The internal meso-scale damage evolution information of concrete under external load can be recorded by CT images. Subtle differences exist in the original digital matrix corresponding to the CT images of the same section at different stress stages before the peak strength. The subtle differences are the key to study the meso-scale damage initiation and evolution of concrete material,however,such differences cannot be intuitively observed with the naked eye in CT imagesIn this study,to dig out the subtle difference information and investigate meso-scale damage evolution process of concrete,a procedure based on the theory of grey level co-occurrence matrices was established,and the effectiveness of this procedure was verified by the CT test results of statically uniaxial compression and tension of concrete specimens. Results shown that this method not only can evaluate the distribution of initial meso-scale damage and the process of meso-scale damage evolution,predict the fracture position of the specimen,and determine the maximum meso-damage section,but it can also establish the quantitative relationship between statistical characteristic value and meso-scale damage. The changes of the four statistical characteristic values (contrast,energy,correlation,and homogeneity) at different stress stages shown that the meso-scale damage has a considerable development between the 42th and 235th cross-sectional CT image when the stress reaches to 96.2% of the peak strength in the specimen subjected to statically compressive load,and the cracks appear in this area after specimen failure. Before peak load,meso-scale damage cannot be found in the statically tensile specimen,and cracks appear between the 89th and 116th cross-sectional CT images after the peak load. In addition,the GLCM of the CT images was illustrated as colored heatmaps. The changes of diagonal distribution,peak number,and bandwidth in the colored heatmap can reflect the intensity of the gray distribution and the degree of damage.