Publication: Detection of label printing defects on ssd drive using image processing approaches
Loading...
Date
2023-07
Authors
Detection of label printing defects on SSD drive using image processing approaches
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Solid-State Drives (SSDs) are integral storage devices that utilize integrated circuit assemblies to permanently store data. SSDs come with diverse specifications, performance characteristics, and safety information, depending on their intended uses. Consequently, the accuracy of the labels on SSDs becomes crucial as they provide vital information about the components. Unfortunately, defects in labels are challenging to avoid during industrial production. To address this issue, the detection of defective labels is necessary to prevent products with such labels from entering the market. This study focuses on employing image processing techniques to detect label defects in SSDs manufactured by Western Digital. The primary image processing approaches utilized in this research are image preprocessing, template matching, and feature extraction. Initially, image preprocessing is performed to enhance the quality of the images. Next, template matching is employed to locate predefined characteristics of the labels accurately. Subsequently, feature extraction is applied to extract relevant information from the labels. After the feature extraction process, the Euclidean distance is calculated to compare the label histograms and distinguish between good and defective labels. The acquired data is then compared with reference samples provided by Western Digital to evaluate the performance of the system. For constant feature detection, the proposed system has produced 53% of measurement accuracy, while for number code defect detection, 78% of accuracy has been achieved. With these low accuracies, improvement has been done to ensure that image
processing approach is suitable to be applied for SSD defects detection.