Abstract: In order to realize the accurate identification of rock heterogeneity，aiming at the problem that traditional methods are difficult to characterize rock heterogeneity efficiently，the research on identify‐ ing rock heterogeneity based on digital image technology is carried out from the statistical theory of rock heterogeneity. The results show that the color heterogeneity can be effectively solved by considering the im‐ age pixels as discrete rock medium micro-elements and using the pixel color distribution characteristics to fit by Weibull function. It is found that there is a good exponential function relationship between the pixel color heterogeneity and the strength heterogeneity of rock micro-units. Therefore，a rock heterogeneity identification method based on digital image technology is established. At the same time，by studying the size effect of rock heterogeneity under color feature characterization，the heterogeneity of rock matrix and the minimum characterization area are determined，and the applicability and reliability of this method in different types of rocks are further discussed. This method not only provides a new way to quickly under‐ stand the uniformity of rock mass in the field，but also provides strong technical support for the application of numerical simulation method based on statistical theory.

Key words: digital image, Weibull distribution, heterogeneity, size effect