Search Results (11)

In recent years, cross-domain few-shot learning (CDFSL) has demonstrated remarkable performance in hyperspectral image classification (HSIC), partially alleviating the distribution shift problem. However, most domain adaptation methods rely on similarity metrics to establish cross-domain class matching, making it difficult to simultaneously account for intra-class sample size variations and inherent inter-class differences. To address this problem, existing studies have introduced a class weighting mechanism within the prototype network framework, determining class weights by calculating inter-sample similarity through distance metrics. However, this method suffers from a dual limitation: susceptibility to noise interference and insufficient capacity to capture global class variations, which may lead to distorted weight allocation and consequently result in alignment bias. To solve these issues, we propose a novel class-discrepancy dynamic weighting-based cross-domain FSL (CDDW-CFSL) framework. It integrates three key components: (1) the class-weighted domain adaptation (CWDA) method dynamically measures cross-domain distribution shifts using global class mean discrepancies. It employs discrepancy-sensitive weighting to strengthen the alignment of critical categories, enabling accurate domain adaptation while maintaining feature topology; (2) the class mean refinement (CMR) method incorporates class covariance distance to compute distribution discrepancies between support set samples and class prototypes, enabling the precise capture of cross-domain feature internal structures; (3) a novel multi-dimensional feature extractor that captures both local spatial details and continuous spectral characteristics simultaneously, facilitating deep cross-dimensional feature fusion. The results in three publicly available HSIC datasets show the effectiveness of the CDDW-CFSL. Full article

Power line segmentation plays a critical role in ensuring the safety of transmission line UAV inspection flights. To address the challenges of small target scale, complex backgrounds, and excessive model parameters in existing deep learning-based power line segmentation algorithms, this paper introduces RGS-UNet, a lightweight segmentation model integrating a residual-like cross-layer feature fusion module. First, ResNet18 is adopted to reconstruct a UNet backbone network as an encoder module to enhance the network’s feature extraction capability for small targets. Second, ordinary convolution in the residual block of ResNet18 is optimized by introducing the Ghost Module, which significantly reduces the computational load of the model’s backbone network. Third, a residual-like addition method is designed to embed the SIMAM attention mechanism module into both encoder and decoder stages, which improves the model’s ability to extract power lines from complex backgrounds. Finally, the Mish activation function is applied in deep convolutional layers to maintain feature extraction accuracy and mitigate overfitting. Experimental results demonstrate that compared with classical UNet, the optimized network achieves 2.05% and 2.58% improvements in F1-Score and IoU, respectively, while reducing the parameter count to 57.25% of the original model. The algorithm achieves better performance improvements in both accuracy and lightweighting, making it suitable for edge-side deployment. Full article

Rainfall significantly affects soil respiration rates by altering microbial activity and organic matter decomposition. In karst regions, it also impacts carbonate dissolution and precipitation, further influencing soil CO₂ flux. Investigating the mechanism of rainfall’s impact on soil respiration is essential for accurately evaluating and predicting changes in terrestrial ecosystems. However, our understanding of the interaction between rainfall and soil respiration in the extensive karst ecosystems of southwestern China remains limited. This study conducted field-based simulated rainfall experiments to examine variations in soil respiration rates and elucidate the associated control mechanisms through stable carbon isotope composition analysis. Simulated rainfall significantly increased the CO₂ release via soil respiration. We observed significant differences in the δ¹³C value of soil-respired CO₂ before and after simulated rainfall. Following the rain, the δ¹³C of soil-respired CO₂ was enriched compared to that before the rain. Through isotope data analysis, we found that the increased soil CO₂ emissions were primarily driven by heterotrophic respiration, likely stimulated via changes in soil moisture, affecting microbial growth conditions. Furthermore, the variation in soil moisture affected carbonate dissolution and precipitation, potentially increasing the soil CO₂ release after rainfall. In conclusion, these findings expand our understanding of rainfall’s effects on soil respiration in the native karst forests of southwestern China, contributing to the prediction of carbon cycling processes in such ecosystems. The data from this study have significant implications for addressing the release of greenhouse gases in efforts to combat climate change. Full article

We propose a simple multifunctional terahertz absorber based on the simulation. The device consists of a gold layer, a SiO₂ dielectric layer, and a VO₂ top layer. The modulation mechanism of this device is to utilize the thermally induced phase transition characteristics of vanadium dioxide material. The simulation results show that when the temperature is 312 K, the device has the effect of complete reflection of terahertz waves. When the temperature is 345 K, the device has almost perfect absorption of terahertz wave in the range of 4.7–9.7 THz, and the spectral absorptivity is modulated in the range of 0~0.999. The electric field conditions at different temperatures were plotted to further explain the reasons for the performance transition of the device. The terahertz device was explained using impedance matching theory. In addition, the influence of different structural parameters on absorption rate was studied, providing reference for practical applications. At the same time, the device is polarization-insensitive and insensitive to the incident angle. When the incident angle changes from 0°to 45°, the device still has a stable absorption effect. The device has great application prospects in terahertz stealth, modulation, and other fields and provides ideas for the design of related devices. Full article

Post-fracturing shut-in, as an important means of improving the energy efficiency of fracturing fluid, has been widely used in the development process of unconventional reservoirs. The determination of the shut-in duration is key to the effectiveness of shut-in measures. However, the distribution characteristics of the fracturing fluid during the post-fracturing shut-in period in unconventional reservoirs, such as the Chang 7 reservoir, were not clear, and the shut-in duration needed further optimization. Therefore, this paper employed low-field nuclear magnetic resonance (NMR) technology to study the distribution characteristics of the fracturing fluid during the post-fracturing shut-in period in unconventional reservoirs and optimized the shut-in duration. The study showed that the Chang 7 reservoir had a complex pore structure and relatively low porosity and permeability. During the shut-in process, the filtrate was distributed in pore throats with radii ranging from 0.0012 μm to 0.025 μm. Pore throats with radii ranging from 0.003 μm to 0.07 μm acted as dynamic pore throats in the process of imbibition displacement. When the shut-in duration for the Chang 7 segment was 7 days, the growth rate of the retained volume of fracturing fluid filtrate was the highest. When the shut-in duration was 10 days, there was no oil displacement in the medium and large pores, and the retained volume of filtrate was lower than that at 7 days shut-in, indicating that an optimal shut-in duration would be 7 days. This study can provide theoretical and technical support for the development of unconventional reservoirs. Full article

Salticidae is the most species-rich family of spiders with diverse morphology, ecology and behavior. However, the characteristics of the mitogenomes within this group are poorly understood with relatively few well-characterized complete mitochondrial genomes. In this study, we provide completely annotated mitogenomes for Corythalia opima and Parabathippus shelfordi, which represent the first complete mitogenomes of the tribe Euophryini of Salticidae. The features and characteristics of the mitochondrial genomes are elucidated for Salticidae by thoroughly comparing the known well-characterized mitogenomes. The gene rearrangement between trnL2 and trnN was found in two jumping spider species, Corythalia opima and Heliophanus lineiventris Simon, 1868. Additionally, the rearrangement of nad1 to between trnE and trnF found in Asemonea sichuanensis Song & Chai, 1992 is the first protein-coding gene rearrangement in Salticidae, which may have an important phylogenetic implication for the family. Tandem repeats of various copy numbers and lengths were discovered in three jumping spider species. The codon usage analyses showed that the evolution of codon usage bias in salticid mitogenomes was affected by both selection and mutational pressure, but selection may have played a more important role. The phylogenetic analyses provided insight into the taxonomy of Colopsus longipalpis (Żabka, 1985). The data presented in this study will improve our understanding of the evolution of mitochondrial genomes within Salticidae. Full article

The key component of electrochemical advanced oxidation technology are high-efficiency anodes, and highly efficient and simple-to-prepare materials have generated a lot of interest. In this study, novel self-supported Ti³⁺-doped titanium dioxide nanotube arrays (R-TNTs) anodes were successfully prepared by a two-step anodic oxidation and straightforward electrochemical reduction technique. The electrochemical reduction self-doping treatment produced more Ti³⁺ sites with stronger absorption in the UV-vis region, a band gap reduction from 2.86 to 2.48 ev, and a significant increase in electron transport rate. The electrochemical degradation effect of R-TNTs electrode on chloramphenicol (CAP) simulated wastewater was investigated. At pH = 5, current density of 8 mA cm⁻², electrolyte concentration of 0.1 M sodium sulfate (Na₂SO₄), initial CAP concentration of 10 mg L⁻¹, CAP degradation efficiency exceeded 95% after 40 min. In addition, molecular probe experiments and electron paramagnetic resonance (EPR) tests revealed that the active species were mainly •OH and SO₄⁻, among which •OH played a major role. The CAP degradation intermediates were discovered using high-performance liquid chromatography-mass spectrometry (HPLC-MS), and three possible degradation mechanisms were postulated. In cycling experiments, the R-TNTs anode demonstrated good stability. The R-TNTs prepared in this paper were an anode electrocatalytic material with high catalytic activity and stability, which could provide a new approach for the preparation of electrochemical anode materials for difficult-to-degrade organic compounds. Full article

In the far-field wake of a ship, the intensity of the scattering of bubbles is relatively weak. In addition, the wake is relatively thin, and the hole phenomenon is prominent. Thus, it is difficult to detect the wake at a long distance. On this basis, this paper studies a broadband 1–3 high-performance composite transceiver sub-array for the improved detection of a ship’s far wake flow field. The content includes frequency characteristics, transmission performance, power tolerance, the beam width of the transmitting array, and the frequency characteristics, reception performance, and beam width of the receiving array. The frequency bandwidth of the transmission array developed in this paper can reach a value of 180 kHz (the center frequency is 390 kHz). The maximum sound source level can reach a value of 228 dB. In the same frequency band, the sensitivity of the receiving array can reach a value of 184 dB, and the fluctuation is less than 5 dB. Compared with the narrowband 1–3 composite array of the same size, the acoustic performance of this sub-array has obvious advantages. Finally, to improve the effective sound path (before the first interface reflection of the sound wave) of the emitted sound wave in the ship’s far-field wake, combined with the speed of the moving carrier and the wide-band detection method of the ship’s wake, the configuration method of the detection array for the width and direction of the ship’s far-field wake is proposed. The results of this research have an important reference value for the research on broadband 1–3 high-performance composite arrays and their application in the far-field wake detection of ships. Full article

With the rapid development of photo-communication technologies, avalanche photodiode (APD) will play an increasingly important role in the future due to its high quantum efficiency, low power consumption, and small size. The monolithic integration of optical components and signal processing electronics on silicon substrate chips is crucial to driving cost reduction and performance improvement; thus, the technical research on InGaAs/Si APD is of great significance. This work is the first to demonstrate the use of a photon-trapping (PT) structure to improve the performance of the InGaAs/Si APD based on an SOI substrate, which exhibits very high absorption efficiency at 1310 nm wavelength while the thickness of the absorption layer is kept at 800 nm. Based on the optical and electrical simulations, an optimized InGaAs/Si PT-APD is proposed, which exhibits a better performance and a higher responsivity compared to the original InGaAs/Si APD. Full article

Seed storability, defined as the ability to remain alive during storage, is an important agronomic and physiological characteristic, but the underlying genetic mechanism remains largely unclear. Here, we report quantitative trait loci (QTLs) analyses for seed storability using a high-density single nucleotide polymorphism linkage map in the backcross recombinant inbred lines that was derived from a cross of a japonica cultivar, Nipponbare, and an indica cultivar, 9311. Seven putative QTLs were identified for seed storability under natural storage, each explaining 3.6–9.0% of the phenotypic variation in this population. Among these QTLs, qSS1 with the 9311 alleles promoting seed storability was further validated in near-isogenic line and its derived-F₂ population. The other locus (qSS3.1) for seed storability colocalized with a locus for germination ability under hydrogen peroxide, which is recognized as an oxidant molecule that causes lipid damage. Transgenic experiments validated that a candidate gene (OsFAH2) resides the qSS3.1 region controlling seed storability and antioxidant capability. Overexpression of OsFAH2 that encodes a fatty acid hydroxylase reduced lipid preoxidation and increased seed storability. These findings provide new insights into the genetic and physiological bases of seed storability and will be useful for the improvement of seed storability in rice. Full article

Herders’ living strategy is a function of the capitals at their disposal which also serve as a buffering mechanism when shocks arise. An insight into the connection between livelihood strategies and capitals owned by herders provides guidance to recognize their living situation. This study evaluated the different livelihood capitals of herders across the five ecological types (meadow, typical, desert, sandy, and desert steppe) in Inner Mongolia region of China, using the sustainable livelihood framework approach. An evaluation index was developed and used to investigate how the livelihood capitals of herders affects preferential selection of livelihood strategies using multinomial logit model. Results indicate that: (1) The stocks of human and social capitals were higher while those for natural, physical, and financial capitals were lower. (2) There were significant regional differences in the livelihood capital stock of herders’ families with zonal horizontal decrease from east to west. (3) Natural capitals affects the preferential selection of livelihood strategies by herders positively implying that possession of more natural capitals by herders leads to selection of livelihood strategies that are devoid of pastoral production; the preferred livelihood strategy of herders was significantly negatively affected by physical and financial capitals, an indication that, when herders possess more physical and financial capitals, they tend to choose livelihood strategies that involve pastoral production. The living strategy of herders was not affected by human and social capitals. (4) Production of rented pasture capital index affected the preferential selection of livelihood strategies by herders positively while cash income capital index had negative influence on how pastoralists select their livelihood strategies. In conclusion, the total livelihood capital of herders in Inner Mongolia is low, and there is perceived benefit in the differentiation of herders families into petty-herders and non-grazing families from the perception of natural resource management and sustainability. This requires income diversification programs such as capacity building and business education that will aid the smooth transition of households to these less resource exploiting livelihood strategies. Full article