Acoustography is a full-field ultrasonic imaging process where a high-resolution two-dimensional acousto-optic (AO) sensor based on liquid-crystal technology is employed to directly convert the ultrasound into a visual image in near real time. Unprocessed acoustography images typically suffer from nonuniformity and a resulting lack of defect sharpness because of spatial variations in the optical brightness response of the acousto-optic sensor field to ultrasonic intensity. These limitations can make acoustography image interpretation difficult, large field applications impractical, and use on samples with a wide range of attenuation difficult. In 2005, a new software methodology was developed and demonstrated to address these limitations.
Prior to developing methodologies, a custom acoustography system was developed jointly by the NASA Glenn Research Center and Santec Systems, Inc. The conventional acoustography system was modified from its original implementation to include a National Instruments Signal Generator card to generate the swept sinusoid function, a National Instruments multifunction (NI PCI-6014) card having analog outputs to generate sine waves for AO sensor erasure, and an 8-bit monochrome digital camera capable of 60 frames/sec video capture. The original frame grabber and charge-coupled device (CCD) analog camera were no longer needed and were removed. A radiofrequency amplifier amplifies the swept sinusoid wave from the signal generator card by approximately 30 dB before it reaches the transducer. An entirely new acoustography operating system was developed using LabVIEW 7.1 both to conduct these experiments and to operate the system for general inspection. This opera-ting system allows control of swept sinusoid voltage amplitude and sweep parameters, AO sensor erasure, camera frame captures per second, and camera gain. Data analysis, image analysis, and image processing software code was developed in LabVIEW 7.1 and the associated IMAQ Vision toolkit to implement the methodologies discussed in this study.
Nonuniformity and flaw-detection enhancement approach in brief. Black arrows demarcate the two threshold optical brightness levels. The image on the right shows results for a plexiglass sample with a seeded defect. Twp, time with part present; Two, time without part present.
After system development, a custom algorithm was developed and tested at Glenn that shows promise for nonuniformity correction and flaw detection enhancement in acoustography images. It is based on the principles shown in the preceding figure. Note that a series of images develops as a function of time in the acoustography process. The algorithm developed is based on a time ratio in which, for each pixel level, times are extracted for a specific brightness level with and without the part present, divided together, and converted to decibels so that each pixel can be assigned a “loss” value. Essentially, normalization is occurring with respect to the underlying AO sensor portion at each pixel. This method is novel in that most ultrasonic (and nondestructive evaluation results in general) use voltage or intensity ratios for decibel assignment (which also was considered in this study). Overall, the custom algorithm is speedier than working with voltage ratios. The top right of the preceding figure shows the results of experiments performed using a plexiglass sample with seeded defect. The loss image shows the seeded defect extremely clearly and appears to have been successful at removing some of the noise pattern at the bottom center of the original brightness-versus-time images.
Next, the same approach was applied for a ceramic matrix composite sample with seeded defects and the final result is shown as the average loss (decibels) in the bottom left image of the following figure. In comparison to the center image (unprocessed acoustography image), the defects are somewhat enhanced.
Acoustography Images for the ceramic matrix composite sample with flat bottom holes. Left: Conventional acoustography image without the sample. Center: Conventional acoustography image with the sample. Right: Average loss (decibels).
Long description of figure 2.
Roth, Don J.; et al.: Approaches for Non-Uniformity Correction and Dynamic Range Extension for Acoustography. SPIE Int. Soc. Opt. Eng., vol. 5770, 2005, pp. 124-134.Glenn contact: Dr. Don J. Roth, 216-433-6017, Donald.J.Roth@nasa.gov
Last updated: October 16, 2006
For additional information, please contact Cynthia L. Dreibelbis at 216-433-2912 or firstname.lastname@example.org.
Responsible NASA Official: Kim Dalgleish-Miller, email@example.com