Rashmi Murthy

Spectral histogram using the minimization algorithm-theory and applications to flaw detection

In ultrasonic flaw detection in large grained materials, backscattered grain noise often masks the flaw signal. To enhance the flaw visibility, a frequency diverse statistical filtering technique known as split-spectrum processing has been developed. This technique splits the received wideband signal into an ensemble of narrowband signals exhibiting different signal-to-noise ratios (SNR). Using a minimization algorithm, SNR enhancement can be obtained at the output. The nonlinear properties of the frequency diverse statistic filter are characterized based on the spectral histogram, which is the statistical distribution of the spectral windows selected by the minimization algorithm. The theoretical analysis indicates that the spectral histogram is similar in nature to the Wiener filter transfer function. Therefore, the optimal filter frequency region can be determined adaptively based on the spectral histogram without prior knowledge of the signal and noise spectra.<<ETX>>

Analysis of a non-linear frequency diverse clutter suppression algorithm

Abstract The target signal is often received in the presence of coherent noise resulting from a large number of complex and randomly distributed scatterers known as clutter, whose echoes can suppress or mask the actual target echo. This is a fundamental limitation which exists in all imaging and detection applications and cannot be removed by conventional techniques, such as time averaging. This paper examines a non-linear signal processing algorithm (polarity thresholding) used in conjunction with a frequency diversity technique (split-spectrum processing), which decorrelates clutter. Since the target echo exhibits significantly smaller amplitude variations with frequency compared to clutter, the polarity thresholding algorithm can achieve flaw enhancement by setting the output to zero at time instants where phase reversal occurs. Theoretical derivations are provided to determine the signal to noise ratio enhancement capabilities of the algorithm. Ultrasonic data obtained from large grained stainless-steel samples with flat-bottom holes are used to verify the theoretical results and demonstrate the grain echo suppression capability of the algorithm in imaging applications.

Flaw detection in stainless steel samples using wavelet decomposition

Wavelet techniques have emerged as useful tools in signal analysis because of their time-frequency localization properties. In this work, wavelet decomposition and reconstruction algorithms are used in ultrasonic nondestructive testing applications to distinguish between the flaw echo and background grain noise. The discrete wavelet transform is applied to reconstruct the signal at scales likely to contain the target. Nonlinear algorithms are used to obtain the output signal from the reconstructed signals. Preliminary results indicate that these methods are quite successful in the detection of single targets but not as effective as split spectrum processing in the resolution of closely spaced multiple targets

Spatial processing for coherent noise reduction in ultrasonic imaging

When the target is received in the presence of coherent noise resulting from a large number of complex and randomly distributed scatterers inherent to the medium, the ability of the system to distinguish between the two types of signals often becomes the most crucial performance consideration. Conventional techniques that are capable of suppressing time‐varying (incoherent) noise are generally not effective when the noise is time invariant (coherent). In recent years, diversity techniques have been developed that allow the decorrelation of the coherent noise term by altering either the transmitted frequency or the position of the transducer. Although the diversity techniques provide some noise suppression when used in conjunction with the conventional averaging algorithms, their potential benefits cannot be fully exploited by such linear techniques alone. In the work presented here, a nonlinear detection scheme based on the polarity of the spatially decorrelated signals is examined. The theoretical and experimental results indicate that the polarity‐thresholding algorithm provides target enhancement that is far superior to the linear techniques previously used in spatial processing. Furthermore, the paper examines the spatial decorrelation properties of the experimental data to determine the desirable parameters for data acquisition and signal processing.

APPLICATION OF BANDPASS FILTERING IN ULTRASONIC NON-DESTRUCTIVE TESTING

Ultrasonic nondestructive testing of large grained materials is limited by the ability of the detection process to distinguish the flaw signals from the backscattered grain boundary echoes. This coherent grain noise often masks the echo from inhomogeneities and defects in the material. Absorption and scattering effects further reduce the ultrasound energy leading to poor signal-to-noise ratio in the received signal. It is not possible to reduce the grain clutter by conventional time averaging techniques due to its coherent nature. Different algorithms utilizing the principles of frequency diversity and spatial diversity have been used in the past for signal-to-noise ratio enhancement. In NDE applications where the noise is primarily due to Rayleigh scattering, it can be shown that flaw detection can be improved significantly by merely bandpass filtering the lower part of the received wideband echo spectrum. Both theoretical and experimental results are presented to support this conclusion. The filtering technique is successfully tested on materials with different grain sizes. The main advantage of this method is its relative simplicity, which eliminates the need for sophisticated and computationally intensive signal processing algorithms. Furthermore, this technique allows simple hardware implementation for real-time applications. The optimal parameters, i.e., the center frequency and bandwidth of the bandpass filter are experimentally determined.

Adaptive and robust filtering techniques for ultrasonic flaw detection

An overview is presented of a number of adaptive flaw enhancement techniques based on the statistical properties of the order-statistic operation (e.g. median or minimization) or the group delay of the received ultrasonic signal. Nonlinear statistical-filtering techniques using frequency-diverse observations that utilize these adaptive properties to improve flaw detection are examined. These concepts are also applied to develop an adaptive-correlation receiver. Furthermore, it is shown that the geometric mean of the decorrelated observations yields outputs that are relatively insensitive to critical-processing parameters. Experimental results using stainless steel samples are presented to demonstrate the robust and adaptive nature of the filters.<<ETX>>

Spectral histogram and its application to flaw detection

The split-spectrum processing (SSP) minimization algorithm has been successfully used to enhance ultrasonic flaw detection. However, this technique is fairly sensitive to the processing parameters, especially the frequency region selected for processing. In the past, the optimal spectral region could be obtained only by trial and error. In the present work, the concept of the spectral histogram has been introduced to achieve the optimal region automatically. In addition, it has been shown that the spectral histogram for the SSP minimization algorithm can be used to approximate the frequency response of this nonlinear system in the statistical sense, which can be utilized for obtaining a linear filter analogous to the Wiener filter. Experimental data obtained by ultrasonic nondestructive testing show that the proposed technique yields performance comparable to that obtained by a trained operator using a trial and error approach, indicating its feasibility in the automation of the SSP minimization algorithm.<<ETX>>

Scaling techniques for medical image enhancement

In previous work, multiresolution representations that provide a hierarchical framework for analyzing the information content of images were used to obtain contrast enhancement in medical sonograms. Image decomposition and reconstruction was implemented using the two dimensional wavelet transform with a computationally efficient quadrature mirror filter bank architecture. The images were reconstructed at scales likely to exhibit high target energy localization indicating the presence of a tumor. The concepts were applied to B-scan images from in vivo liver images. It was observed that the reconstructed images at scale 1 provided the most enhancement but did not achieve sufficient image contrast. In order to utilize the information present in the adjacent scales, the reconstructed signals are nonlinearly combined using algorithms which exploit the observation that the abnormalities are less echogenic and less sensitive to frequency shifts than the surrounding healthy tissue. It was observed that the frequency diversity techniques outperform the reconstructed images.

Detection of ultrasonic flaw signals using wavelet transform techniques

Ultrasonic detection and identification of flaws embedded in large-grained materials is often limited by the presence of high amplitude interfering echoes due to unresolvable grain boundaries. The split spectrum processing (SSP) technique using nonlinear algorithms is very effective in grain noise suppression and flaw detection. The wavelet transform technique is used to perform spectral decomposition, followed by the application of various nonlinear algorithms to obtain the output signal. The wavelet transform is based on the principle of constant Q or constant relative bandwidth frequency. Experimental results for the constant-Q SSP technique are presented. The experimental data indicate improved performance in identifying and extracting multiple targets compared to the conventional fixed bandwidth SSP.<<ETX>>

Temporal and spatial spectral features of B-scan images

B-scan images of metals are frequently used in ultrasonic nondestructive evaluation for flaw detection purposes. Often, the backscattered grain noise masks the flaw echo, making it difficult to visually distinguish the flaw signal from the grain echoes. Previously, spatial and frequency compounding, as well as decorrelation techniques have been used to enhance the flaw echo. The partial spectra, i.e., the temporal or spatial frequency components of the B-scan image, are used individually to obtain information regarding the flaw location. It is observed that the partial spectra behave in a distinct manner when the B-scan contains a flaw, and provide more insight than the conventional two-dimensional spectrum. Using this information, a two-dimensional filter that significantly suppresses grain noise can be designed. Experimental results using data from stainless steel samples are presented.<<ETX>>