Sub-Pixel Level Defect Detection Based on Notch Filter and Image Registration

Guo, Lei; Li, ShiNan; Hu, Wenjing; Wu, Jian; Tu, Bing; He, Wei; Ou, Xianfeng; Zhang, Guoyun

doi:10.1142/s0218001418540162

Cited by 9 publications

(3 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The emergence of traditional vision methods comes from manual visual inspection, which mainly uses machines to replace human eyes for defect detection. Traditional image processing operators are used to learn some features [8] and process the target, such as image enhancement, image segmentation, image matching, and other methods. At present, some traditional vision-based methods mainly use these algorithms or improve these algorithms for defect detection.…”

Section: Traditional Visual Methodsmentioning

confidence: 99%

Strip steel surface defect detecting method combined with a multi-layer attention mechanism network

Liu

2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

In the production of strip steel, defect detection is a crucial step. However, current inspection techniques frequently suffer from issues like low detection accuracy and subpar real-time performance. We provide a deep learning-based strip steel surface defect detection technique to address the aforementioned issues. The algorithm is also implemented in three specific ways: as the backbone, the neck, and the detection head. Backbone employs an enhanced cross stage partial in conjunction with ResNet to effectively mine defect information and minimize the issue of adjoining feature maps' neighboring feature maps losing information. Neck is a better structure, we propose and use the Dilated Weighted Across Stages-feature pyramid network (DWAS-FPN) in the network to adjust the receptive field and attention weight preference of the output feature maps at different scales and to improve the utilization of defect features by the algorithm to enhance the detection of abnormal size defects. We use four detection heads in the detection head so that the network can learn the features of defects of various sizes. Finally, we use the decoupled head to separate the classification work from the regression work before combining the prediction. Two datasets of surface flaws in strip steel are used in our experiments (GC10-DET and NEU-DET). In addition, it has been shown that our proposed algorithm's mAP in GC10-DET and NEU-DET reaches 79.93% and 72.76%, respectively, resulting in a better detection impact.

show abstract

Section: Traditional Visual Methodsmentioning

confidence: 99%

Strip steel surface defect detecting method combined with a multi-layer attention mechanism network

Liu

2023

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Then, the second-level HT achieves an accurate center at the pixel level, with little time consumption. Finally, the least-squares fitting method is adopted to upgrade the positioning accuracy to the sub-pixel level 20 . In general, the TSHT algorithm avoids discrete coordinates and peak diffusion, greatly enhancing the credibility of the test results.…”

Section: Methodsmentioning

confidence: 99%

Center detection algorithm for printed circuit board circular marks based on image space and parameter space

Wang

Chen

et al. 2022

J. Electron. Imag.

View full text Add to dashboard Cite

A highly efficient circle positioning algorithm, called the two-step optimization Hough transform (TSHT), based on multi-resolution segmentation is proposed to solve the problems of the offset Hough transform, namely, its large memory overhead, long time consumption, and low recognition accuracy. First, using the image feature of the printed circuit board (PCB) circular identifier, the target circle is obtained using adaptive image preprocessing, and then, images of an acceptable quality are separated by shape quality inspection to improve their robustness. Second, using effective interval sampling strategies and gradually controlling the accumulative interval of parameters, the TSHT algorithm reduces the memory overhead and quickly locates the center at the pixel level. Finally, the center at the sub-pixel level is found by the least-squares method for circle fitting. The experiments prove that TSHT, as a result of its high robustness, strong anti-noise capability, fast recognition speed, and accuracy, can be successfully applied to a vision positioning system of a solder paste printing machine.

show abstract

“…Few studies have focused on the detection of defective pixels in TFT-LCD panels. Guo et al [15] developed a process that used notch filter and threshold segmentation method to detect subpixel defects on LCD screen images. Çelik et al [16] proposed a real-time defective pixel detection system for LCD TV products using a deep learning algorithm.…”

Section: Introductionmentioning

confidence: 99%

A Novel Multicategory Defect Detection Method Based on the Convolutional Neural Network Method for TFT-LCD Panels

Chang

Meng

et al. 2022

Mathematical Problems in Engineering

View full text Add to dashboard Cite

Defects on thin film transistor liquid crystal display (TFT-LCD) panel could be divided into either macro- or microdefects, depending on if they are easy to be detected by the naked eye or not. There have been abundant studies discussing the identification of macrodefects but very few on microones. This study proposed a multicategory classification model using a convolutional neural network model to work with automatic optical inspection (AOI) for identifying defective pixels on the TFT-LCD panel. Since the number of nondefective pixels outnumbered the defective ones, there exists a very serious class-imbalanced problem. To deal with that, this study designed a special training strategy that worked with data augmentation to increase the effectiveness of the proposed model. Actual panel images provided by a mobile manufacturer in Taiwan are used to demonstrate the efficiency and effectiveness of the proposed approach. After validation, the model constructed by this study had 98.9% total prediction accuracy and excellent specificity and sensitivity. The model could finish the detection and classification process automatically to replace the human inspection.

show abstract

Sub-Pixel Level Defect Detection Based on Notch Filter and Image Registration

Cited by 9 publications

References 11 publications

Strip steel surface defect detecting method combined with a multi-layer attention mechanism network

Strip steel surface defect detecting method combined with a multi-layer attention mechanism network

Center detection algorithm for printed circuit board circular marks based on image space and parameter space

A Novel Multicategory Defect Detection Method Based on the Convolutional Neural Network Method for TFT-LCD Panels

Contact Info

Product

Resources

About