A review on pipeline integrity management utilizing in-line inspection data

Ultrasonic detection and characterization of targets concealed by scattering noise is remarkably challenging. In this study, a neural network (NN) coupled to split-spectrum processing (SSP) is examined for target echo visibility enhancement using experimental measurements with input signal-to-noise ratio around 0 dB. The SSP-NN target detection system is trainable and consequently is capable of improving the target-to-clutter ratio by an average of 40 dB. The proposed system is exceptionally robust and outperforms the conventional techniques such as minimum, median, average, geometric mean, and polarity threshold detectors. For realtime imaging applications, a field-programmable gate array (FPGA)-based hardware platform is designed for system-onchip (SoC) realization of the SSP-NN target detection system. This platform is a hardware/software co-design system using parallel and pipelined multiplications and additions for highspeed operation and high computational throughput.

/pdf/system-on-chip-design-for-ultrasonic-target-detection-using-4yn6ymey9v.pdf

System-on-chip design for ultrasonic target detection using split-spectrum processing and neural networks

Split-spectrum processing technique for SNR enhancement of ultrasonic guided wave.

Using ultrasonic arrays to detect and characterise surface breaking cracks is important in the non-destructive evaluation (NDE) field. It can provide early warning of failure and useful information for component integrity assessment. Typically, cracks are approximated by machined slots and used to examine and assess defect detection and characterisation methods. In this paper, real surface breaking cracks are fabricated in 3-point bending specimens following ASTM standard E1820 and used to examine the performance of two array defect characterisation methods: image-based and scattering matrix sizing. In both cases, an array is used to record the full matrix capture (FMC). In image-based sizing, the total focusing method (TFM) is used to form an image from which the defect size is measured directly. This approach is shown to work well for cracks greater than two wavelengths in size. The FMC is also used to extract the defect scattering matrix which is then compared to a pre-computed smooth-crack scattering matrix database. The best match between experiment and this database is found by cross-correlation and used to characterise the defect. This approach is shown to work well for defects in the range of 0.78–1.84 wavelengths. Within these ranges of applicability, both methods show excellent agreement between the known crack length and that measured ultrasonically, with errors less than 19% in all cases.

https://research-information.bris.ac.uk/files/165662689/PDF.pdf

The sizing of small surface-breaking fatigue cracks using ultrasonic arrays

In this paper we compare the performances of different variants of split-spectrum algorithms and propose some new extensions based on the use of variable bandwidth filters equally spaced in frequency and energy gain equalized. Signal-to-Noise Ratio Gain and Flaw-to-Clutter Ratio Gain factors were selected as the figures of merit to make the comparisons among the different methods. We considered simulated ultrasonic signals using both stationary and non-stationary models for the grain noise, and real scans obtained in laboratory from low dispersive (aluminum) and high dispersive (cement) materials. Frequency Multiplication (FM) recombination method is revealed as the best option when combined with the new extensions.

New analysis and extensions of split-spectrum processing algorithms

Ultrasonic flaw detection and classification is an essential method for many NDE applications. Frequency diverse algorithms such as split spectrum processing (SSP) can be used effectively for flaw detection since they decompose the signal in different subbands and induce significant statistical variation in scattering noise or speckles. Following the subband decomposition, neural networks can be used as post-processors for enhanced detection performance. In this study, it has been shown that using neural networks (NN), the flaw-to-clutter ratio (FCR) improvement up to 13dB can be achieved when the input experimental signal has FCR equal to 0dB or less. The experimental ultrasonic flaw signals masked by grain scattering noise (i.e., clutter) are obtained from a steel block using a 5MHz transducer and 100MHz sampling rate. Neural network algorithms are computationally demanding and require substantial hardware resources for implementation due to their complex connectivity. Therefore, we have designed an efficient and adaptable neural network architecture for real-time realization of ultrasonic flaw detection applications. The hardware implementation takes full advantage of the inherent parallelism of neural networks and is dynamically reconfigurable; indicating the system is flexible and can be suitable for various NDE applications

P1G-2 Dynamically Reconfigurable Neural Network Hardware Design for Ultrasonic Target Detection

In this study, a robust flaw detection algorithm using Neural Networks (NN) is presented for NDE applications. A three-layer feedforward NN which can perform a complex nonlinear mapping process has been used as a detection processor following the subband decomposition of the measured signal. The neural network architecture is trained to suppress the clutter echoes while maintaining the integrity of flaw echoes. The training process allows the neural network to learn about the statistics and the variation of the clutter signal. The robustness of the NN method is examined through testing materials with different grain sizes and multiple flaws. It has been shown that NN can improve the flaw-to-clutter (FCR) ratio significantly when the input experimental signal has FCR equal to 0 or less. Experimental results show that a typical FCR improvement of 40dB can be achieved using NN post detectors as opposed to 15dB with the conventional techniques including minimum, median, average, geometric mean and polarity detectors. The experimental results also confirm that the NN detector is capable of distinguishing two adjacent flaw echoes whereas the conventional techniques detect the presence of a single anomaly only. Furthermore, due its trainability, NN performs robustly when some of the subband signals used for detection have little or no flaw information.

P2D-3 Performance Evaluation of Neural Network Based Ultrasonic Flaw Detection

A theory-based approach to the usability of augmented reality technology: A cost-benefit perspective

이미 촬영된 동영상의 품질을 카메라의 물리적 한계를 극복하여 향상하는 연구가 활발하게 진행되어왔다. 특히 동영상 안정화는, 외부적인 요인에 의해 흔들리는 카메라로부터 획득한 불안정한 동영상을 흔들리지 않는 동영상으로 변환하는 기술이며, 고품질의 동영상을 획득하기 위한 필수적인 연구 분야이다. 본 논문에서는 가시광선 정보가 부족한 저조도 환경에서 중요하게 사용되고 있는 적외선 동영상이, 차량 혹은 소형 비행 물체 등에서 촬영되어 물리적인 작은 흔들림이 발생하였을 때 이를 실시간으로 안정화할 수 있는 두 가지 딥러닝 기반 모델을 제안한다. 제안하는 모델은 불안정한 동영상과 정답 안정한 동영상 학습 데이터 페어가 없는 환경에서 학습할 수 있는 자기 지도 학습 모델이다. 우리가 제안하는 안정화 모델 두 가지는, 적외선 동영상(해상도: 640x512)에 대하여 실시간 프레임 처리율(＞30 fps)과 함께 기존 딥러닝 기반 온라인 안정화 모델보다 더 좋은 안정화 정도를 보이는 것을 수치적 평가 및 결과 영상을 통해 확인하였다. 

Sungjoon Yoon

Papers

System-on-chip design for ultrasonic target detection using split-spectrum processing and neural networks

P1G-2 Dynamically Reconfigurable Neural Network Hardware Design for Ultrasonic Target Detection

P2D-3 Performance Evaluation of Neural Network Based Ultrasonic Flaw Detection

A theory-based approach to the usability of augmented reality technology: A cost-benefit perspective

Deep Learning-based Real-time Video Stabilization for Infrared Videos