Autoencoder-Based Anomaly Detection in Industrial X-ray Images

Within many quality-critical industries, e.g. the aerospace industry, industrial X-ray inspection is an essential as well as a resource intense part of quality control. Within such industries the X-ray image interpretation is typically still done by humans, therefore, increasing the interpretation automatization would be of great value. We claim, that safe automatic interpretation of industrial X-ray images, requires a robust confidence estimation with respect to out-of-distribution (OOD) data. In this work we have explored if such a confidence estimation can be achieved by comparing input images with a model of the accepted images. For the image model we derived an autoencoder which we trained unsupervised on a public dataset with X-ray images of metal fusion-welds. We achieved a true positive rate at 80–90% at a 4% false positive rate, as well as correctly detected an OOD data example as an anomaly.

Research on the Measurement Accuracy of Ultrasonic Phased Array Time-of-Flight Method

In special public pipelines, High Density Polyethylene (HDPE) had been widely applied, making integrity and safety priority. In the periodic inspection of pressure vessel and pressure pipeline, the safety level shall be determined by the defect geometric size. Therefore, it is necessary to get the defect size data as accurately as possible. Digital radiography inspection is a feasible method for detecting butt fusion welds. Defect size can be measured accurately by digital radiographic image. Due to factors such as damage to human body caused by ray, it is difficult to detect HDPE pipelines in some occasions in field. Ultrasonic phased array testing has high accuracy in flaw inspection, high efficiency and little artificial criteria error. But conventional phased array has difficulties in detecting butt fusion welds due to acoustic attenuation and noise affection. In this paper, PE butt fusion welding is inspected by immersion phased array. The absolute time-of-flight method is used to measure the defect size, and the accuracy is qualified with the requirements of the butt fusion welding inspection. For more important, inspector does not need to move equipment probe, avoiding artificial errors and improving the efficiency.