The distribution of defects on the inner and outer surfaces of the capsule has an important impact on the success rate and efficiency of the laser inertial confinement fusion experiment. At present, the only way to observe the inner surface is to use direct X-ray imaging technology to obtain local two-dimensional information, while this paper uses micro-CT to photograph the capsule to non-destructively obtain the full surface morphology information of its inner and outer surfaces. On this basis, a technique is proposed for the observation of defects on the inner and outer surfaces of the capsule. First, the defect area was segmented by edge detection and ideal surface fitting. Then the defects were divided into blocks by connected domain analysis, and the width and height of each area were calculated separately. In addition, a stereoscopic image of the capsule was interactively drawn using a volume rendering method, especially highlighting the colored height field in the defect areas and allowing the inner surface of the capsule to be observed by virtual cutting. This method can identify most surface defects, and can calculate the defect size parameters more accurately. Unidentified defects can be observed as a supplement through volume rendering, providing researchers with a good aid for analysis.