Development of fuzzy logic based differentiation algorithm and fast line-scan imaging system for chicken inspection

A fuzzy logic-based algorithm was developed for a hyperspectral line-scan imaging system for differentiation of wholesome and systemically diseased fresh chickens. The hyperspectral imaging system consisted of an electron-multiplying charge-coupled-device (EMCCD) camera and an imaging spectrograph. The imaging system acquired line-scan images of chicken carcasses as they passed through the pixel-wide vertical linear field of view. The chickens were hung on a closed-loop laboratory processing line moving at a speed of 70 birds per minute. The use of light-emitting-diode (LED) line lights was selected following comparative evaluation of LED and quartz-tungsten-halogen (QTH) line lights. From analysis of wholesome and systemically diseased chicken spectra, four key wavelengths for differentiating between wholesome and systemically diseased chickens were selected: 413, 472, 515, and 546 nm; a reference wavelength at 626 nm was also selected. The ratio of relative reflectance between each key wavelength and the reference wavelength was calculated for use as input image features. A fuzzy logic-based algorithm utilising the image features was developed to identify individual pixels on the chicken surface exhibiting symptoms of systemic disease. Two differentiation methods utilising the fuzzy logic-based algorithm were tested using two separate image sets, the first containing 65 wholesome and 74 systemically diseased chickens, and the second containing 48 wholesome and 42 systemically diseased chickens. The first method achieved 100% accuracy in identifying chickens in both image sets. The second method achieved 96% and 100% accuracy in identifying chickens in the first and second image sets, respectively.

[1]  Kuanglin Chao,et al.  Systemically diseased chicken identification using multispectral images and region of interest analysis , 2004, SPIE Optics East.

[2]  L. Krummenacher,et al.  Noncontact surface roughness measurement of engineering surfaces by total integrated infrared scattering , 1997 .

[3]  Roy B. Dodd,et al.  COMPARISON OF DIFFERENT TYPES OF LIGHT SOURCES FOR OPTICAL COTTON MASS MEASUREMENTA NEURAL NETWORK FOR SETTING TARGET CORN YIELDS , 2001 .

[4]  Gerrit Polder,et al.  Hyperspectral image analysis for measuring ripeness of tomatoes. , 2000 .

[5]  Chun-Chieh Yang,et al.  Development of multispectral image processing algorithms for identification of wholesome, septicemic, and inflammatory process chickens ☆ , 2005 .

[6]  W. R. Windham,et al.  Hyperspectral Imaging for Detecting Fecal and Ingesta Contaminants on Poultry Carcasses , 2002 .

[7]  Kurt C. Lawrence,et al.  Hyperspectral Imaging for Detecting Fecal and Ingesta Contamination on Poultry Carcasses , 2001 .

[8]  Y R Chen,et al.  Analysis of visible reflectance spectra of stored, cooked and diseased chicken meats. , 2001, Meat science.

[9]  W. R. Windham,et al.  CALIBRATION OF A PUSHBROOM HYPERSPECTRAL IMAGING SYSTEM FOR AGRICULTURAL INSPECTION , 2003 .

[10]  D. R. Massie,et al.  Visible/Near-infrared Reflectance and Interactance Spectroscopy for Detection of Abnormal Poultry Carcasses , 1993 .

[11]  Anthony W. Parker,et al.  Time-resolved fluorescence imaging of photosensitiser distributions in mammalian cells using a picosecond laser line-scanning microscope , 2001 .

[12]  Ted Heske,et al.  Fuzzy Logic for Real World Design , 1996 .

[13]  Rasmus Nyholm Jørgensen,et al.  Modelling nitrogen and phosphorus content at early growth stages in spring barley using hyperspectral line scanning , 2004 .