Quality Improvement of Defects Region in Weld Radiography Image by the Wavelet Method Based on Adaptive Thresholding

In this paper, adaptive thresholding at the wavelet transform is utilized for improving the industrial radiography images. The quality of radiographic images is a very effective parameter in the defect determining by the experts. Therefore, the defect detection capabilities can be improved by the image processing algorithms. In this research, two-stage adaptive thresholding method has been used to improve the contrast of the inspected areas. The radiographic image is decomposed to several sub-bands using the wavelet function and the obtained coefficients are corrected by the threshold function. Then, the inverse wavelet transform is applied for obtaining the corrected image. Unlike the usual methods, in the threshold function of this method, the coefficient of under the threshold level is not zero and weakened by the multi-polyminal function. The advantages of this method are the continuity and derivability at threshold level. The proposed algorithm is implemented to the several radiographs of standard welded objects with known defects. The results have been evaluated by industrial radiography experts and show that the defect regions are clearer in reconstructed images than the original radiograph according to the operator perception analysis. Mean while, the dimensions and style of defects can be evaluated more precisely by this method.