Aiming at the real-time, continuous, and quantitative evaluation of the homogeneity of asphalt concrete, taking the Jinan-Qimen Highway of Dangshan section in Anhui as an example, this study adopted the digital image processing technology, statistical method, and the principle of shape from shading to establish the homogeneity evaluation method and standard of asphalt concrete based on the quadrangular static moment variability and the structural depth variability in paving state and compaction state. Moment variation coefficient and structure depth variation coefficient were proposed as evaluation indexes to evaluate the homogeneity of asphalt concrete. To obtain accurate image information of asphalt concrete surface, the processes of digital image acquisition and pre-processing were standardized, and the illumination intensity and shooting height of image acquisition were recommended. The study showed that the proposed digital image evaluation method for the homogeneity of asphalt concrete was feasible and efficient, and the results were quantitative and objective, which could reflect homogeneity in the construction quality of asphalt concrete well. To achieve homogeneity of asphalt concrete paving, the moment variation coefficient should be controlled within (0,1.5%]. When AC-13 and AC-25 asphalt concretes were homogeneous in compaction state, the structure depth variation coefficient should be controlled within (0,4.0%] and (0,8.0%], respectively. Considering the method of subdividing asphalt concrete compaction image into 16 equal areas as the optimum, the correlation equation of the digital image structure depth variation coefficient in different equal subdivision ways was established by coefficient simulation, and the correlations were remarkable.