In order to evaluate the segregation level of bituminous pavement quickly and effectively, the digital image processing technique was adopted in studying the heterogeneity of bituminous pavement. The drawbacks of common evaluation methods on image processing of bituminous pavement segregation were firstly discussed, and a new method of evaluating the bituminous pavement segregation was proposed, which is based on the analysis of the macroscopic construction width of the concave area in pavement texture. This method is to get the binary image by taking photos of the concave areas in macroscopic construction, and defines the average width of the related concave area in the binary image as the macroscopic construction width K of bituminous pavement, and then builds a model based on the construction width K and depth TD to evaluate the aggregate segregation level of bituminous mixture. It is found that the construction width K is correlated with the conventional construction depth measured by the traditional manual sand patch method, which better characterizes the aggregate segregation level of bituminous pavement.