ALBERT

Beschreibung: Dilute ensembles of granular matter (so-called granular gases) are nonlinear systems which exhibit fascinating dynamical behavior far from equilibrium, including non-Gaussian distributions of velocities and rotational velocities, clustering, and violation of energy equipartition. In order to understand their dynamic properties, microgravity experiments were performed in suborbital flights and drop tower experiments. Up to now, the experimental images were evaluated mostly manually. Here, we introduce an approach for automatic 3D tracking of positions and orientations of rod-like particles in a dilute ensemble, based on two-view video data analysis. A two-dimensional (2D) localization of particles is performed using a Mask R-CNN neural network trained on a custom data set. The problem of 3D matching of the particles is solved by minimization of the total reprojection error, and finally, particle trajectories are tracked so that ensemble statistics are extracted. Depending on the required accuracy, the software can work fully self-sustainingly or serve as a base for subsequent manual corrections. The approach can be extended to other 3D and 2D particle tracking problems.

Beschreibung: Adequate mechanical stimulation is essential for cellular health and tissue maintenance, including articular cartilage, which lines the articulating bones in joints. Chondrocytes, which are the sole cells found in articular cartilage, are responsible for matrix synthesis, maintenance and degradation. It is generally believed that chondrocytes require mechanical stimuli through daily physical activity for adequate cartilage homeostasis. However, to date, the molecular mechanisms of cellular force sensing (mechanotransduction) are not fully understood. Among other mechanisms, the cytoskeleton is thought to play a key role. Despite that gravity is a very small force at the cellular level, cytoskeletal adaptations have been observed under altered gravity conditions of a parabolic flight in multiple cell types. In this study, we developed a novel hardware which allowed to chemically fix primary bovine chondrocytes at 7 time points over the course of a 31-parabola flight. The samples were subsequently stained for the microtubules and vimentin network and microscopic images were acquired. The images showed a large heterogeneity among the cells in morphology as well as in the structure of both networks. In all, no changes or adaptions in cytoskeleton structure could be detected over the course of the parabolic flight.

Beschreibung: Samples of European beech (Fagus sylvatica) were used for this study. Logs of these samples covered a scatter of mild-to-strong curvatures and the boards of these samples covered strong fiber deviations. This study consists of two separate parts: (1) log reconstruction and optimization of the cutting pattern, and (2) board reconstruction and strength prediction. Information about the internal quality of the logs is missing in this study, as laser scanning has been used for surface reconstruction of logs. Therefore, two separate steps were implemented here. (1) Influence of cutting pattern and board-dimensions on yield were analyzed. For this step, 50 logs were checked. (2) A more advanced numerical method based on the finite element (FE) analysis was developed to improve the accuracy of tensile strength predictions. This step was performed, because visual grading parameters were relatively weak predictors for tensile strength of these samples. In total, 200 beech boards were analyzed in this step. However, due to the geometrical configuration of some knots, the reconstruction and numerical strength prediction of 194 boards out of 200 boards were possible. By performing tensile tests numerically, stress concentration factors (SCFs) were derived, considering the average and maximum stresses around the imperfections. SCFs in combination with the longitudinal stress wave velocity were the numerical identifying parameters (IPs), used in the nonlinear regression model for tensile strength prediction. The influence of the combination of different numerical parameters in the developed non-linear model on improving the quality of the strength prediction was analyzed. For this reason, improvement of coefficient of determination (R2) after adding each parameter to the multiple regression analysis was checked. Performance of the developed numerical method was compared to the typical grading approaches [using knottiness and the dynamic MoE (MoEdyn)], and it was shown that the coefficient of determination is higher, when using the virtual methods for tensile strength predictions.