With increased use of light-weight materials with low factors of safety, non-destructive testing becomes increasingly important. Thanks to the advancement of infrared camera technology, pulse thermography is a cost efficient way to detect subsurface defects non-destructively. However, currently available evaluation algorithms have either a high computational cost or show poor performance if any geometry other than the most simple kind is surveyed. We present an extension of the thermographic signal reconstruction technique which can automatically segment and image defects from sound areas, while also estimating the defect depth, all with low computational cost. We verified our algorithm using real world measurements and compare our results to standard active thermography algorithms with similar computational complexity. We found that our algorithm can detect defects more accurately, especially when more complex geometries are examined.

Ground-penetrating radar (GPR) is a non-invasive technology that uses electromagnetic pulses for subsurface exploration. In the railroad sector, it is crucial to assessing soil layers and infrastructure, offering insights into soil stratification and geological features and aiding in identifying subsurface hazards. However, the automation of radargram analysis is impeded by the lack of ground truth—accurate real-world data used to validate machine learning models—thus affecting the deployment of advanced algorithms. This study focuses on generating high-quality simulated data to address the shortage of real-world data in the context of object detection along railroad tracks and presents a fully automated pipeline that includes data generation, algorithm training, and validation using real-world data. By doing so, it paves the way for significantly easing the future task of object detection algorithms in the railway sector. A simulation environment, including the digital twin of a GPR antenna, was developed for artificial data generation. The process involves pre- and post-processing techniques to transform the three-dimensional data from the multichannel GPR system into two-dimensional datasets. This ensures minimal information loss and suitability for established two-dimensional object detection algorithms like the well-known YOLO (You Only Look Once) framework. Validation involved real-world measurements on a track with predefined buried objects. The entire pipeline, encompassing data generation, processing, training, and application, was automated for efficient algorithm testing and implementation. Artificial data show promise for better performance with increased training. Future AI and sensor advancements will enhance subsurface exploration, contributing to safer and more reliable railroad operations.

In the past decades, developments in the field of computer vision have made both the software and hardware more capable and more easily accessible. This has enabled otherwise complex vision systems to be used in other fields, such as autonomous robotics. Although vision systems in the visible light spectrum are commonplace in robotics nowadays, the thermal spectrum is still rarely used, even though it offers certain advantages. A thermal camera can sense the temperature of objects, is independent of illumination and can actually see through heavy smoke and fog. This makes it a useful tool in particular in the field of rescue robotics, where poor vision conditions are to be expected. In this paper, the feasibility of using two thermal cameras in a stereo vision setup to map indoor scenes is to be examined. It is meant to allow an autonomous robot to perceive its indoor surroundings as a 3D space, even in poor vision conditions. The biggest challenges are the calibration of thermal cameras and the proper filtering of the raw image and the resulting disparity map. Simple and easily implemented solutions are proposed for each of these issues.

Ground penetrating radar (GPR) has been used for several years as a non-contact and non-destructive measurement method for rail track analysis with the aim of recording the condition of ballast and substructures. As the recorded data sets typically cover a distance of many kilometers, the evaluation of these data involves considerable effort and costs. For this reason, there is an increasing need for automated support in the evaluation of GPR measurement data. This paper presents an image segmentation pipeline based on 2D Gabor filter texture analysis, which can assist users in GPR data-based track condition assessment. Gabor filtering is used to transform a radargram image (or B-scan) into a high-dimensional, multi-resolution representation. Principal component analysis (PCA) is then applied to reduce the data content to three characteristic dimensions (namely amplitude, frequency, and local scattering) to finally obtain a segmented radargram image representing different classes of relevant image structures. From these results, quantitative measures can be derived that allow experts an improved condition assessment of the rail track.

The present paper deals with the application of atomic force microscopy (AFM) as a tool for morphological characterization of histological brain tumor samples. Data mining techniques will be applied for automatic identification of brain tumor tissues based on AFM images by means of classifying grade II and IV tumors. The rapid advancement of AFM in recent years turned it into a valuable and useful tool to determine the topography of surface nanoscale structures with high precision. Therefore, it is used in a variety of applications in life science, materials science, electrochemistry, polymer science, biophysics, nanotechnology, and biotechnology. Minkowski functionals are used (in particular the Euler- Poincaré characteristic) as a feature descriptor to characterize global geometric structures in images related to the topology of the AFM image. In order to improve classification accuracy on the one hand, but to infer interpretable information from AFM images for domain experts on the other hand, feature analysis and reduction will be applied. From a data mining point of view, Genetic Programming will be introduced as a sophisticated method for both feature analysis and reduction as well as for producing highly accurate and interpretable models. Support Vector Machines will be used for comparison reasons when talking about reachable model accuracy.

This paper presents an automated high-speed rolling mark recognition system for railroad rails utilizing image processing techniques. Rolling marks, which consist of numbers, letters, and special characters, were engraved into the rail web as 3D information. These rolling marks provide crucial details regarding the rail manufacturer, steel quality, year of production, and rail profile. As a result, they empower rail infrastructure managers to gain valuable insights into their infrastructure. The rolling marks were captured using a standard color camera under dark field illumination. The recognition of individual numbers, letters, and special characters was achieved through state-of-the-art deep neural network object detection, specifically employing the YOLO architecture. By leveraging reference rolling marks, the detected characters can then be accurately interpreted and corrected. This correction process involves calculating a weighted Levenshtein distance, ensuring that the system can identify and rectify partially misidentified rolling marks. Through the proposed system, the accurate and reliable identification of rolling marks was achieved, even in cases in which there were partial errors in the detection process. This novel system thus has the potential to substantially improve the management and maintenance of railroad infrastructure.