Welding is a critical process in the fabrication of oil and chemical storage tanks, serving as the primary method for joining metal plates. Different Non-Destructive inspection methods can be implemented for investigating of weldments in the storage tanks. Among them, industrial radiography is one of the best methods because of generated images which can give a fingerprint of welding and heat affected zone. However, radiographic images obtained using gamma sources of welded areas of large tanks are often of poor quality. Radiographic testing is the most widely employed technique for the evaluation of weld integrity in such structures. However, radiographic images acquired with a Gamma-ray source from welded regions of large tanks often suffer from poor quality. This problem is mainly geometrical effect that caused by radiation scattering, which occurs due to the distance between the film and the weld as a result of the insulation, leading to a reduction in image contrast. Quantifying haze in radiographic images can contribute to enhancing image quality. The Goldstein–Fattal (GF) dehazing method offers a valuable solution in this context, as it operates without requiring precise a priori information about the weld structure. Instead, it relies on the general statistical properties of the image, making it particularly well-suited for weld radiography. This property enables effective enhancement of low-quality radiographs. In the present study, the GF method was applied to remove haze and improve the detectability of weld defects in radiographic images of tanks of chemical material. These radiographs were obtained using traditional film-based techniques with Ir-192. Experimental findings show that the GF method successfully enhanced image contrast in several regions containing defects. By improving image clarity, this approach reduces the likelihood of misidentifying weld flaws, thereby increasing the reliability and safety of storage tanks. Furthermore, when the results of GF method compared with the Perona–Malik technique, the GF method exhibited superior performance, particularly in the root regions of welds where contrast improvement is most critical. Independent evaluations by radiographic experts corroborated these results.