A New Image-based Safety Monitoring Technique with Applications to a Steel Coil Painting Process

Abstract

This article develops a brand new technique to perform safety monitoring using images for a type of complex industrial processes. The new safety monitoring technique is developed with the aid of a modified nonnegative matrix factorization (NMF) algorithm that is called multiple-centroid strictly convex nonnegative matrix factorization (MCSCNMF). Specifically, MCSCNMF is designed deliberately in an all-new manner such that it can learn more accurate centroids for each group of samples than the existing NMF-like algorithms. Moreover, MCSCNMF takes advantage of multiple centroids instead of single centroid to describe the complicated distribution of each type of samples. Both properties give MCSCNMF a powerful clustering performance and make it beneficial for developing high-performance safety monitoring techniques than the existing NMF-like algorithms. We compare different types of monitoring methods in terms of false negative, false positive, and average approximation error in an experiment on a steel coil painting process. Comparison results demonstrate that the MCSCNMF-based monitoring technique can obtain the best experimental results compared with other monitoring methods all the time.