In the industrial radiography of large storage tanks, the presence of noise, thickness non-uniformity, photon scattering within the radiation beam, unwanted background components, and the distance between the radiographic film and the tank wall can significantly reduce defect detectability and degrade radiographic image quality. In this study, a method based on Stochastic Factor Graphs (SFGs) is proposed for background removal and quality enhancement of radiographic images. In the proposed approach, the statistical dependencies among pixels and the structural characteristics of the image are modeled using a factor graph framework. Through optimal parameter estimation, the background components are separated from regions containing weld and defect-related information. The resulting background-corrected image is then subtracted from the original radiograph on a pixel-by-pixel basis to enhance image contrast. The results demonstrated that the proposed method substantially improved defect detectability. According to the evaluation performed by radiography experts, the contrast in the Image Quality Indicator (IQI) region increased from 20 % in the original image to 89 % in the reconstructed image. Furthermore, the detection accuracy for the weld line increased from 40 % to 83 %, while defect detection accuracy improved from 47 % to 94 %. The detection of lead letters increased from 83 % to 100 %, and the detectability of the lead ruler improved from 90 % to 100 %. These results demonstrate a significant improvement in image contrast, effective suppression of background effects, and enhanced visibility of fine structural details in radiographic images of large tanks. The findings indicate that the proposed SFG-based approach can serve as an effective tool for background removal and image-quality enhancement in industrial radiography. Moreover, the method has the potential to improve the accuracy of non-destructive testing (NDT) inspections, reduce operator-dependent interpretation errors, and enhance the reliability of the evaluation of large tanks and welded structures.