Towards Enhancing Automated Defect Recognition (ADR) in Digital X-ray Radiography Applications: Synthesizing Training Data through X-ray Intensity Distribution Modeling for Deep Learning Algorithms

Abstract

Industrial radiography is a pivotal non-destructive testing (NDT) method that ensures quality and safety in a wide range of industrial sectors. Conventional human-based approaches, however, are prone to challenges in defect detection accuracy and efficiency, primarily due to the high inspection demand from manufacturing industries with high production throughput. To solve this challenge, numerous computer-based alternatives have been developed, including Automated Defect Recognition (ADR) using deep learning algorithms. At the core of training, these algorithms demand large volumes of data that should be representative of real-world cases. However, the availability of digital X-ray radiography data for open research is limited by non-disclosure contractual terms in the industry. This study presents a pipeline that is capable of modeling synthetic images based on statistical information acquired from X-ray intensity distribution from real digital X-ray radiography images. Through meticulous analysis of the intensity distribution in digital X-ray images, the unique statistical patterns associated with the exposure conditions used during image acquisition, type of component, thickness variations, beam divergence, anode heel effect, etc., are extracted. The realized synthetic images were utilized to train deep learning models, yielding an impressive model performance with a mean intersection over union (IoU) of 0.93 and a mean dice coefficient of 0.96 on real unseen digital X-ray radiography images. This methodology is scalable and adaptable, making it suitable for diverse industrial applications.

Advances in Metal Casting Technology: A Review of State of the Art, Challenges and Trends—Part II: Technologies New and Revived

“…Providing synthetic training data may elegantly solve this problem by virtually generating component models-or submodels of specific regions of interest-containing varied defect distributions and simulating the X-ray scanning or CT process. A secondary advantage of this concept is that it may help solve the issue of missing ground truth data, as it eliminates the need for manually labeled training data and, thus, the possibility of errors running up in the labeling process, which is tedious enough as it stands [316,317]. Fuchs et al discuss this issue in detail, highlighting and even showcasing the variations observed in labeling, even when entrusting domain experts with this task.…”