The replacement of film technology with digital in radiographic monitoring during the construction and repair of pipelines is a pressing issue. The new technology is based on the use of phosphorous plates (computed radiography) and digital array detectors (digital radiography). However, there is the problem of defining requirements for digital radiograms. Selection of image quality class in digital radiographic testing is carried out in coordination with the customer, even though the requirements for the quality of film-based images are set out in regulations. The authors have carried out a study and provided a method making it possible to minimize the influence of the human factor in assessing the descriptiveness of radiographic images and to impartially determine the requirements for a digital radiogram when the use of computed and digital radiography may be recommended as a replacement for film technology. In order to solve this problem there has been a proposal to use ROC analysis, making it possible to estimate binary classification, including decryption of radiograms based on the «defect – no defect» principle. A test sample containing areas with defects (defect simulators) and areas with no such defects, which made it possible to use a binary classification for the analysis of its radiograms, has been developed and manufactured. A minimum sample size of independent experts sufficient to exclude the influence of the human factor on decoding test sample images has been identified. A film radiogram corresponding to regulatory requirements has been made in order to compare film-based and digital technologies. Following the analysis of this image, a base ROC curve has been plotted. This formed the basis for comparing ROC curves plotted following the results of decoding digital images with different quality parameters. It made it possible to establish values for quality parameters when the descriptiveness of digital radiograms was higher than of film-based images, and to draw conclusions about the circumstances in which it is possible to replace film-based technology with digital and computed radiography.
Optimization of Dose and Image Quality for Computed Radiography and Digital Radiography