Automatic Recognition for Nuclear Fuel Rod Weld Defect Based on X-ray Digital Image

摘要

We have researched the automatic recognition method for nuclear fuel rod weld defects and developed the corresponding software platform. The method is consisted of two parts. They are image preprocessing and defect recognition. Firstly, we analyzed the characteristics of the nuclear fuel rod welding image such as the welding structure and the image noise. And then we researched the appropriate preprocessing method and the defects recognition method based on the waveform analysis method. Finally we realized the automatic recognition of nuclear fuel rod weld defects based on the X-ray digital image.