An Investigation in Applying Image Retrieval Techniques to X-Ray Engineering Pictures

摘要

Using image retrieval techniques in analysing Non-destructive testing results is a new challenge in both computing science and engineering applications. Objective of this research is to develop an image retrieval system to analyse X-ray images for welding industry. The content based image retrieval has been used in this investigation, particularly in feature vector paradigm and similarity as well as detailed analysis towards single defects. It is found that X-ray images can be digitally analysed qualitatively and quantitatively easily. It concludes that the use of existing CBIR techniques can provide a platform to quickly develop new image analysis tools.