Skip to main content
ARS Home » Research » Publications at this Location » Publication #194754
Title: LINE-SCAN MACHINE VISION SYSTEM FOR ONLINE POULTRY CARCASS INSPECTION

Author
item YANG, CHUN-CHIEH - VIS SY UNIV OF KENTUCKY
item Chao, Kuanglin - Kevin Chao
item Chen, Yud

Submitted to: BARC Poster Day
Publication Type: Abstract Only
Publication Acceptance Date: 4/6/2006
Publication Date: 4/26/2006
Citation: Yang, C.C., Chao, K., Chen, Y.R. 2006. Line-scan machine vision system for online poultry carcass inspection. BARC Poster Day 2006.

Interpretive Summary:

Technical Abstract: It is essential to ensure food safety for poultry products, including a zero tolerance standard for chickens with infectious conditions such as septicemia and toxemia, which must be removed from the processing line. A hyperspectral line-scan imaging system for online automated inspection of wholesome and diseased chickens was developed and tested in a poultry processing plant in December 2005. The hyperspectral imaging system consisted of an electron-multiplying charge-coupled-device camera and an imaging spectrograph. Using the spectrograph, the system collected spectral measurements across a pixel-wide vertical linear field of view which chicken carcasses are passed through. The camera took line-scan images with an exposure time of 0.1 msec. Unlike most imaging systems, this hyperspectral system can also function as a line-scan multispectral imaging system using the same features and detection algorithm. This characteristic allows methods used to collect and process huge amounts of hyperspectral data to be easily transferred for implementation of high-speed multispectral acquisition and analysis without the need for cross-system calibration. Another original feature of this system is the use of light-emitting-diode line lights that can provide higher intensity illumination than conventional quartz-tungsten-halogen lights do at short wavelengths, from which more essential information for systemical disease identification can be extracted. From spectral analysis, several key wavebands were selected. An algorithm was developed to determine whether or not the region of interest for a carcass has entered or passed beyond the field of view, in order to trigger the beginning and end of the image differentiation process. A fuzzy logic-based algorithm utilizing the key wavebands, as another innovation of this system, was developed to identify individual pixels on the chicken surface exhibiting symptoms of systemic disease. Based on the mean and standard deviation values of sample chickens, the fuzzy membership functions for the key wavelength were built for wholesome and systemically diseased chickens, respectively. By the defuzzification functions, a decision output was made, which ranged from one for existence of systemic disease to zero for the evidence of being wholesome. The successful result showed that the system could quickly acquire clear line-scan images from the processing line and accurately differentiate systemically diseased chickens from wholesome ones. The system can be applied to ensure food safety and for preventing food safety hazards in the inspection process for poultry products. With this automated machine vision system, the poultry plants can meet government food safety regulations while maintaining their competitiveness to satisfy consumer demand.