X-radiography inspection techniques are generally used in non-destructive evaluation

industry. Manual assessment of the inspection may fail and turn into false assessment due

to a large number of examining while inspection process. Accordingly, incorrect

assessment through human vision may cause immense industry disaster. Therefore, it is

essential to examine through machine vision, which is capable to avoid false assessment.

Digital X-radiography has widely been used in Non-Destructive Testing (NDT) and

particularly in weld defect detection. Weld defects occurrence is an unavoidable problem

during the welding process. Most common weld defects some of which are porosity, gas

pore, tungsten inclusion, longitudinal crack, lack of penetration, and slag inclusion.

Digital image processing techniques are a foremost way to experiment in the NDT. The

detection and classification of weld defects depend on the quality of the digitized image,

which is subjected to certain factors, such as noise, the mode of the image histogram,

defects of different dimensions, indiscernible defects in the image background, and low

contrast or unevenly illuminated image. More accuracy can be achieved during classification

of weld defects are always be subject to the deliverables of low and mid-level image

processing techniques. Therefore, it is desirable to provide more importance for these levels

in the weld X-radiography image.