Abstract: In order to improve the speed and accuracy of cone beam CT image quality assurance, an automatic image quality assurance software based on feature extraction is designed and implemented. This paper introduces how the software can realize the positioning of Catphan500 phantoms and the selection of ROI in each module through Canny algorithm and Hough transform circle detection. For the different modules in the Catphan500 phantom, this paper describes the calculation methods of HU accuracy, spatial linearity, HU uniformity and spatial linearity in the software. Finally, the article verifies the feasibility of the software through experiments.