Development of an Intelligent and Automated Detection Systems for Functional Clothing Labels

• Main Authors: Yu-Wei Hsu, 許祐瑋
• Other Authors: Pin-Chuan Chen
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/ayj7aq

碩士 === 國立臺灣科技大學 === 機械工程系 === 107 === This study successfully developed an optical detection system for functional clothing labels. The system hardware includes a motor-driven transport platform with an image capture module. Also explore the impact of the light source on the detection, and then select the desired light source. The software is written based on the reason the label is judged to be defective, including deformation, defects, and underline marks. For deformation and defect labels, the system uses the reference subtraction method for detection. A label that is judged to be qualified by the human eye as a standard film. Then, overlapping and subtracting the standard label and the label to be tested. Divide the image into front end, tail end, and all detections, and judge from the remaining pixel values. If the label is severely deformed, the position of the center point will be far apart, and the remaining pixel values will be larger. When the judgment threshold of each area is exceeded, the defect can be effectively determined. The label of the underline mark is not easily visible under the illumination of white light. Therefore, it is attempted to use different color lights for illumination, and the experimental results are most appropriate for blue light. However, the small underline mark at the tip of the label is still difficult to display, so different image processing algorithms are written for detection in the area. Finally, samples including normal and defected parts have been tested to verify the capability of the inspection system developed.