Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain
Capacitive touch panels (CTPs) have the merits of being waterproof, antifouling, scratch resistant, and capable of rapid response, making them more popular in various touch electronic products. However, the CTP has a multilayer structure, and the background is a directional texture. The inspection w...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-02-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/3/1737 |
| _version_ | 1797623155773145088 |
|---|---|
| author | Hong-Dar Lin Huan-Hua Tsai Chou-Hsien Lin Hung-Tso Chang |
| author_facet | Hong-Dar Lin Huan-Hua Tsai Chou-Hsien Lin Hung-Tso Chang |
| author_sort | Hong-Dar Lin |
| collection | DOAJ |
| description | Capacitive touch panels (CTPs) have the merits of being waterproof, antifouling, scratch resistant, and capable of rapid response, making them more popular in various touch electronic products. However, the CTP has a multilayer structure, and the background is a directional texture. The inspection work is more difficult when the defect area is small and occurs in the textured background. This study focused mainly on the automated defect inspection of CTPs with structural texture on the surface, using the spectral attributes of the discrete cosine transform (DCT) with the proposed three-way double-band Gaussian filtering (3W-DBGF) method. With consideration to the bandwidth and angle of the high-energy region combined with the characteristics of band filtering, threshold filtering, and Gaussian distribution filtering, the frequency values with higher energy are removed, and after reversal to the spatial space, the textured background can be weakened and the defects enhanced. Finally, we use simple statistics to set binarization threshold limits that can accurately separate defects from the background. The detection outcomes showed that the flaw detection rate of the DCT-based 3W-DBGF approach was 94.21%, the false-positive rate of the normal area was 1.97%, and the correct classification rate was 98.04%. |
| first_indexed | 2024-03-11T09:24:38Z |
| format | Article |
| id | doaj.art-ef69dbb1bebf490c89d92c1ed7a22887 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T09:24:38Z |
| publishDate | 2023-02-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-ef69dbb1bebf490c89d92c1ed7a228872023-11-16T18:05:27ZengMDPI AGSensors1424-82202023-02-01233173710.3390/s23031737Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine DomainHong-Dar Lin0Huan-Hua Tsai1Chou-Hsien Lin2Hung-Tso Chang3Department of Industrial Engineering and Management, Chaoyang University of Technology, Taichung 413310, TaiwanDepartment of Industrial Engineering and Management, Chaoyang University of Technology, Taichung 413310, TaiwanDepartment of Civil, Architectural, and Environmental Engineering, The University of Texas at Austin, Austin, TX 78712-0273, USADepartment of Industrial Engineering and Management, Chaoyang University of Technology, Taichung 413310, TaiwanCapacitive touch panels (CTPs) have the merits of being waterproof, antifouling, scratch resistant, and capable of rapid response, making them more popular in various touch electronic products. However, the CTP has a multilayer structure, and the background is a directional texture. The inspection work is more difficult when the defect area is small and occurs in the textured background. This study focused mainly on the automated defect inspection of CTPs with structural texture on the surface, using the spectral attributes of the discrete cosine transform (DCT) with the proposed three-way double-band Gaussian filtering (3W-DBGF) method. With consideration to the bandwidth and angle of the high-energy region combined with the characteristics of band filtering, threshold filtering, and Gaussian distribution filtering, the frequency values with higher energy are removed, and after reversal to the spatial space, the textured background can be weakened and the defects enhanced. Finally, we use simple statistics to set binarization threshold limits that can accurately separate defects from the background. The detection outcomes showed that the flaw detection rate of the DCT-based 3W-DBGF approach was 94.21%, the false-positive rate of the normal area was 1.97%, and the correct classification rate was 98.04%.https://www.mdpi.com/1424-8220/23/3/1737touch paneldefect detectiondirectional texturediscrete cosine transformfrequency domain filtering |
| spellingShingle | Hong-Dar Lin Huan-Hua Tsai Chou-Hsien Lin Hung-Tso Chang Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain Sensors touch panel defect detection directional texture discrete cosine transform frequency domain filtering |
| title | Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain |
| title_full | Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain |
| title_fullStr | Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain |
| title_full_unstemmed | Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain |
| title_short | Optical Panel Inspection Using Explicit Band Gaussian Filtering Methods in Discrete Cosine Domain |
| title_sort | optical panel inspection using explicit band gaussian filtering methods in discrete cosine domain |
| topic | touch panel defect detection directional texture discrete cosine transform frequency domain filtering |
| url | https://www.mdpi.com/1424-8220/23/3/1737 |
| work_keys_str_mv | AT hongdarlin opticalpanelinspectionusingexplicitbandgaussianfilteringmethodsindiscretecosinedomain AT huanhuatsai opticalpanelinspectionusingexplicitbandgaussianfilteringmethodsindiscretecosinedomain AT chouhsienlin opticalpanelinspectionusingexplicitbandgaussianfilteringmethodsindiscretecosinedomain AT hungtsochang opticalpanelinspectionusingexplicitbandgaussianfilteringmethodsindiscretecosinedomain |