High Contrast Ultrasound Images of Defects in Food Package Seals
Previous work to detect defects in food packaging seals using pulse-echo ultrasound inspired
the backscattered amplitude integral (BAI) imaging techniques, which could reliably identify channels
with diameters 38 micron or larger at a center frequency of 17.3 MHz (lamda = 86 micron). The current
study presents two new processing techniques that more reliably reveal smaller channels (about 6 micron in
diameter). The RF sampling technique (RFS) displays a single, time-gated, pressure value from the
received (not envelope-detected) RF waveform at each transducer position. The RF correlation technique
(RFC) calculates the correlation coefficients of the RF signals with a reference signal that does not
pass through a channel. The correlation coefficient can be calculated for the entire RF signal (RFCE)
or over a short segment of the RF signal (RFCS). The performance of these imaging methods for detecting
channel defects is investigated for plastic and aluminum foil trilaminate films with 6, 10, 15, 38, and
50 micron channels filled with water or air. Data are collected with a focused ultrasound transducer (17.3
MHz, 6.35 mm in diameter, f/2, 173 micron -6 dB pulse-echo lateral beamwidth the focus) scanned over a
rectangular grid, keeping the package in the focus. Performance is measured using detection rates, image
contrast, and contrast-to-noise ratio (CNR). Both RFS and RFCS have improved detection rates relative to
BAI for channels 15 micron or smaller. The RFCS technique is the most effective at smoothing the
background, leading to the greatest CNR improvements.
|