Abstract: The primary work of defect detection of Digital Radiography（DR） image is the segmentation and measurement of defect regions.The precision of defect measurement，to a large extent，depends on the accuracy of segmentation.Based on the curve evolution theory and level set method for segmentation，C-V model has taken regional information into account so that a globally optimal solution can be got，on the other hand，this model also can naturally handle contour topology changes.Aiming at the traits of DR image，this paper researches on a C-V method of defects detection in DR image.In the method，firstly，C-V model is applied to defects segmentation for DR image，on this basis，defects measurement is achieved.The experiment shows that the C-V method can accurately divide defect regions of DR image，and get its geometric parameters.