Abstract: In the present study, typical black soil in northeastern China was selected as the test object, and the simulated image processing was adopted combined with computerized tomography (CT) scanning. With the extracted image feature by gray level co-occurrence matrix and principal component analysis (PCA), the Euclidean distance between the feature vector of test image and verify image was calculated, which built the basis for automatic discrimination of different soil physical states caused by freeze-thaw. It was shown that it could realize automatic identification of different soil physical state by image features extracted by either gray level co-occurrence matrix or PCA. And the identification accuracy of gray-level co-occurrence matrix method was higher than that of principal component analysis method for the same soil CT tomography image database.

Key words: soil computed tomography images, gray level co-occurrence matrix, principal component analysis, Euclidean distance, identification accuracy