Title: Improved modified pressure imaging and software for egg micro-crack detection and egg quality grading Authors
Submitted to: Applied Engineering in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: December 18, 2011
Publication Date: April 25, 2012
Citation: Yoon, S.C., Lawrence, K.C., Jones, D.R., Heitschmidt, G.W. 2012. Improved modified pressure imaging and software for egg micro-crack detection and egg quality grading. Applied Engineering in Agriculture. 28(2):283-293. Interpretive Summary: Cracks in the egg shell are the first line of defense against the external bacterial contamination. Thus, cracked eggs increase the food safety risk to consumers. Especially, eggs with very fine hairline cracks (micro-cracks) often escape the detection of the mass-scanning equipment and professional graders alike. The problem with these micro-cracks is that they become enlarged in cool storage after grading them. Therefore, improved accuracy in crack detection is important in ensuring the food safety and quality for consumers. Regardless of egg grading methods being adopted in individual processing plants, trained USDA graders still use hand candling of using light in a darkened booth for assuring accuracy in grading of mass-scan equipment and personnel. In order to aid trained USDA graders in crack detection, researchers at the Agricultural Research Service (ARS) of the USDA developed an imaging system using the modified negative pressure to detect checked (cracked but membranes are intact) and cracked eggs with high accuracy. In a recent project to improve the egg capacity from 15 eggs to 20 eggs, the system was also upgraded with a stepper motor to image the entire surface of an egg automatically, a hinged lid to help faster loading/unloading of eggs, and data management software to manually enter other grading values and to automatically compile and save the grading results as an Microsoft Excel spreadsheet conformed to the PY-75 form used by USDA graders. The improved system was tested in a commercial egg processing facility in a 3-day study in which two professional USDA graders helped grade 3,000 eggs (thirty 100-egg lots) by using the hand candling technique and the imaging system, independently. The study found about 11% difference in crack detection accuracy between the system (95%) and the human grader (84%). This study has implications for potential use of the imaging system to increase the grading accuracy of USDA graders. Also, the developed data management software is applicable to a database program for gathering and analyzing local and national grading trends on a daily basis.
Technical Abstract: Cracks in the egg shell increase a food safety risk. Especially, eggs with very fine, hairline cracks (micro-cracks) are often undetected during the grading process because they are almost impossible to detect visually. A modified pressure imaging system was developed to detect eggs with micro-cracks and regular cracks alike without adversely affecting egg quality. The imaging system forces a change in pressure from the atmospheric pressure to the negative pressure (vacuum) in an enclosure to open the cracked shell surface momentarily, while intensity changes in the open shell surface are recorded by a high-resolution digital camera and detected by an image processing algorithm. In this paper, modifications and improvements to the system are reported. The previous design for 15 eggs was modified to accommodate 20 eggs and improved to add a stepper motor to image the entire surface of eggs automatically and the data management software to aid a professional grader in egg grading. In a three-day study with 3,000 eggs in a commercial egg processing plant, a professional grader operated the 20-egg imaging system after a short training on the first day and another professional grader graded the eggs with hand candling. The accuracy in crack detection of both graders and the system itself was compared in order to test the crack-detection performance of the system and to explore the possibility of the use of the system for egg grading by USDA graders. The system automatically detected checked and cracked eggs with 94.53% accuracy with 0.29% of a false positive rate while a hand candler had 83.58% accuracy in crack detection. The grader operating the system had 98.51% accuracy in crack detection without any false positives. Other quality factors were also graded by both graders. The study demonstrated the potential of the system for egg grading in high accuracy.