combined Wavelet-Fourier Technique for Advancing Manufacturing Quality Control

This study focuses on utilizing image data for statistical process control and improving quality monitoring in manufacturing and service systems. The effectiveness of individual and combined feature extraction methods is evaluated, with the Wavelet-Fourier approach identified as the most suitable. The proposed method not only identifies image processing issues but also provides valuable information for estimating change points, fault locations, and fault sizes. This enables the resolution and prediction of faults, leading to cost and time savings in production. To perform evaluation of the proposed method, an image from a tile production line is subjected to Wavelet transform, followed by Fourier transform on the obtained coefficients. The results demonstrate the superiority of the Wavelet-Fourier method over individual methods such as Fourier transform and Wavelet transform. The proposed method exhibits comparable or improved performance in fault detection and localization compared to similar research. This study highlights the potential of utilizing image data for statistical process control and quality monitoring, offering a comprehensive solution for fault detection and analysis. The findings contribute to advancements in image processing techniques and have practical implications for enhancing quality monitoring in various industries. By leveraging image data, manufacturers can make informed decisions, enhance process performance, and improve overall product quality.

Keywords:

Quality performance , Imaging technologies , Feature Extraction , Statistical process control , Online process monitoring , Wavelet-Fourier Method , Fault detection and localization