Irregular pipes play an extremely important role in special industries. The detection of subtle defects on their inner surface faces structural characteristics, such as small space and free change of radial size, which makes the traditional in-pipeline imaging detection method has the problem of out-of-focus caused by the large change of depth of field scale when imaging in succession, and affects the system′s high-resolution imaging. To solve this problem, a method for the cross-mode full depth of field imaging is proposed, which is by focusing the imaging surface depth data and imaging system parameter data. The current focus mode is determined adaptively, and the corresponding aggregate of the focal plane depth data will be output. A radial cross-mode imaging device with a depth-sensing function is developed, and an imaging experiment is implemented for the features of fine lines in the inner wall of an irregular pipe. Results show that the designed imaging device can obtain the single focus image or fold focal sequence images adaptively, a multi-focus fusion operation in the wavelet domain is performed on the stacked sequence images to obtain the full depth of field image, and the obtained full depth of field image can distinguish tiny fine lines with a width of 0. 5 mm under the condition of a 40~ 1 000 mm depth span. The method and device can provide reliable image support for defect detection and damage analysis of irregular pipes and other similar scenes.