Learn how to choose and evaluate the best signal and image processing methods for nondestructive testing (NDT) applications based on key aspects and criteria.

Implementing Advanced Non-Destructive Testing (NDT) Methods: Explain the role of NDT in ensuring project safety and compliance. Provide a detailed overview of different NDT techniques such as ultrasonic testing (UT), radiographic testing (RT), and dye penetrant testing (DPT). Share case studies from your work where NDT significantly impacted project outcomes.

How do you compare and benchmark different NDT signal and image processing methods?

Nondestructive testing (NDT) is a broad field that uses various methods to inspect and evaluate the properties, quality, and integrity of materials, components, and structures without causing damage or altering them. Signal and image processing are essential techniques for enhancing, analyzing, and interpreting the data obtained from NDT methods, such as ultrasonic, eddy current, magnetic, thermal, or optical testing. However, how do you compare and benchmark different signal and image processing methods for NDT applications? In this article, we will discuss some key aspects and criteria that you should consider when choosing and evaluating the best signal and image processing methods for your NDT problems.

1 Signal and image characteristics

The first step is to understand the characteristics of the signals and images that you are dealing with in your NDT application. For example, what are the sources and types of noise, artifacts, and interference that affect the quality and reliability of the data? What are the spatial and temporal resolutions, dynamic ranges, and contrasts of the signals and images? What are the features and patterns of interest that you want to extract, measure, or classify from the signals and images? These characteristics will help you determine the appropriate signal and image processing methods that can enhance, filter, segment, or transform the data for further analysis.

Add your perspective

Abdur Rasheed Khan

Senior Inspection Engineer QA/QC at IRCLASS Systems & Solutions Pvt. Ltd. ASNT Level II (RT, VT, UT, DPT)
Report contribution
Implementing Advanced Non-Destructive Testing (NDT) Methods: Explain the role of NDT in ensuring project safety and compliance. Provide a detailed overview of different NDT techniques such as ultrasonic testing (UT), radiographic testing (RT), and dye penetrant testing (DPT). Share case studies from your work where NDT significantly impacted project outcomes.

Like
Hamed Malekmohammadi

Project Engineer @ Applus+ RTD | Ph.D.
Report contribution
Should be added that in almost all of the NDT methods (except pure optical techniques such as VT/endoscope, etc.), the basis to generate an image is the initial point-by-point signal from ultrasound, electromagnetic or other types of sensors/probes. Therefore, it is essential to process the signal in the first place and then deal with images.

Like

2 Signal and image processing methods

The next step is to review the available signal and image processing methods that can address your NDT problem. There are many methods and algorithms that have been developed and applied for different NDT applications, such as wavelet analysis, Fourier analysis, principal component analysis, neural networks, machine learning, deep learning, computer vision, and artificial intelligence. Each method has its own advantages and limitations, depending on the nature and complexity of the data, the computational resources and time required, the accuracy and robustness of the results, and the ease of implementation and interpretation. You should compare and benchmark different methods based on these factors and select the most suitable ones for your NDT problem.

Add your perspective

Hamed Malekmohammadi

Project Engineer @ Applus+ RTD | Ph.D.
Report contribution
Generally, in most NDT methods, signals are in the form of time series. These time series have to be analysed either in time or in the frequency domain in the first stage. The signals are usually transformed from time- to frequency-domain using the Fourier Transform. Also, real and imaginary parts of the signal in both time and frequency domains as well as phase information can be quite useful for further analysis. Techniques such as Principal Component Analysis (PCA), or AI/ML are usually used to extract features, transform data to different spaces and classify, cluster or perform regression on these data to find additional features of interest.

Like

3 Signal and image processing tools

The third step is to choose the signal and image processing tools that can help you implement and evaluate the selected methods. There are various software and hardware tools that can facilitate the signal and image processing tasks for NDT applications, such as MATLAB, Python, LabVIEW, OpenCV, TensorFlow, PyTorch, GPUs, FPGAs, and cloud computing platforms. These tools can provide you with libraries, frameworks, modules, functions, and interfaces that can simplify the coding, debugging, testing, and optimization of the signal and image processing algorithms. You should compare and benchmark different tools based on their compatibility, functionality, performance, scalability, and usability for your NDT problem.

Add your perspective

4 Signal and image quality metrics

The fourth step is to define and measure the signal and image quality metrics that can quantify the effectiveness and efficiency of the signal and image processing methods and tools. There are various metrics that can be used to evaluate the signal and image quality, such as signal-to-noise ratio (SNR), peak signal-to-noise ratio (PSNR), mean squared error (MSE), structural similarity index (SSIM), contrast-to-noise ratio (CNR), edge detection rate (EDR), feature extraction rate (FER), classification accuracy (CA), precision, recall, F1-score, and confusion matrix. These metrics can help you compare and benchmark different methods and tools based on their impact on the signal and image quality and the NDT performance.

Add your perspective

Hamed Malekmohammadi

Project Engineer @ Applus+ RTD | Ph.D.
Report contribution
In my experience, NDT performance is mainly affected by the physical principles in the first place that are mainly influenced by the hardware and instruments, sensors, and objects under test and test conditions. Having these factors fixed, one can optimize the signal and image processing algorithms based on the mentioned merit factors for any specific NDT method.

Like

5 Signal and image validation methods

The fifth step is to validate the signal and image processing results using reliable and independent methods. Validation is an important step to ensure that the signal and image processing methods and tools are producing accurate, consistent, and meaningful results that can support the NDT decision making process. There are different ways to validate the signal and image processing results, such as using reference standards, ground truth data, synthetic data, simulation models, cross-validation techniques, statistical tests, expert opinions, and feedback from end-users. You should compare and benchmark different validation methods based on their relevance, feasibility, and credibility for your NDT problem.

Add your perspective

6 Signal and image optimization methods

The final step is to optimize the signal and image processing methods and tools to achieve the best possible results for your NDT problem. Optimization is an iterative process that involves fine-tuning the parameters, settings, and configurations of the signal and image processing methods and tools to improve their performance and efficiency. There are different methods and techniques that can help you optimize the signal and image processing tasks, such as grid search, random search, genetic algorithms, gradient descent, evolutionary computation, swarm intelligence, and reinforcement learning. You should compare and benchmark different optimization methods based on their suitability, complexity, and convergence for your NDT problem.

Add your perspective

7 Here’s what else to consider

This is a space to share examples, stories, or insights that don’t fit into any of the previous sections. What else would you like to add?

Add your perspective

Hamed Malekmohammadi

Project Engineer @ Applus+ RTD | Ph.D.
Report contribution
It has to be taken into account that the selection of proper signal and image processing algorithms and approaches is dependent on the NDT method, as well. There is no standard approach to be used for all NDT methods. Moreover, it is important to take into account the input signals, hardware and the whole inspection chain as a system, to better understand the output signals and implement the best possible approach to process and analyse them to generate user-friendly and high-quality data.

Like

How do you compare and benchmark different NDT signal and image processing methods?

1

2

3

4

5

6

7

1 Signal and image characteristics

2 Signal and image processing methods

3 Signal and image processing tools

4 Signal and image quality metrics

5 Signal and image validation methods

6 Signal and image optimization methods

7 Here’s what else to consider

Nondestructive Testing (NDT)

Rate this article

Thanks for your feedback

More articles on Nondestructive Testing (NDT)

More relevant reading