Laser & Optoelectronics Progress, Volume. 48, Issue 7, 73001(2011)
Detection of Fecal Contaminants on Chicken Carcasses Using Segmented Principal Component Analysis and Band Ratio Algorithm
Using chicken carcasses as the research subject, fecal contaminants on chicken carcasses are detected by using hyperspectral imaging technology and combining segmented principal component analysis and band ratio algorithm. Firstly, hyperspectral images of chicken carcasses from 400 to 1000 nm are collected. Secondly, seven characteristic wavelengths (520.64,542.12,561.61,577.04,703.82,595.6 and 852.1 nm) are obtained by segmented principal component analysis, and the images obtained using 577.04/520.64 nm band ratio image added by 852.1/703.82 nm band ratio image are selected as the characteristic images of the detection of fecal contaminants on chicken carcasses. Lastly, the fecal contaminants on chicken carcasses are extracted using the threshold segmentation and mathematical morphology. The experimental results show that the accuracy rates of the detection for the fecal contaminants of ceca, colon and duodenum are 100%, 100% and 96% respectively, and the total accuracy rate of the detection is 93.3% using 60 samples of chicken carcasses.
Get Citation
Copy Citation Text
Zhao Jinhui, Yu Fang, Wu Ruimei, Liu Muhua, Yao Mingyin. Detection of Fecal Contaminants on Chicken Carcasses Using Segmented Principal Component Analysis and Band Ratio Algorithm[J]. Laser & Optoelectronics Progress, 2011, 48(7): 73001
Category: Spectroscopy
Received: Mar. 28, 2011
Accepted: --
Published Online: Jun. 2, 2011
The Author Email: Jinhui Zhao (zjhxiaocao@sina.com)