Foreign object detection using noncontact ultrasound image . . .
of Vermont sharp and extra sharp cheddar cheese were scanned to disorders in foods. To discriminate the foreign objects from food
investigate their non-uniform characteristics since air bubble and materials, quantitative analysis needs to be performed by using the
cracks were frequently observed in the cheese samples examined. absolute values of ultrasound parameters of foods and foreign
The sporadic holes and a relatively long crack embedded in the objects. In addition, a better scan system and a higher frequency
cheese blocks were detected with NCU images as demonstrated in ultrasound transducer integrated with advanced signal processing
Figure 7e and 7f.
technique need to be developed to detect smaller foreign fragments
NCU images of relative attenuation and attenuation coefficient at higher scanning speeds.
of metal and glass fragments in poultry are shown in Figures 8a
The results of NCU images demonstrate its potential as a nonde-
through 8d. Since the second peak of through-transmission NCU structive, rapid, and economic tool to detect the presence of foreign
signal is clearly obtained in boneless poultry measurements, which objects and defects inside food materials. Further study is needed
is hardly observed in cheese measurements, the images of attenu- on NCU imaging of food materials of uneven or irregular shapes for
ation coefficient could be constructed. The images of attenuation calibration and use in online detection under actual production
coefficient and relative attenuation provide a clear indication of setting. To the authors’ knowledge, this is the first application of
metal and glass fragments in poultry. When the steel rod is placed noncontact air instability compensation ultrasound imaging of
in the poultry, the images of attenuation coefficient show the po- foreign object detection in solid foods.
sition and dimension of the incursion much more clearly than those
References
of relative attenuation. This is to be expected because of the large
difference in attenuated energy between the second and the first
peak. The second peak is from a signal that passes through the for-
eign incursion 3 times due to an internal reflection while the first
peak is from the signal that passes through it only once. The large
scatterings occur when the second peak signal is propagated
through the incursion repeatedly whereby distinguishing the posi-
tion and dimension of the foreign incursion clearly.
Bhardwaj MC. 2000. High transduction piezoelectric transducers and introduction
of non-contact analysis. The e-Journal of Nondestructive Testing & Ultrasonics
[serial online]. 1:1-21. Available from NDT net <http://www.ndt.net>. Posted Jan-
uary 2000.
Cho B, Irudayaraj J. 2002. Design and application of a non-contact air instability
compensation ultrasound transducer using spatial impulse response. Trans ASAE
(
Forthcoming).
Cho B, Irudayaraj J, Bhardwaj MC. 2001. Rapid measurement of physical properties
of cheddar cheese using a non-contact ultrasound technique. Trans ASAE
44(6):1759-62.
Gan TH., Hutchins DA, Billson DR. 2002. Preliminary studies of a novel air-coupled
ultrasonic inspection system for food containers. J Food Eng 53:315-23.
Experiments on the step size of scanning on image quality indi-
cated that when the step size was reduced to 0.5 mm instead of 1 Haney MJ, O’Brien WD Jr. 1986. Temperature dependency of ultrasonic propaga-
tion properties in biological materials. In: Greenleaf JF, Editors. Tissue character-
ization with ultrasound. Boca Raton, Fla.: CRC Press. p 15-55.
mm, more identical images were observed as shown in Figure 8e and
8
f. The scanning time takes 90 s with a 1-mm step size for a 20 ϫ 20 MacAndrew A, Harris C. 1991. Sensors detect food contamination. Sensor Rev 11:23-
2
5.
mm image. However, the period is increased to 6 min, which is 4
times when a 0.5-mm step size was used. Hence, the scanning region
and step size need to be optimized for each application.
McFarlane N, Bull C, Tillett R, Speller R, Royle G. 2001. Time constraints on glass detec-
tion in food materials using Compton scattered X-rays. J Agric Eng Res 79:407-18.
Nielsen M. 1996. Texture measurement of fruits and vegetables by the use of ultra-
sonics. [DPhil thesis]. Telefon, Denmark: The Royal Veterinary and Agricultural
Univ. 137 p. Available from: Univ. Library, Telefon, Denmark.
Conclusions
Povey, MJW, McClements DJ. 1988. Ultrasonics in food engineering. Part 1: Intro-
duction and experimental methods. J Food Eng 8:217-45.
Tao Y, Chen Z, Jing H, Walk J. 2001. Internal inspection of deboned poultry using X-
ray imaging and adaptive thresholding. Trans ASAE 44:1005-9.
Tao Y, Ibarra J. 2000. Thickness-compensated X-ray imaging detection of bone frag-
ments in deboned poultry – Model analysis. Trans ASAE 43:453-9.
HE NONCONTACT AIR INSTABILITY COMPENSATION ULTRASOUND
T
transducers improved the velocity and thickness measurements
of the calibration standard in a varying temperature (26.0 ЊC to
2
8.5 ЊC) environment. Even though the advantages of the air insta- Zeqiri B. 1989. Reference liquid for ultrasonic attenuation. Ultrasonics. 27:314-5.
MS 20020501 Submitted 8/30/02, Revised 12/15/02, Accepted 1/9/03, Received
bility compensation transducer did not affect the images of velocity
due to low-quality images, accurate images of relative attenuation
and attenuation coefficient could be obtained with the stable thick-
1/10/03
The authors express their sincere appreciation to Dr. Mahesh Bhardwaj and Gary Stead of
SecondWave Systems for providing generous access to the image scanning equipment and
technical support.
2
ness calculation. The minimum 3 ϫ 3 mm foreign fragments in
cheese and poultry and 1.5-mm-dia cylindrical objects in the food
Authors Cho and Irudayaraj are with the Dept. of Agriculture and Biologi-
materials could be clearly detected, even though the images were cal Engineering, 227, Agricultural Engineering Building, The Pennsylva-
nia State Univ., University Park, PA 16802. Direct inquiries to author
Irudayaraj (E-mail: josephi@psu.edu).
distorted by diffraction and refraction effects. However, it was dif-
ficult to recognize the differences between fragments and internal
9
74 JOURNAL OF FOOD SCIENCE—Vol. 68, Nr. 3, 2003