Road surface extraction from remote sensing images using deep learning methods has achieved good performance, while most of the existing methods are based on fully supervised learning, which requires ...a large amount of training data with laborious per-pixel annotation. In this article, we propose a scribble-based weakly supervised road surface extraction method named ScRoadExtractor, which learns from easily accessible scribbles such as centerlines instead of densely annotated road surface ground truths. To propagate semantic information from sparse scribbles to unlabeled pixels, we introduce a road label propagation algorithm, which considers both the buffer-based properties of road networks and the color and spatial information of super-pixels, to produce a proposal mask with categories road, nonroad, and unknown. The proposal mask, along with the auxiliary boundary prior information detected from images, is utilized to train a dual-branch encoder-decoder network which we designed for precise road surface segmentation. We perform experiments on three diverse road data sets that are comprised of high-resolution remote sensing satellite and aerial images across the world. The results demonstrate that ScRoadExtractor exceeds the classic scribble-supervised segmentation method by 20% for the intersection over union (IoU) indicator and outperforms the state-of-the-art scribble-based weakly supervised methods at least 4%.
Accurate and up-to-date road maps are of great importance in a wide range of applications. Unfortunately, automatic road extraction from high-resolution remote sensing images remains challenging due ...to the occlusion of trees and buildings, discriminability of roads, and complex backgrounds. To address these problems, especially road connectivity and completeness, in this article, we introduce a novel deep learning-based multistage framework to accurately extract the road surface and road centerline simultaneously. Our framework consists of three steps: boosting segmentation, multiple starting points tracing, and fusion. The initial road surface segmentation is achieved with a fully convolutional network (FCN), after which another lighter FCN is applied several times to boost the accuracy and connectivity of the initial segmentation. In the multiple starting points tracing step, the starting points are automatically generated by extracting the road intersections of the segmentation results, which then are utilized to track consecutive and complete road networks through an iterative search strategy embedded in a convolutional neural network (CNN). The fusion step aggregates the semantic and topological information of road networks by combining the segmentation and tracing results to produce the final and refined road segmentation and centerline maps. We evaluated our method utilizing three data sets covering various road situations in more than 40 cities around the world. The results demonstrate the superior performance of our proposed framework. Specifically, our method's performance exceeded the other methods by 7% and 40% for the connectivity indicator for road surface segmentation and for the completeness indicator for centerline extraction, respectively.
Metasurfaces as artificially nanostructured interfaces hold significant potential for multi-functionality, which may play a pivotal role in the next-generation compact nano-devices. The majority of ...multi-tasked metasurfaces encode or encrypt multi-information either into the carefully tailored metasurfaces or in pre-set complex incident beam arrays. Here, we propose and demonstrate a multi-momentum transformation metasurface (i.e., meta-transformer), by fully synergizing intrinsic properties of light, e.g., orbital angular momentum (OAM) and linear momentum (LM), with a fixed phase profile imparted by a metasurface. The OAM meta-transformer reconstructs different topologically charged beams into on-axis distinct patterns in the same plane. The LM meta-transformer converts red, green and blue illuminations to the on-axis images of "R", "G" and "B" as well as vivid color holograms, respectively. Thanks to the infinite states of light-metasurface phase combinations, such ultra-compact meta-transformer has potential in information storage, nanophotonics, optical integration and optical encryption.
This paper deals with quantitative domain theory via fuzzy sets. It examines the continuity of fuzzy directed complete posets (dcpos for short) based on complete residuated lattices. First, we show ...that a fuzzy partial order in the sense of Fan and Zhang and an
L-order in the sense of Bělohlávek are equivalent to each other. Then we redefine the concepts of fuzzy directed subsets and (continuous) fuzzy dcpos. We also define and study fuzzy Galois connections on fuzzy posets. We investigate some properties of (continuous) fuzzy dcpos. We show that a fuzzy dcpo is continuous if and only if the fuzzy-double-downward-arrow-operator has a right adjoint. We define fuzzy auxiliary relations on fuzzy posets and approximating fuzzy auxiliary relations on fuzzy dcpos. We show that a fuzzy dcpo is continuous if and only if the fuzzy way-below relation is the smallest approximating fuzzy auxiliary relation.
Jumping spiders (Salticidae) rely on accurate depth perception for predation and navigation. They accomplish depth perception, despite their tiny brains, by using specialized optics. Each principal ...eye includes a multitiered retina that simultaneously receives multiple images with different amounts of defocus, and from these images, distance is decoded with relatively little computation. We introduce a compact depth sensor that is inspired by the jumping spider. It combines metalens optics, which modifies the phase of incident light at a subwavelength scale, with efficient computations to measure depth from image defocus. Instead of using a multitiered retina to transduce multiple simultaneous images, the sensor uses a metalens to split the light that passes through an aperture and concurrently form 2 differently defocused images at distinct regions of a single planar photosensor. We demonstrate a system that deploys a 3-mm-diameter metalens to measure depth over a 10-cm distance range, using fewer than 700 floating point operations per output pixel. Compared with previous passive depth sensors, our metalens depth sensor is compact, single-shot, and requires a small amount of computation. This integration of nanophotonics and efficient computation brings artificial depth sensing closer to being feasible on millimeter-scale, microwatts platforms such as microrobots and microsensor networks.
Allylic amines are versatile building blocks in organic synthesis and exist in bioactive compounds, but their synthesis via hydroaminoalkylation of alkynes with amines has been a formidable ...challenge. Here, we report a late transition metal Ni-catalyzed hydroaminoalkylation of alkynes with N-sulfonyl amines, providing a series of allylic amines in up to 94% yield. Double ligands of N-heterocyclic carbene (IPr) and tricyclohexylphosphine (PCy
) effectively promote the reaction.
Abstract
Non-Maxwellian
κ
electron energy distributions (EEDs) have been proposed in recent years to resolve the so-called “electron temperature and abundance discrepancy problem” in the study of ...planetary nebulae (PNs). Thus the need to develop diagnostic tools to determine from observations the EED of PNs is raised. Arising from high-energy levels, the ultraviolet (UV) emission lines from PNs present intensities that depend sensitively on the high-energy tail of the EED. In this work, we investigate the feasibility of using the C
ii
λ
2326/C
ii
λ
1335 intensity ratios as a diagnostic of the deviation of the EED from the Maxwellian distribution (as represented by the
κ
index). We use a Maxwellian decomposition approach to derive the theoretical
κ
-EED-based collisionally excited coefficients of C
ii
, and then compute the C
ii
UV intensity ratio as a function of the
κ
index. We analyze the archival spectra acquired by the International Ultraviolet Explorer and measure the intensities of C
ii
UV lines from 12 PNs. By comparing the observed line ratios and the theoretical predictions, we can infer their
κ
values. With the Maxwellian-EED hypothesis, the observed C
ii
λ
2326/C
ii
λ
1335 ratios are found to be generally lower than those predicted from the observed optical spectra. This discrepancy can be explained in terms of the
κ
EED. Our results show that the
κ
values inferred range from 15 to infinity, suggesting a mild or modest deviation from the Maxwellian distribution. However, the
κ
-distributed electrons are unlikely to exist throughout the whole nebulae. A toy model shows that if just about 1%–5% of the free electrons in a PN had a
κ
EED as small as
κ
= 3, it would be sufficient to account for the observations.
A brief review of the recent advances in kerosene-fueled supersonic combustion modeling is present by comparing the fuels, reviewing the kinetic mechanisms, and introducing recent modeling results. ...The advantages and disadvantages of hydrogen and kerosene for the scramjet combustor are compared to show that kerosene is a more viable fuel option for a Mach number range of 4
–
8. However, detailed kinetic mechanisms for kerosene, which usually contain thousands of elementary reactions, must be significantly reduced for use in modeling. As of this writing, the smallest skeletal kerosene mechanism has only 19 species and 53 reversible reactions. In contrast to pioneer models based on global chemistry, the current kerosene-fueled supersonic combustion models based on reduced/skeletal chemistry are classified as second-stage. The influence of kinetic mechanisms, global equivalence ratios, inlet Mach number, geometric shape, and domain symmetry are reviewed based on high-fidelity models and available measurements. With the advances in computational technology, models with accurate descriptions of both flow and chemistry are becoming a promising, indispensable approach for the study of supersonic combustion.