Driven by recent technological breakthroughs, the electric vertical take-off and landing (eVTOL) aircraft has gained considerable attention. The widespread demand for eVTOL aircraft can be attributed to their potential use in the commercialisation of urban air mobility (UAM) in low-altitude urban airspaces. However, the urban low-altitude airspace environment is complex. UAM has a high traffic density and the eVTOL aircraft specifications are not uniform. Particularly in commercial scenarios, controlling eVTOL aircraft and ensuring safety in UAMs are the two major problems that should be addressed in future studies. The design of reasonable traffic rules is a potential solution; hence, we organised a UAM traffic rule system and proposed several alternative UAM traffic rules from three perspectives: a single eVTOL aircraft, a certain route, and key control areas. In addition, we validated these traffic rules using multi-rotor and fixed-wing eVTOL aircraft. The results show that designing reasonable traffic rules can facilitate attaining the primary objectives of commercialisation of UAM.
It is a challenging task to accurately segment damaged road markings from images, mainly due to their fragmented, dense, small-scale, and blurry nature. This study proposes a multi-scale spatial kernel selection net named M-SKSNet, a novel model that integrates a transformer and a multi-dilated large kernel convolutional neural network (MLKC) block to address these issues. Through integrating multiple scales of information, the model can extract high-quality and semantically rich features while generating damage-specific representations. This is achieved by leveraging both the local and global contexts, as well as self-attention mechanisms. The performance of M-SKSNet is evaluated both quantitatively and qualitatively, and the results show that M-SKSNet achieved the highest improvement in F1 by 3.77% and in IOU by 4.6%, when compared to existing models. Additionally, the effectiveness of M-SKSNet in accurately extracting damaged road markings from images in various complex scenarios (including city roads and highways) is demonstrated. Furthermore, M-SKSNet is found to outperform existing alternatives in terms of both robustness and accuracy.
To tackle the complexity and limited applicability of high-precision segmentation models for damaged road markings, this study proposes a Multi-level Adaptive Lightweight Network (MALNet) based on knowledge distillation. By incorporating multi-scale dilated convolution and adaptive spatial channel attention fusion modules, the MALNet model significantly enhances the precision, integrity, and robustness of its segmentation branch. Furthermore, it employs an intricate knowledge distillation strategy, channeling rich, layered insights from a teacher model to a student model, thus elevating the latter’s segmentation ability. Concurrently, it streamlines the student model by markedly reducing its parameter count and computational demands, culminating in a segmentation network that is both high-performing and pragmatic. Rigorous testing on three distinct data sets for damaged road marking detection—CDM_P (Collective Damaged road Marking—Public), CDM_H (Collective Damaged road Marking—Highways), and CDM_C (Collective Damaged road Marking—Cityroad)—underscores the MALNet model’s superior segmentation abilities across all damage types, outperforming competing models in accuracy and completeness. Notably, the MALNet model excels in parameter efficiency, computational economy, and throughput. After distillation, the student model’s parameters and computational load decrease to only 31.78% and 27.40% of the teacher model’s, respectively, while processing speeds increase to 1.9 times, demonstrating a significant improvement in lightweight design.
Faxon fir (Abies fargesii var. faxoniana), as a dominant tree species in the subalpine coniferous forest of Southwest China, has strict requirements regarding the temperature and humidity of the growing environment. Therefore, the dynamic and continuous monitoring of Faxon fir distribution is very important to protect this highly sensitive ecological environment. Here, we combined unmanned aerial vehicle (UAV) imagery and convolutional neural networks (CNNs) to identify Faxon fir and explored the identification capabilities of multispectral (five bands) and red-green-blue (RGB) imagery under different months. For a case study area in Wanglang Nature Reserve, Southwest China, we acquired monthly RGB and multispectral images on six occasions over the growing season. We found that the accuracy of RGB imagery varied considerably (the highest intersection over union (IoU), 83.72%, was in April and the lowest, 76.81%, was in June), while the accuracy of multispectral imagery was consistently high (IoU > 81%). In April and October, the accuracy of the RGB imagery was slightly higher than that of multispectral imagery, but for the other months, multispectral imagery was more accurate (IoU was nearly 6% higher than those of the RGB imagery for June). Adding vegetation indices (VIs) improved the accuracy of the RGB models during summer, but there was still a gap to the multispectral model. Hence, our results indicate that the optimized time of the year for identifying Faxon fir using UAV imagery is during the peak of the growing season when using a multispectral imagery. During the non-growing season, RGB imagery was no worse or even slightly better than multispectral imagery for Faxon fir identification. Our study can provide guidance for optimizing observation plans regarding data collection time and UAV loads and could further help enhance the utility of UAVs in forestry and ecological research.
Rapid urbanization and the rapid accumulation of multi-source and multi-temporal Earth observation data are creating an increasing demand for remote sensing models that can flexibly support fine-grained monitoring beyond fixed label taxonomies. Open-vocabulary remote sensing image semantic segmentation (OVRSIS) aims to segment text-specified categories beyond a fixed label space with vision–language foundation models. However, dense remote sensing scenes make pixel–text matching highly vulnerable to semantic confusion and misalignment, owing to extreme scale variation, thin structures, repetitive textures, and prompt sensitivity. To address these challenges, we propose UGDMoE, an uncertainty-guided mixture-of-experts framework for OVRSIS. First, we design a domain-specific MoE decoder with three geometrically specialized experts—for slender structures, mid-scale objects, and large-region context—routed by the alignment-risk cue . Second, we introduce a lightweight prompt–response estimation strategy that quantifies prediction dispersion across semantically equivalent prompts to derive in an annotation-free manner. Third, we develop prompt ensemble-based likelihood calibration (PELC), which takes the shared alignment-risk cue as input to calibrate prompt-specific logits before refinement. Finally, we design a lightweight uncertainty-aware structure refinement module that, guided by , selectively fuses early visual features with segmentation logits to restore boundary continuity and connectivity of thin structures. We conduct extensive experiments on eight OVRSIS benchmarks under cross-dataset evaluation protocols. Trained on DLRSD, it achieves 46.97 m-mIoU and 63.31 m-mACC, surpassing the strongest baseline by 0.76 and 0.62 points; trained on iSAID, it reaches 37.47 m-mIoU and 58.52 m-mACC, improving over the strongest competitor by 0.71 and 0.61 points. UGDMoE consistently achieves state-of-the-art performance and remains robust under training-source changes.