Page Banner

United States Department of Agriculture

Agricultural Research Service

Title: Hyperspectral Image Classification for Fecal and Ingesta Identification by Spectral Angle Mapper

Authors
item Park, Bosoon
item Windham, William
item Lawrence, Kurt
item Smith, Douglas

Submitted to: ASAE Annual International Meeting
Publication Type: Proceedings
Publication Acceptance Date: June 8, 2004
Publication Date: August 1, 2004
Citation: Park, B., Windham, W.R., Lawrence, K.C., Smith, D.P. 2004. Hyperspectral image classification for fecal and ingesta identification by spectral angle mapper. Transaction of American Society of Agricultural Engineers. American Society of Agricultural Engineers, St. Joseph, MI. Annual International Meeting. Technical Paper No.043032.

Interpretive Summary: Food safety has become an important concern for public health because reductions in health risks to consumers are important food safety issues. While a number of factors can influence bacterial contamination of chicken carcasses, one of the leading causes is fecal contamination at the processing plant. Improper plant handling and equipment that is not good working condition can also contribute to the problem. Hyperspectral imaging techniques may be used in poultry safety inspection to detect fecal contamination. The hyperspectral image classification method was developed to identify the types and sources of various carcass contaminants. This new inspection method can improve poultry safety inspection programs by incorporating scientific testing and thereby reducing the amount of processed chicken carcasses that are contaminated by feces.

Technical Abstract: This paper describes the performance of spectral angle mapper (SAM) supervised classification method for hyperspectral poultry images to classify fecal and ingesta contaminants on the surface of broiler carcasses. Spatially averaged spectra of three different types of feces from the duodenum, ceca, colon, and ingesta of corn/soybean diet were used for classification data. SAM classifier using reflectance of hyperspectral data with 512 narrow bands from 400 to 900 nm was able to classify three different types of feces and ingesta on the surface of poultry carcasses. Based on the comparison with ground truth region of interest, both classification accuracy and kappa coefficient increased when spectral angle increased. The overall mean accuracy and corresponding mean kappa coefficient used to classify fecal and ingesta contaminants were 90.13% (standard deviation = 5.40%) and 0.8841 (standard deviation = 0.0629) when a spectral angle of 0.3 radians was used as a threshold.

Last Modified: 10/1/2014