Using ResNets to perform automated defect detection for Fluorescent Penetrant Inspection

Shipway, N.J.; Huthwaite, Peter; Lowe, M. J. S.; Barden, Tim

doi:10.1016/j.ndteint.2020.102400

Cited by 39 publications

(15 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, TFM images are often contaminated by noise and various strategies have been offered to modify the TFM algorithm to eliminate false indications [7,8] and reduce noise [8][9][10], enabling real-time imaging with portable NDT devices [8,11]. Researchers also began to explore application of machine learning to NDT [12][13][14][15]. However, at present, standard machine learning approaches have limited value: Firstly, most researchers have no access to big data such approaches require and even a few laboratory datasets used below have required a considerable effort and expense to collect.…”

Section: Introductionmentioning

confidence: 99%

Towards Explainable Augmented Intelligence (AI) for Crack Characterization

Fradkin¹,

Altinbasak²,

Darmon

2021

Applied Sciences

View full text Add to dashboard Cite

Crack characterization is one of the central tasks of NDT&E (the Non-destructive Testing and Evaluation) of industrial components and structures. These days data necessary for carrying out this task are often collected using ultrasonic phased arrays. Many ultrasonic phased array inspections are automated but interpretation of the data they produce is not. This paper offers an approach to designing an explainable AI (Augmented Intelligence) to meet this challenge. It describes a C code called AutoNDE, which comprises a signal-processing module based on a modified total focusing method that creates a sequence of two-dimensional images of an evaluated specimen; an image-processing module, which filters and enhances these images; and an explainable AI module—a decision tree, which selects images of possible cracks, groups those of them that appear to represent the same crack and produces for each group a possible inspection report for perusal by a human inspector. AutoNDE has been trained on 16 datasets collected in a laboratory by imaging steel specimens with large smooth planar notches, both embedded and surface-breaking. It has been tested on two other similar datasets. The paper presents results of this training and testing and describes in detail an approach to dealing with the main source of error in ultrasonic data—undulations in the specimens’ surfaces.

show abstract

Section: Introductionmentioning

confidence: 99%

Towards Explainable Augmented Intelligence (AI) for Crack Characterization

Fradkin¹,

Altinbasak²,

Darmon

2021

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…Dengan demikian, teknik ini pun hanya dapat digunakan untuk mendeteksi keberadaan retakan pada permukaan objek. Otomatisasi teknik penetrant testing dilakukan pada banyak industri dengan memanfaatkan perkembangan metode deep learning untuk meningkatkan keandalannya [3]. Teknik NDT yang dinamakan magnetic flux leakage (MFL) dapat mendeteksi keberadaan cacat pada permukaan dan sekitar permukaan objek terbuat dari logam.…”

Section: Pendahuluanunclassified

Electromagnetic Acoustic Transducer (EMAT) sebagai Transduser Ultrasonik untuk Nondestructive Testing (NDT): Ulasan Artikel

Nurmalia¹

2021

joki

View full text Add to dashboard Cite

Nondestructive testing (NDT) digunakan untuk mendeteksi keberadaan cacat pada suatu objek tanpa merusak objek tersebut, sehingga ideal untuk digunakan pada komponen-komponen infrastruktur yang sudah terpasang dan beroperasi. Diantara berbagai teknik, NDT ultrasonik memiliki kemampuan untuk menginspeksi cacat di sepanjang ketebalan objek, tidak hanya cacat permukaan. Metode pembangkitan gelombang ultrasonik yang paling populer adalah menggunaan piezoelectric transducer (PZT). Transduser ini memiliki efisiensi yang tinggi tetapi merupakan transduser kontak dan memerlukan pelumas (couplant), yang sering menyebabkan keterbatasan fleksibilitas untuk aplikasinya. Electromagnetic acoustic transducer (EMAT) adalah transduser ultrasonik yang dapat membangkitkan gelombang ultrasonik pada objek konduktif secara non-kontak. Transduser ini membangkitkan gelombang ultrasonik langsung pada objek uji, sehingga tidak memerlukan couplant. Hal ini menyebabkan EMAT ideal untuk inspeksi pada temperatur tinggi dimana couplant akan meleleh, dan pengujian yang bergerak. Dua jenis EMAT dibahas pada ulasan artikel ini: untuk membangkitkan gelombang shear horizontal (SH) dan gelombang torsional. Kedua jenis gelombang ini memiliki kelebihan dibanding moda gelombang lainnya, tetapi sulit untuk dibangkitkan menggunakan PZT. Kedua moda gelombang ini kemudian digunakan untuk menginspeksi penipisan pada pelat dan dinding pipa, yang dapat disebabkan oleh korosi maupun kontak mekanik. Hasil penelitian menunjukkan potensi EMAT dengan metoda konversi moda untuk inspeksi kuantitatif penipisan pada pelat dan pipa.

show abstract

“…The visualization of the test results and the maturity of the testing system are their main advantages. However, for these technologies, it is difficult to realize automatic inspection [ 7 , 8 , 9 ]. X-ray detection methods have high accuracy and high inspection efficiency.…”

Section: Introductionmentioning

confidence: 99%

Resolution Enhanced Array ECT Probe for Small Defects Inspection

Long

Zhang

Tao

et al. 2023

Sensors

View full text Add to dashboard Cite

It is a continual and challenging problem to detect small defects in metallic structures for array eddy current testing (ECT) probes, which require the probe to have ultra-high resolution and sensitivity. However, the spatial resolution of an ECT array probe is limited by the size of the induction coils. Even if it is possible to increase the spatial resolution by using smaller coils, the sensitivity of the sensor also decreases. To obtain finer spatial resolution without sacrificing sensitivity, this paper proposes a resolution enhanced ECT array probe with four rows of coils attached to a flexible printed circuit board (FPCB). The distance between each two adjacent coils in a row is 2 mm and the position of each row is offset by 0.5 mm along the horizontal direction related to its prior row. The outputs of the four rows are aligned and interpolated in a line, and in this way the image resolution of the probe is increased to 0.5 mm. The probe is configured to operate with the differential setting, namely two differential coils operate simultaneously at each time. The currents in the two coils can be controlled to have the same flowing direction or opposite flowing direction, resulting in different distributions of the induced eddy current and two sets of output images. A patch-image model and an image fusion method based on discrete wavelet transforms are employed to suppress the noise and highlight the defects’ indications. Experimental results show that small defects with dimensions as small as length × width × depth = 1 mm × 0.1 mm × 0.3 mm on a 304 stainless-steel sample can be detected from the fused image, demonstrating that the probe has super sensitivity for small defects inspection.

show abstract

Using ResNets to perform automated defect detection for Fluorescent Penetrant Inspection

Cited by 39 publications

References 21 publications

Towards Explainable Augmented Intelligence (AI) for Crack Characterization

Towards Explainable Augmented Intelligence (AI) for Crack Characterization

Electromagnetic Acoustic Transducer (EMAT) sebagai Transduser Ultrasonik untuk Nondestructive Testing (NDT): Ulasan Artikel

Resolution Enhanced Array ECT Probe for Small Defects Inspection

Contact Info

Product

Resources

About