In order to achieve realistic simulations of the chip formation, high quality input data regarding the flow stress and damage behavior of the materials are required. The split Hopkinson pressure bar (SHPB) test setup for the characterization of highly dynamic material properties offers a suitable method for generating high strain rates, similar to those in the chip formation zone. However, the strain measurement in SHPB is usually performed by means of strain gauges. This leads to an unreliable evaluation of strain rate and flow stress/shear flow stress in the case of an inhomogeneous material deformation, since this method presents the total strain whilst excluding local deformations. Inhomogeneous deformations are induced deliberately in special shear specimens, as they are also observed in the investigated cylindrical specimens. The present work deals with this issue by providing two additional measurement techniques, which are applied in SHPB tests for cylindrical specimens made of AISI 1045 and Ti6Al4V. To enable a local strain resolution, digital image correlation (DIC) is applied to high-speed images of the deformation process. In order to allow for the detection of shear bands in the specimens, a deep-learning-based approach is presented. The two measurement methods (strain gauges and DIC) are compared and discussed. In particular, the findings regarding the inhomogeneous deformation of Ti6Al4V allow for future improvements in the result quality of SHPB tests. The presented algorithm shows promising predictions for shear band detection and creates the basis for an automated evaluation of shear sample results, as well as an AI-based pre-selection of frames for the DIC evaluation of SHPB tests.

Monitoring corrosion in industrial components is essential for ensuring structural integrity, operational safety, and economic efficiency. However, the nonlinear nature of corrosion leads to irregular visual and physical surface changes that conventional image-processing methods struggle to capture. Supervised deep-learning approaches require extensive annotated datasets, often unavailable for diverse material conditions. This study uses the pre-trained CLIP (Contrastive Language–Image Pre-training) model for semantic analysis of nonlinear corrosion progression. By embedding images and descriptive text into a shared latent space, the method detects subtle optical variations, identifies characteristic corrosion features, and derives a directional metric representing progression from initiation to advanced states. Optical analyses of corrosion test plates provide a controlled dataset for evaluating surface evolution and validating the semantic similarity metrics. The results indicate that latent space analysis offers a practical, data-driven approach to automated corrosion monitoring.