Multispectral Fluorescence Imaging Technique for On-Line Inspection of Fecal Residues on Poultry Carcasses

Abstract

Rapid and reliable inspection of food is essential to ensure food safety, particularly in mass production and processing environments. Many studies have focused on spectral imaging for poultry inspection; however, no research has explored the use of multispectral fluorescence imaging (MFI) for on-line poultry inspection. In this study, the feasibility of MFI for on-line detection of fecal matter from the ceca, colon, duodenum, and small intestine of poultry carcasses was investigated for the first time. A multispectral line-scan fluorescence imaging system was integrated with a commercial poultry conveying system, and the images of chicken carcasses with fecal contaminants were scanned at processing line speeds of one, three, and five birds per second. To develop an optimal detection and classification algorithm to distinguish upper and lower feces-contaminated parts from skin, the principal component analysis (PCA) and partial least square discriminant analysis (PLS-DA) were first performed using the spectral data of the selected regions, and then applied in spatial domain to visualize the feces-contaminated area based on binary images. Our results demonstrated that for the spectral data analysis, both the PCA and PLS-DA can distinguish the high and low feces-contaminated area from normal skin; however, the PCA analysis based on selected band ratio images (F630 nm/F600 nm) exhibited better visualization and discrimination of feces-contaminated area, compared with the PLS-DA-based developed chemical images. A color image analysis using histogram equalization, sharpening, median filter, and threshold value (1) demonstrated 78% accuracy. Thus, the MFI system can be developed utilizing the two band ratios for on-line implementation for the effective detection of fecal contamination on chicken carcasses.

Keywords: food safety; multispectral fluorescence imaging; online measurement; poultry inspection.

Figure 1

Body substances extracted from the chicken organs such as ceca, colon, small intestine, and duodenum.

Figure 2

Schematic of the multispectral fluorescence imaging system (a) and a real-time multispectral fluorescence imaging system for detecting fecal matters on chicken surface (b).

Figure 3

Schematic flowchart for multispectral fluorescence image acquisition and image classification. CASE1 shows data acquisition and spectral classification using multivariate analysis and CASE2 shows image classification based on the optimal principal component (PC). CASE3 is for color image classification.

Figure 4

Mean and standard deviation plot of three ROI groups from upper/bottom-ROI (fecal spots) and skin-ROI. Upper-ROI (red line, a) and bottom-ROI (blue line, b) represent ceca, colon spots, and small intestine, duodenum spots. Skin-ROI (green line, c) covers up whole pixels of skin image.

Figure 5

Score plot of three groups with PC2 and PC3 space in 2 dimensional. Red asterisks denote upper-ROI, blue ones are bottom-ROI, and green ones are skin-ROI.

Figure 6

Result of color image classification. A RGB color image (a) was split into 8-bit RGB component as red (b), green (c), and blue (d) image. Blue image was selected as a reference image to find the optimal threshold. Image processing methods were applied such as histogram equalization (e), sharpening (f), and median filter (g) (radius = 2.0). The result of auto threshold method and Shanbhag algorithm (threshold = 17) were the most acceptable result (h–j). The revised threshold value (threshold = 1) was employed and its pseudo colored and black/white result image (k,l).

Figure 7

Band ratios for discrimination of upper-ROI (UR), skin-ROI (SR), and bottom-ROI (BR) with the fluorescence intensity of 620 and 600 nm (a), 512 and 492 nm (b), and 630 and 600 nm (c). Density plot (d) of the kernel density of band ratio 630/600 (c). Black and white image of the band ratio 630/600 nm (e) and its projection onto the sample images (f).

Figure 8

The resultant binary image of partial least square discriminant analysis (PLS-DA) to the multispectral images (a) and its projection to the sample image with a pseudo color (b). Beta coefficient of PLS-DA of 2nd latent variable (c).

Figure 9

PC images of on-line scan multispectral fluorescence imaging (MFI) images for chicken with fecal spots. PC images (a–d) and four boxplots (e–h) collected pixels from red (1), green (2), and blue (3) rectangles.

Figure 10

Threshold algorithms with PC2 image using ImageJ. PC2 image (a) was applied to determine the threshold value (155) using the Huang method. The threshold employed to the PC2 image (b) and its binary image (c).

Figure 11

In order to find the optimal threshold value 30 samples (a) were used. PC2 employed (b) and selected the optimal threshold value (155) and its validation using the threshold value with artificial color (c).

Figure 12

Evaluation of principal component analysis (PCA) for fecal contamination detection on chicken carcasses moving at three different speeds of conveyor line: (a) 1 bird/s, (b) 3 birds/s, and (c) 5 birds/s. PCA detection and isolation accuracy for fecal spots on chicken carcasses is 97.5%.