Abstract: Since inclusions will inevitably be introduced during the granulation of powder materials and they will definitely affect the performance and quality of the product, it is necessary to conduct non-destructive testing of the inclusions introduced therein. A multi-angle automatic detection method for inclusions in DR images of powder materials is proposed. Firstly, aiming at the problems of low contrast and random noise in DR images, the image is enhanced by adjusting the window width and level and image filtering, and then aiming at the multi-angle detection of the same inclusion, The inclusions were feature-matched using the SIFT algorithm, and then the inclusions were segmented from the background. The maximum value of the measured size at different scanning angles is regarded as the approximate value of the inclusion size, and finally the inclusion volume is obtained by calculating the size, and the quality of the powder material is evaluated on this basis. The feasibility of the method in this paper is verified through the metal wire measurement experiment, and the method in the follow-up actual test verifies that the method in this paper can meet the actual testing needs.