Deep-learning-based object detection for automated IQI evaluation in analog film radiographic images of piping welds
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Radiographic testing (RT) is a nondestructive testing method for evaluating internal defects in welds, and image quality is assessed using image quality indicators (IQIs). In this study, two deep-learning-based object detection models, YOLOv8x and Faster R-CNN, were applied to detect IQI wires in analog RT images automatically. YOLOv8x demonstrated excellent performance, achieving a precision of 0.987, recall of 0.972, and mAP@50 of 0.992. The wire count match rate in the low-contrast images was 55.0% for YOLOv8x and 50% for Faster R-CNN. This study demonstrates the feasibility of automating IQI wire detection in analog RT images, and provides a foundation for future digital RT systems.
Abstract
Radiographic testing (RT) is a nondestructive testing method for evaluating internal defects in welds, and image quality is assessed using image quality indicators (IQIs). In this study, two deep-learning-based object detection models, YOLOv8x and Faster R-CNN, were applied to detect IQI wires in analog RT images automatically. YOLOv8x demonstrated excellent performance, achieving a precision of 0.987, recall of 0.972, and mAP@50 of 0.992. The wire count match rate in the low-contrast images was 55.0% for YOLOv8x and 50% for Faster R-CNN. This study demonstrates the feasibility of automating IQI wire detection in analog RT images, and provides a foundation for future digital RT systems.