Development of an image analysis-based technique for precise evaluation of print mottle in the paper industry

Abstract

This research aims to develop a technique for evaluating print mottle, a critical defect that affects print quality, particularly in solid-printed areas. The proposed method combines image processing with a mathematical model aligned with the human visual system (HVS) to achieve greater accuracy and consistency than visual inspection alone. The experimental work comprised (1) the creation of simulated images and (2) the use of actual printed samples obtained from commercial printing facilities. These samples were analyzed using multiple techniques, and the resulting measurements were compared with expert visual assessments to determine correlation values. The results indicate that bandpass filtering and the modified coefficient of variation yielded correlations below 0.8, while the tile cell method achieved correlations in the range of 0.8-0.9. In contrast, an integration model in the form of a simple model achieved the higher correlation values of more than 0.9, indicating the closest alignment with human visual assessment. Furthermore, the proposed approach can be implemented using common equipment such as flatbed scanners and image-analysis software, making it highly suitable for quality control in paper and packaging production processes.