摘要： The chemical composition of cheese is significantly related to quality, as it is responsible for its shelf life, yield and texture. However, conventional methods for determination of chemical composition are laborious and time-consuming. Fast assessment tools such as NIR spectroscopy and spectral imaging could be used for the quality control of cheeses. The advantage of this technique is that it allows several constituents to be measured simultaneously in a quick and nondestructive way. The objective of this work was to investigate spectral imaging combined with principal component analysis (PCA) for assessment of cheese samples. Spectral information of six varieties of cheese (Cheddar, coalho, Minas, mozzarella, prato and block processed cheese) were obtained using a spectral imaging system, between 928 and 2524 nm, with 6 nm intervals, resulting in 256 analyzed wavelengths. The first two principal components were responsible for 98.2% of the variation among samples, and the score plot presented good separation among samples of coalho cheese, mozzarella and Minas cheese. Loadings show that some peaks are strongly influenced by wavebands, associated to chemical bonds related to protein and fat. Spectral imaging combined with multivariate analysis can be a potential tool for fast cheese quality assessment.