Submitted to: Sensing and Instrumentation for Food Quality and Safety
Publication Type: Peer reviewed journal
Publication Acceptance Date: 2/1/2008
Publication Date: 3/14/2008
Citation: Lawrence, K.C., Yoon, S.C., Heitschmidt, G.W., Jones, D.R., Park, B. 2008. Imaging system with modified-pressure chamber for crack detection in shell-eggs. Sensing and Instrumentation for Food Quality and Safety.2:116-122 Interpretive Summary: In the U.S., most shell eggs are processed with high-speed processing equipment and many companies sell their eggs as a shielded product. A shielded egg product is one that has been inspected by a USDA grader and determined to have certain quality attributes. In common terms, the processors want to sell USDA graded eggs because the consumers trust the quality of graded eggs and buy them. Thus, to make sure the processing equipment is operating within specifications of a particular grade, representative egg samples are pulled from the processing line and humans inspect these sample eggs for various defects. One of the biggest challenges for a grader is detecting very small hair-line cracks or checks, known as micro-cracks. When the grader inspects the eggs, these cracks are generally very small and hard to see (and hear). However, as the egg cools over time, these cracks tend to grow and become more noticeable. This results in egg lots that can exceed the number of allowable checks and/or cracks for their assigned grade. To aid the grader, a system was developed to detect micro-cracks in shell eggs. The system consists of a computer controlled digital camera positioned above a clear chamber which holds the eggs and is illuminated from underneath. An egg is placed in the chamber and an image of the egg at atmospheric pressure was captured. Next a short, quick, and relatively low negative pressure was applied to the chamber and a second image was captured. Taking the negative pressure image and dividing it by the atmospheric pressure image results in a ratio image that can easily detect cracks in shell eggs because the system detects the difference in the light emitting from the crack. A few additional common image-processing steps were also used to remove background noise and reduce false positives (intact egg-shell features incorrectly predicted as a crack). For single eggs, the system was able to detect over 98% of the cracked eggs and 100% of the intact eggs. The system shows great promise and now needs to be scaled up to image multiple eggs at a time.
Technical Abstract: To detect checks and/or cracks in shell eggs, the egg industry is using high-speed acoustic systems. Prior to shipment, human graders candle a small subset of eggs to ensure that the high-speed systems are operating within specifications for a given grade of egg (e.g., Grade A Large Eggs). In addition to visual inspection, graders also listen for a dull, flat sound as an indicator of a crack when tapping eggs together. However, very small cracks, or micro-cracks can go undetected by the human graders. A method to detect egg checks/cracks with an imaging camera was developed. The system consisted of an imaging camera positioned above a clear inspection chamber that housed an egg and was illuminated from underneath. The chamber was designed so that a short, quick vacuum could be pulled to enhance the crack detection. High resolution monochromatic images were collected at atmospheric pressure and under negative pressure. The negative pressure gradient was used to briefly open any existing cracks without inducing any new cracks in an intact egg. Initially, eggs were manually rotated three times to image the whole surface of the egg. The ratio of a negative-pressure image divided by an atmospheric-pressure image was used to highlight the cracks and simple image processing was used to identify a crack. Eighty cracked and 80 intact eggs were imaged with the system. Only one cracked egg was not detected, and this was because the crack was located on the air-cell end of the egg and was not visible by the camera. Thus, the system was 98.75% accurate in identifying cracked eggs and 100% accurate in identifying intact eggs. This initial phase of the research was successful and the system is being scaled up to image multiple eggs at a time.