Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning
For many decades, ultrasonic imaging inspection has been adopted as a principal method to detect multiple defects, e.g., void and corrosion. However, the data interpretation relies on an inspector’s subjective judgment, thus making the results vulnerable to human error. Nowadays, advanced...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2018-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/18/11/3820 |
| _version_ | 1828281528761712640 |
|---|---|
| author | Jiaxing Ye Shunya Ito Nobuyuki Toyama |
| author_facet | Jiaxing Ye Shunya Ito Nobuyuki Toyama |
| author_sort | Jiaxing Ye |
| collection | DOAJ |
| description | For many decades, ultrasonic imaging inspection has been adopted as a principal method to detect multiple defects, e.g., void and corrosion. However, the data interpretation relies on an inspector’s subjective judgment, thus making the results vulnerable to human error. Nowadays, advanced computer vision techniques reveal new perspectives on the high-level visual understanding of universal tasks. This research aims to develop an efficient automatic ultrasonic image analysis system for nondestructive testing (NDT) using the latest visual information processing technique. To this end, we first established an ultrasonic inspection image dataset containing 6849 ultrasonic scan images with full defect/no-defect annotations. Using the dataset, we performed a comprehensive experimental comparison of various computer vision techniques, including both conventional methods using hand-crafted visual features and the most recent convolutional neural networks (CNN) which generate multiple-layer stacking for representation learning. In the computer vision community, the two groups are referred to as shallow and deep learning, respectively. Experimental results make it clear that the deep learning-enabled system outperformed conventional (shallow) learning schemes by a large margin. We believe this benchmarking could be used as a reference for similar research dealing with automatic defect detection in ultrasonic imaging inspection. |
| first_indexed | 2024-04-13T08:16:31Z |
| format | Article |
| id | doaj.art-15f9b41b316e407282b69b11a19f4986 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-13T08:16:31Z |
| publishDate | 2018-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-15f9b41b316e407282b69b11a19f49862022-12-22T02:54:45ZengMDPI AGSensors1424-82202018-11-011811382010.3390/s18113820s18113820Computerized Ultrasonic Imaging Inspection: From Shallow to Deep LearningJiaxing Ye0Shunya Ito1Nobuyuki Toyama2National Metrology Institute of Japan (NMIJ), National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 2, Tsukuba 305-8568, JapanNational Metrology Institute of Japan (NMIJ), National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 2, Tsukuba 305-8568, JapanNational Metrology Institute of Japan (NMIJ), National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba Central 2, Tsukuba 305-8568, JapanFor many decades, ultrasonic imaging inspection has been adopted as a principal method to detect multiple defects, e.g., void and corrosion. However, the data interpretation relies on an inspector’s subjective judgment, thus making the results vulnerable to human error. Nowadays, advanced computer vision techniques reveal new perspectives on the high-level visual understanding of universal tasks. This research aims to develop an efficient automatic ultrasonic image analysis system for nondestructive testing (NDT) using the latest visual information processing technique. To this end, we first established an ultrasonic inspection image dataset containing 6849 ultrasonic scan images with full defect/no-defect annotations. Using the dataset, we performed a comprehensive experimental comparison of various computer vision techniques, including both conventional methods using hand-crafted visual features and the most recent convolutional neural networks (CNN) which generate multiple-layer stacking for representation learning. In the computer vision community, the two groups are referred to as shallow and deep learning, respectively. Experimental results make it clear that the deep learning-enabled system outperformed conventional (shallow) learning schemes by a large margin. We believe this benchmarking could be used as a reference for similar research dealing with automatic defect detection in ultrasonic imaging inspection.https://www.mdpi.com/1424-8220/18/11/3820nondestructive evaluationultrasonic imagingcomputer visiondeep learninglocal descriptorconvolutional neural networks |
| spellingShingle | Jiaxing Ye Shunya Ito Nobuyuki Toyama Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning Sensors nondestructive evaluation ultrasonic imaging computer vision deep learning local descriptor convolutional neural networks |
| title | Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning |
| title_full | Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning |
| title_fullStr | Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning |
| title_full_unstemmed | Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning |
| title_short | Computerized Ultrasonic Imaging Inspection: From Shallow to Deep Learning |
| title_sort | computerized ultrasonic imaging inspection from shallow to deep learning |
| topic | nondestructive evaluation ultrasonic imaging computer vision deep learning local descriptor convolutional neural networks |
| url | https://www.mdpi.com/1424-8220/18/11/3820 |
| work_keys_str_mv | AT jiaxingye computerizedultrasonicimaginginspectionfromshallowtodeeplearning AT shunyaito computerizedultrasonicimaginginspectionfromshallowtodeeplearning AT nobuyukitoyama computerizedultrasonicimaginginspectionfromshallowtodeeplearning |