Consumer-grade UAV imagery facilitates semantic segmentation of species-rich savanna tree layers.

Conventional forest inventories are labour-intensive. This limits the spatial extent and temporal frequency at which woody vegetation is usually monitored. Remote sensing provides cost-effective solutions that enable extensive spatial coverage and high sampling frequency. Recent studies indicate that convolutional neural networks (CNNs) can classify woody forests, plantations, and urban vegetation at the species level using consumer-grade unmanned aerial vehicle (UAV) imagery. However, whether such an approach is feasible in species-rich savanna ecosystems remains unclear. Here, we tested whether small data sets of high-resolution RGB orthomosaics suffice to train U-Net, FC-DenseNet, and DeepLabv3 + in semantic segmentation of savanna tree species. We trained these models on an 18-ha training area and explored whether models could be transferred across space and time. These models could recognise trees in adjacent (mean F1-Score = 0.68) and distant areas (mean F1-Score = 0.61) alike. Over time, a change in plant morphology resulted in a decrease of model accuracy. Our results show that CNN-based tree mapping using consumer-grade UAV imagery is possible in savanna ecosystems. Still, larger and more heterogeneous data sets can further improve model robustness to capture variation in plant morphology across time and space.

Space-borne Bose-Einstein condensation for precision interferometry.

Owing to the low-gravity conditions in space, space-borne laboratories enable experiments with extended free-fall times. Because Bose-Einstein condensates have an extremely low expansion energy, space-borne atom interferometers based on Bose-Einstein condensation have the potential to have much greater sensitivity to inertial forces than do similar ground-based interferometers. On 23 January 2017, as part of the sounding-rocket mission MAIUS-1, we created Bose-Einstein condensates in space and conducted 110 experiments central to matter-wave interferometry, including laser cooling and trapping of atoms in the presence of the large accelerations experienced during launch. Here we report on experiments conducted during the six minutes of in-space flight in which we studied the phase transition from a thermal ensemble to a Bose-Einstein condensate and the collective dynamics of the resulting condensate. Our results provide insights into conducting cold-atom experiments in space, such as precision interferometry, and pave the way to miniaturizing cold-atom and photon-based quantum information concepts for satellite-based implementation. In addition, space-borne Bose-Einstein condensation opens up the possibility of quantum gas experiments in low-gravity conditions1,2.

Navigation Aiding by a Hybrid Laser-Camera Motion Estimator for Micro Aerial Vehicles.

Micro Air Vehicles (MAVs) equipped with various sensors are able to carry out autonomous flights. However, the self-localization of autonomous agents is mostly dependent on Global Navigation Satellite Systems (GNSS). In order to provide an accurate navigation solution in absence of GNSS signals, this article presents a hybrid sensor. The hybrid sensor is a deep integration of a monocular camera and a 2D laser rangefinder so that the motion of the MAV is estimated. This realization is expected to be more flexible in terms of environments compared to laser-scan-matching approaches. The estimated ego-motion is then integrated in the MAV's navigation system. However, first, the knowledge about the pose between both sensors is obtained by proposing an improved calibration method. For both calibration and ego-motion estimation, 3D-to-2D correspondences are used and the Perspective-3-Point (P3P) problem is solved. Moreover, the covariance estimation of the relative motion is presented. The experiments show very accurate calibration and navigation results.

FDG PET/MR for lymph node staging in head and neck cancer.

OBJECTIVE: To assess the diagnostic value of PET/MR (positron emission tomography/magnetic resonance imaging) with FDG (18F-fluorodeoxyglucose) for lymph node staging in head and neck cancer. MATERIALS AND METHODS: This prospective study was approved by the local ethics committee; all patients signed informed consent. Thirty-eight patients with squamous cell carcinoma of the head and neck region underwent a PET scan on a conventional scanner and a subsequent PET/MR on a whole-body hybrid system after a single intravenous injection of FDG. The accuracy of PET, MR and PET/MR for lymph node metastases were compared using receiver operating characteristic (ROC) analysis. Histology served as the reference standard. RESULTS: Metastatic disease was confirmed in 16 (42.1%) of 38 patients and 38 (9.7%) of 391 dissected lymph node levels. There were no significant differences between PET/MR, MR and PET and MR (p>0.05) regarding accuracy for cervical metastatic disease. Based on lymph node levels, sensitivity and specificity for metastatic involvement were 65.8% and 97.2% for MR, 86.8% and 97.0% for PET and 89.5% and 95.2% for PET/MR. CONCLUSIONS: In head and neck cancer, FDG PET/MR does not significantly improve accuracy for cervical lymph node metastases in comparison to MR or PET.

PET/MR for therapy response evaluation in malignant lymphoma: initial experience.

OBJECT: To evaluate the feasibility of positron emission tomography/magnetic resonance imaging (PET/MR) with (18)fluoro-2-deoxyglucose (FDG) for therapy response evaluation of malignant lymphoma. MATERIALS AND METHODS: Nine patients with malignant lymphoma who underwent FDG-PET/MR before and after chemotherapy were included in this retrospective study. Average time between the two scans was 70 days. The scans were evaluated independently by two nuclear medicine physicians. The Ann Arbor classification was used to describe lymphoma stage. Furthermore, the readers also rated PET image quality using a five point scale. Weighted kappa (κ) was used to calculate interrater agreement. RESULTS: The initial scan showed foci of increased FDG uptake in all patients, with Ann Arbor stage varying between I and IV. In the follow-up examination, all but one patient showed complete response to chemotherapy. PET image quality was rated as very good or excellent for all scans. Interrater agreement was excellent regarding Ann Arbor stage (κ = 0.97) and good regarding image quality (κ = 0.41). CONCLUSION: PET/MR shows promising initial results for therapy response evaluation in lymphoma patients.