![]() ![]() In large-scale industrial production, an optical character recognition (OCR) system-based computer vision approach is widely used to read serial numbers on the head slider acquired from more than a hundred slider attachment machines ( Figure 1B), with a readable rate of 99.87%. ( A) OCR image acquisition, ( B) OCR-based computer vision, and ( C) OCR-based concatenated deep learning technique. System configuration of the proposed approach for a defect inspection of serial numbers on the slider of HDD. The serial number is typically captured with a digital camera and saved as an image format ( Figure 1A). ![]() ![]() Serial numbers of 12-character length are composed primarily of both numbers (0–9) and letters (A–Z) (A to Z). The Auto ID system employs laser-based serial number printing on individual head sliders. As a result, automatic identification (Auto ID) of sliders is essential for tracking and identifying component process history, processing defects, and product recalls. During the mechanical slider assembly process, certain defects, such as contamination or scratches, are prevalent. Wafer fabrication and mounting, dicing, row chopping, and head parting were all part of the general manufacturing process. A circular wafer is sectioned into 700-micron pieces and tied to a suspension assembly. A slider was produced using a wafer made of aluminum and titanium. ![]() The head slider is a magnetic head sensor mounted to the tip of a suspension arm, also known as a head gimbal assembly, of a hard disk drive (HDD) that assists in reading and writing components while flying above the hard disk magnetic media surface ( Figure 1). The EfficientNet-B0 network outperformed highly qualified human readers with the best recovery rate (98.4%) and fastest inference time (256.91 ms). Experimenting on almost 15,000 photographs yielded accuracy greater than 99% on four CNN networks, proving the feasibility of the proposed technique. For character classification, four convolutional neural networks (CNN) were compared for accuracy and effectiveness: DarkNet-19, EfficientNet-B0, ResNet-50, and DenseNet-201. Our approach starts with image preprocessing methods that deal with differences in lighting and printing positions and then progresses to deep learning character detection based on the You-Only-Look-Once (YOLO) v4 algorithm and finally character classification. A deep-learning-based technique is proposed for determining the printed identity of a slider serial number from images captured by a digital camera. This paper outlines a system for detecting printing errors and misidentifications on hard disk drive sliders, which may contribute to shipping tracking problems and incorrect product delivery to end users. ![]()
0 Comments
Leave a Reply. |
Details
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |