Improvement of defect size Measurement in Radiography Images by Recursive Image Processing Method

Measuring defect sizes is very important in determining weld strength. This size is also used to evaluate the object on the test for accept or reject criteria according to the relative standards. One of the important methods for measuring different defects' sizes is industrial radiography testing (RT), a non-destructive testing method. Radiography is carried out using penetrating X or Gamma rays. Radiography is a volumetric method and can give information from the inside the different objects. Various factors, such as the magnification of the radiographic image, the fogginess of the radiographs, the non-point source, and the inherent scattering of X-rays, affect on the measurement and the accuracy of the defect sizing. In particular, size measurement for defects with larger distance respect to the film or radiography detector can be with more uncertainty. This is due to more shadowiness of rays with larger distances. For radiography imaging, an industrial Computed Radiography System (CR) was used. General purpose imaging phosphor plates with a laser resolution of 50 micrometers have been used. The X-ray source was an industrial powerful X-ray tube with a voltage of up to 300 kilovolts. In this research, utilizing the distance measurement between the lines in the duplex image quality indicator (DIQI) tool as a known length, the size of defects in standard parts in Sonakit educational test kit that have specific defects is estimated and compared with the original value. To decrease the blurring, a recursive filter method is used to make the edges sharper to better estimate the defect sizes. The results show that the measurement of defects is related to the accuracy of the user in estimating the pixels of the defect regions and the sharpness of the edges. The measurement error is reported between 6% and 19% for the defect measurement in the standard parts examined.