Contaminant Classification of Poultry Hyperspectral Imagery using a Spectral Angle Mapper Algorithm

Abstract

Since hyperspectral imaging technique has been demonstrated to be a potential tool for poultry safety inspection, particularly faecal contamination, a hyperspectral image classification method was developed for identifying the type and source of faecal contaminants. Spectral angle mapper (SAM) supervised classification method for hyperspectral poultry imagery was performed for classifying faecal and ingesta contaminants on the surface of broiler carcasses. Spatially averaged spectra of three different faeces from the duodenum, caecum, colons, and ingesta of maize/soya bean diet were used for classification data. The SAM classifier using reflectance of hyperspectral data with 512 narrow bands from 400 to 900 nm was able to classify three different faeces and ingesta on the surface of poultry carcasses. Based on the comparison with ground truth region of interest, both classification accuracy and kappa coefficient, which quantifies the agreement of classification, increased when spectral angle increased. The overall mean accuracy and corresponding mean kappa coefficient to classify faecal and ingesta contaminants were 90·13% (standard deviation of 5·40%) and 0·8841 (standard deviation of 0·0629) when a spectral angle of 0·3 radians was used as a threshold.