The Study On Fast Detection Of Fecal Contaminants And Disease Of Lamb Meat Using Hyperspectral Imaging Technique

the aim of this study is to use Vis-Nir and NIR hyperspectral and spectral image technology to detect fecal contaminants and disease of lamb meat from Ningxia. Combined chemometrics methods with image processing method,the fecal contaminants and disease of lamb meat was established,the detection algorithms offered theoretical basis for developing a real-time.fast and on-line nondestructive detection system.The results are as below:(1)Hyperspectral imaging technique operated in the visible and near-infrared(Vis-NIR) region(400-1000nm) was developed for non-contact measurement of fecal contaminants of lamb meat.Principal Component Analysis(PCA) was performed on hyperspectral imaging for suitable PC images.Fecal contaminants on lamb’s surface were segemented by median filtering, square root transformation, binarization and morphological imaging processing.and120samples(include40normal samples,40fecal of intestinum crassum,40fecal of intestinum tenue) were identificated one by one. Experimental results showed that the fecal on the lambs can be detected with an accuracy of98.5%in the region of Vis-NIR.(2)five wavelengths were selected as the optimum wavelengthes for fecal of intestinum crassum and fecal of intestinum tenue using PCA for developing a on-line nondestructive detection system, Experimental results showed that the fecal on the lambs can be detected with an accuracy of98.5%in the optimum wavelengthes.(3)The pattern recognition of two disease lamb meats was developed using Hyperspectral imaging technique in the near-infrared(NIR) region(900-1700nm). multiple scattering correction and SG was adopted to normal lamb meats and two disease lamb meats for spectral preprocessing. unsupervised pattern of clustering analysis was performed on the samples and predicion model were built by LDA, identification rate of validity was99%.(4)the optimal wavelength was obtained by PCA and Second derivative. the resulting wavelengths from the2nd derivative for classification of meat samples with LDA yielding98%.and the key wavelengths from PCA for classification of meat samples with LDA yielding96%.