The accurate and timely monitoring of asphalt mixture paving process is critical for reducing segregation and effective construction. In this study, 1085 construction site images were pre-processed by digital image processing techniques (i.e., grayscale, filtration, and binarization) to explore and build up a real-time evaluation method for paving asphalt mixture during construction process. The sum and standard deviation of the static moment of each particle larger than 9.5 mm was calculated and the min-max normalized standard deviation were utilized to evaluate the paving asphalt mixture distribution uniformity. Results show that the mix proportion value from the proposed image technique was close to the value from the reality. By comparing the value of min-max normalized static moment standard deviation with the corresponding image, a uniformity indicator was proposed for the evaluation of asphalt pavement uniformity during paving process. The paper provides a new method for paving asphalt mixture segregation detection.