A Study on Automatic Flaw Detection using MSSIM in Ultrasound Imaging of Steel Plate

Ultrasound imaging had been widely used in nondestructive evaluation (NDE) of materials due to its operational safety and portability. The interpretation of result depends on specialist's judgement which can sometimes be ambiguous and subjective. In image quality assessment, Mean Structural Similarity index (MSSIM) is a proven method to quantitatively evaluate image quality, which is highly matched with qualitative evaluation of human. This method can statistically calculate difference between two images and represent as a number. Therefore, applying MSSIM on ultrasound images will quantify difference between compared imaged into number. In case ultrasound image of a benchmark material which contains no flaw is available, MSSIM has high potential to create a machine that can automatically recognize presence of flaw in materials. As a fundamental study of this technique, this paper tested the proposed technique in detection of steel plate bottom flaw.