Adaptive pseudo-colour image enhancement algorithm for high-grey X-ray film based on pixel self-transformation

Abstract

12-bit X-ray film images with high greyscale values and low contrast, produced in the field of X-ray non-destructive testing, have the problems of a poor visualisation effect and serious greyscale loss when displayed on conventional monitors. In order to solve these problems, an APEA_PST algorithm for pseudo-colour image enhancement is proposed, which is suitable for enhancing super-8-bit X-ray film images. Firstly, in order to solve the problem that super-8-bit images cannot be displayed directly on 8-bit monitors, a RAW-Optical preprocessing algorithm, based on a log function, is designed. Secondly, in order to improve the overall contrast of the preprocessed images, a super-8-bit non-linear superposition gain compensation model, called G-NLM, is designed based on a combination of log mapping and grey-level change. Then, the novel super-8-bit APEA_PST pseudo-colour algorithm, based on pixel self-transformation, is proposed, which solves the problems of poor adaptive ability and a poor image display effect, enhanced by a traditional pixel self-transformation algorithm. Finally, focusing on super-8-bit greyscale images, generated in the field of industrial welding manufacturing, the APEA_PST algorithm and other classical pseudo-colour algorithms are used to enhance a specific radiographic image. The qualitative and quantitative results show that the image, as enhanced by the APEA_PST algorithm, is not only more suitable for human eyes to observe, but that the algorithm also performs well in quantitative experiments involving Entropy and IL-NIQE, showing good universality and superiority.