特征增强的Sparse Transformer目标跟踪算法

针对Transformer的自注意力机制计算量大、容易被背景分心,导致有效信息抓取不足,从而降低跟踪性能的问题,提出特征增强的Sparse Transformer目标跟踪算法。基于孪生网络骨干进行特征提取;特征增强模块利用多尺度特征图生成的上下文信息,增强目标局部特征;利用Sparse Transformer的最相关特性生成目标聚焦特征,并嵌入位置编码提升跟踪定位的精度。提出的跟踪模型以端到端的方式进行训练,在OTB100,VOT2018和LaSOT等5个数据集上进行了大量实验,实验结果表明所提算法取得了较好的跟踪性能,实时跟踪速度为34帧/s。

Characteristics and Influencing Factors of Microplastics in Snow in the Inner Mongolia Plateau, China

Microplastics(MPs;<5 mm)have become one of the most prominent global environmental pollution problems.MPs can spread to high altitudes through atmospheric transport and can be deposited by rainfall or snowfall,potentially threatening the structure and function of natural ecosystems.MPs in terrestrial and aquatic ecosystems alter the growth and functional characteristics of organisms.However,little attention has been given to the possible harm associated with MPs deposited in snow,particularly in the context of global climate warming.MPs collected from surface snow in the Inner Mongolia Plateau,China,were used for quantitative analysis and identification.The results showed that MPs were easily detected,and the related concentration was approximately(68±10)–(199±22)MPsL1 in snow samples.Fibers were the most common morphology,the polymer composition was largely varied,and the abundance and composition of MPs were linked to human activity to a great extent.High-throughput sequencing results showed that the composition and abundance of microorganisms also differed in snow samples from areas with different MP pollution characteristics,indicating a considerable difference in microbial functional diversity.MPs may have an interference effect on the individual growth and functional expression of microorganisms in snow.In addition,the results showed that functional living areas(e.g.,landfills and suburban areas)in cities play an important role in the properties of MPs.For instance,the highest abundance of MPs was found in thermal power plants,whereas the abundance of polymers per sample was significantly lower in the suburban area.The MP contaminants hidden in snow can alter microbial structure and function and are therefore a potential threat to ecosystem health.

A progress review of black carbon deposition on Arctic snow and ice and its impact on climate change

The rapid warming of the Arctic,accompanied by glacier and sea ice melt,has significant consequences for the Earth’s climate,ecosystems,and economy.Black carbon(BC)deposition on snow and ice can trigger a significant reduction in snow albedo and accelerate melting of snow and ice in the Arctic.By reviewing the published literatures over the past decades,this work provides an overview of the progress in both the measurement and modeling of BC deposition and its impact on Arctic climate change.In summary,the maximum value of BC deposition appears in the western Russian Arctic(26 ng·g^(–1)),and the minimum value appears in Greenland(3 ng·g^(–1)).BC records in the Arctic ice core already peaked in 1920s and 1970s,and shows a regional difference between Greenland and Canadian Arctic.The different temporal variations of Arctic BC ice core records in different regions are closely related to the large variability of BC emissions and transportation processes across the Arctic region.Model simulations usually underestimate the concentration of BC in snow and ice by 2–3 times,and cannot accurately reflect the seasonal and regional changes in BC deposition.Wet deposition is the main removal mechanism of BC in the Arctic,and observations show different seasonal variations in BC wet deposition in Ny-Ålesund and Barrow.This discrepancy may result from varying contributions of anthropogenic and biomass burning(BB)emissions,given the strong influence by BC from BB emissions at Barrow.Arctic BC deposition significantly influences regional climate change in the Arctic,increasing fire activities in the Arctic have made BB source of Arctic BC more crucial.On average,BC in Arctic snow and ice causes an increase of+0.17 W·m^(–2)in radiative forcing and 8 Gt·a^(–1)in runoff in Greenland.As stressed in the latest Arctic Monitoring and Assessment Programme report,reliable source information and long-term and high-resolution observations on Arctic BC deposition will be crucial for a more comprehensive understanding and a better mitigation strategy of Arctic BC.In the future,it is necessary to collect more observations on BC deposition and the corresponding physical processes(e.g.,snow/ice melting,surface energy balance)in the Arctic to provide reliable data for understanding and clarifying the mechanism of the climatic impacts of BC deposition on Arctic snow and ice.

Inter-comparison of field snow measurements using different instruments in Türkiye

Snow is important in Türkiye especially in the mountainous eastern areas where it may stay on the ground for more than half of the year.This region plays a vital role in feeding the water resources of the trans-boundary Euphrates-Tigris Basin,supporting crucial dams for water supply,irrigation and energy production.Thus,easy,frequent,correct and economical ways of measuring the snowpack is crucial.The snow properties at specific locations in the mountainous eastern regions over the two snow seasons(2018 and 2019)were studied by using different instruments and techniques,snow pit(box/cylinder/wedge cutter types),snow tube(Federal Sampler)and SnoTel(Snowpack Analyzer).The results point out a 1.7%-7.1%variation between different cutter type snow density measurements within snow pit analysis and the long-term utilized snow tube observations show a closer relation to box/cylinder type cutters.As for the continuous SnoTel observations,a variation of 2.4%-9.8%with various cutter types and a 5.9%difference regarding the snow tube density results are detected.These findings indicate a close range among different instruments,but it is the best when all three systems complement each other to characterize the snowpack effectively in the complex terrain since each has its own advantages.

Numerical Simulations of Snow Accumulation in the Bogie Region of a Train Considering Snow Particle Rotation

To investigate the influence of snow particle rotational motion on the accumulation of snow in the bogie region of high-speed trains,an Euler‒Lagrange numerical approach is adopted.The study examines the effects of snow particle diameter and train speed on the ensuing dynamics.It is shown that considering snow particle rotational motion causes significant deviation in the particle trajectories with respect to non-rotating particles.Such a deviation increases with larger snow particle diameters and higher train speeds.The snow accumulation on the overall surface of the bogie increases,and the amount of snow on the vibration reduction device varies greatly.In certain conditions,the amount of accumulated snow can increase by several orders of magnitudes.

A Two-Layer Encoding Learning Swarm Optimizer Based on Frequent Itemsets for Sparse Large-Scale Multi-Objective Optimization

Traditional large-scale multi-objective optimization algorithms(LSMOEAs)encounter difficulties when dealing with sparse large-scale multi-objective optimization problems(SLM-OPs)where most decision variables are zero.As a result,many algorithms use a two-layer encoding approach to optimize binary variable Mask and real variable Dec separately.Nevertheless,existing optimizers often focus on locating non-zero variable posi-tions to optimize the binary variables Mask.However,approxi-mating the sparse distribution of real Pareto optimal solutions does not necessarily mean that the objective function is optimized.In data mining,it is common to mine frequent itemsets appear-ing together in a dataset to reveal the correlation between data.Inspired by this,we propose a novel two-layer encoding learning swarm optimizer based on frequent itemsets(TELSO)to address these SLMOPs.TELSO mined the frequent terms of multiple particles with better target values to find mask combinations that can obtain better objective values for fast convergence.Experi-mental results on five real-world problems and eight benchmark sets demonstrate that TELSO outperforms existing state-of-the-art sparse large-scale multi-objective evolutionary algorithms(SLMOEAs)in terms of performance and convergence speed.

Spatiotemporal Changes of Snow Depth in Western Jilin,China from 1987 to 2018

Seasonal snow cover is a key global climate and hydrological system component drawing considerable attention due to glob-al warming conditions.However,the spatiotemporal snow cover patterns are challenging in western Jilin,China due to natural condi-tions and sparse observation.Hence,this study investigated the spatiotemporal patterns of snow cover using fine-resolution passive mi-crowave(PMW)snow depth(SD)data from 1987 to 2018,and revealed the potential influence of climate factors on SD variations.The results indicated that the interannual range of SD was between 2.90 cm and 9.60 cm during the snowy winter seasons and the annual mean SD showed a slightly increasing trend(P>0.05)at a rate of 0.009 cm/yr.In snowmelt periods,the snow cover contributed to an increase in volumetric soil water,and the change in SD was significantly affected by air temperature.The correlation between SD and air temperature was negative,while the correlation between SD and precipitation was positive during December and March.In March,the correlation coefficient exceeded 0.5 in Zhenlai,Da’an,Qianan,and Qianguo counties.However,the SD and precipitation were neg-atively correlated over western Jilin in October,and several subregions presented a negative correlation between SD and precipitation in November and April.

Enhancing Evolutionary Algorithms With Pattern Mining for Sparse Large-Scale Multi-Objective Optimization Problems

Sparse large-scale multi-objective optimization problems(SLMOPs)are common in science and engineering.However,the large-scale problem represents the high dimensionality of the decision space,requiring algorithms to traverse vast expanse with limited computational resources.Furthermore,in the context of sparse,most variables in Pareto optimal solutions are zero,making it difficult for algorithms to identify non-zero variables efficiently.This paper is dedicated to addressing the challenges posed by SLMOPs.To start,we introduce innovative objective functions customized to mine maximum and minimum candidate sets.This substantial enhancement dramatically improves the efficacy of frequent pattern mining.In this way,selecting candidate sets is no longer based on the quantity of nonzero variables they contain but on a higher proportion of nonzero variables within specific dimensions.Additionally,we unveil a novel approach to association rule mining,which delves into the intricate relationships between non-zero variables.This novel methodology aids in identifying sparse distributions that can potentially expedite reductions in the objective function value.We extensively tested our algorithm across eight benchmark problems and four real-world SLMOPs.The results demonstrate that our approach achieves competitive solutions across various challenges.

Impact of Climate Change on Mountaineering Activities in Haba Snow Mountain Reserve and the Development of Surrounding Communities

In the context of global climate change, this study reviews and discusses the three aspects of ecology, economic development of surrounding communities, ecological balance and snow mountain activities in the Haba Snow Mountain Reserve through literature collation and research. 1) The Hengduan Mountain Plate of Haba Snow Mountain is affected by the high altitude temperate monsoon and is sensitive to climate change. There has been continuous glacier melting and snow line fluctuations. Although there is no forest line movement, the vegetation at the junction of the forest line has increased. 2) Human activities in the Haba Snow Mountain Reserve have shown an active trend, and the Biomass in various ecosystems in the region is inversely correlated. 3) Climate change will have a negative impact on landscape attraction and tourism safety in snowy mountain areas. 4) Haba Snow Mountain Reserve needs more perfect biological species statistical research and dynamic vegetation research to support the establishment of a perfect ecological protection strategy and ecological early warning in the region. 5) As the frequency of tourist activities in the Haba Protected Area increases, corresponding environmental protection signage, garbage cleaning methods, and tourist education have not been synchronizedly improved.

Changes in Spring Snow Cover over the Eastern and Western Tibetan Plateau and Their Associated Mechanism

The spring snow cover(SC)over the western Tibetan Plateau(TP)(TPSC)(W_TPSC)and eastern TPSC(E_TPSC)have displayed remarkable decreasing and increasing trends,respectively,during 1985–2020.The current work investigates the possible mechanisms accounting for these distinct TPSC changes.Our results indicate that the decrease in W_TPSC is primarily attributed to rising temperatures,while the increase in E_TPSC is closely linked to enhanced precipitation.Local circulation analysis shows that the essential system responsible for the TPSC changes is a significant anticyclonic system centered over the northwestern TP.The anomalous descending motion and adiabatic heating linked to this anticyclone leads to warmer temperatures and consequent snowmelt over the western TP.Conversely,anomalous easterly winds along the southern flank of this anticyclone serve to transport additional moisture from the North Pacific,leading to an increase in snowfall over the eastern TP.Further analysis reveals that the anomalous anticyclone is associated with an atmospheric wave pattern that originates from upstream regions.Springtime warming of the subtropical North Atlantic(NA)sea surface temperature(SST)induces an atmospheric pattern resembling a wave train that travels eastward across the Eurasian continent before reaching the TP.Furthermore,the decline in winter sea ice(SIC)over the Barents Sea exerts a persistent warming influence on the atmosphere,inducing an anomalous atmospheric circulation that propagates southeastward and strengthens the northwest TP anticyclone in spring.Additionally,an enhancement of subtropical stationary waves has resulted in significant increases in easterly moisture fluxes over the coastal areas of East Asia,which further promotes more snowfall over eastern TP.

Study on Image Recognition Algorithm for Residual Snow and Ice on Photovoltaic Modules

The accumulation of snow and ice on PV modules can have a detrimental impact on power generation,leading to reduced efficiency for prolonged periods.Thus,it becomes imperative to develop an intelligent system capable of accurately assessing the extent of snow and ice coverage on PV modules.To address this issue,the article proposes an innovative ice and snow recognition algorithm that effectively segments the ice and snow areas within the collected images.Furthermore,the algorithm incorporates an analysis of the morphological characteristics of ice and snow coverage on PV modules,allowing for the establishment of a residual ice and snow recognition process.This process utilizes both the external ellipse method and the pixel statistical method to refine the identification process.The effectiveness of the proposed algorithm is validated through extensive testing with isolated and continuous snow area pictures.The results demonstrate the algorithm’s accuracy and reliability in identifying and quantifying residual snow and ice on PV modules.In conclusion,this research presents a valuable method for accurately detecting and quantifying snow and ice coverage on PV modules.This breakthrough is of utmost significance for PV power plants,as it enables predictions of power generation efficiency and facilitates efficient PV maintenance during the challenging winter conditions characterized by snow and ice.By proactively managing snow and ice coverage,PV power plants can optimize energy production and minimize downtime,ensuring a sustainable and reliable renewable energy supply.

THE STABLE RECONSTRUCTION OF STRONGLY-DECAYING BLOCK SPARSE SIGNALS

In this paper,we reconstruct strongly-decaying block sparse signals by the block generalized orthogonal matching pursuit(BgOMP)algorithm in the l2-bounded noise case.Under some restraints on the minimum magnitude of the nonzero elements of the strongly-decaying block sparse signal,if the sensing matrix satisfies the the block restricted isometry property(block-RIP),then arbitrary strongly-decaying block sparse signals can be accurately and steadily reconstructed by the BgOMP algorithm in iterations.Furthermore,we conjecture that this condition is sharp.

Classifying moisture sources associated with snowfall in the mountains of Lesotho

An average of eight snowfall events occur each year in the eastern Lesotho Highlands.These snowfall events are typically associated with cut-off low(CoLs)systems and mid-latitude cyclones.However,the moisture sources of the snowfall are unclassified and unclear.The Hybrid Single-Particle Lagrangian Integrated Trajectory(HYSPLIT)model,an air mass back trajectory model,has been used to evaluate moisture source waters locally in southern Africa and internationally in China and Europe.This study uses HYSPLIT to determine the source moisture of snow in Lesotho.A list of all 82 snowfall events in Lesotho spanning 2017 to 2022 was compiled using the Snow Report SA Instagram page,including the date and location of snowfall.A 72-hour back trajectory for each snowfall event was initiated for both Afriski and the whole of Lesotho.This amounts to models of moisture source trajectories for 28 and 82 snowfall days,respectively.These air mass pathways are classified according to their frequency per snowfall event,per month in the snow season,per year and for the full period.From this,associated moisture source regions and dominant air mass trajectories were identified.This study reports that the air mass trajectories associated with Afriski and Lesotho as a whole are very similar.The most common pathway of air mass trajectories transporting snow-bearing moisture to Lesotho was an inland trajectory from the northern regions of southern Africa.This pathway makes up 16.6%of all trajectories reported and is associated with the Angola Low,the Congo Air Boundary and the St.Helena High Pressure.

Sparse Reconstructive Evidential Clustering for Multi-View Data

Although many multi-view clustering(MVC) algorithms with acceptable performances have been presented, to the best of our knowledge, nearly all of them need to be fed with the correct number of clusters. In addition, these existing algorithms create only the hard and fuzzy partitions for multi-view objects,which are often located in highly-overlapping areas of multi-view feature space. The adoption of hard and fuzzy partition ignores the ambiguity and uncertainty in the assignment of objects, likely leading to performance degradation. To address these issues, we propose a novel sparse reconstructive multi-view evidential clustering algorithm(SRMVEC). Based on a sparse reconstructive procedure, SRMVEC learns a shared affinity matrix across views, and maps multi-view objects to a 2-dimensional humanreadable chart by calculating 2 newly defined mathematical metrics for each object. From this chart, users can detect the number of clusters and select several objects existing in the dataset as cluster centers. Then, SRMVEC derives a credal partition under the framework of evidence theory, improving the fault tolerance of clustering. Ablation studies show the benefits of adopting the sparse reconstructive procedure and evidence theory. Besides,SRMVEC delivers effectiveness on benchmark datasets by outperforming some state-of-the-art methods.

Cambricon-QR:a sparse and bitwise reproducible quantized training accelerator

Quantized training has been proven to be a prominent method to achieve deep neural network training under limited computational resources.It uses low bit-width arithmetics with a proper scaling factor to achieve negligible accuracy loss.Cambricon-Q is the ASIC design proposed to efficiently support quantized training,and achieves significant performance improvement.However,there are still two caveats in the design.First,Cambricon-Q with different hardware specifications may lead to different numerical errors,resulting in non-reproducible behaviors which may become a major concern in critical applications.Second,Cambricon-Q cannot leverage data sparsity,where considerable cycles could still be squeezed out.To address the caveats,the acceleration core of Cambricon-Q is redesigned to support fine-grained irregular data processing.The new design not only enables acceleration on sparse data,but also enables performing local dynamic quantization by contiguous value ranges(which is hardware independent),instead of contiguous addresses(which is dependent on hardware factors).Experimental results show that the accuracy loss of the method still keeps negligible,and the accelerator achieves 1.61×performance improvement over Cambricon-Q,with about 10%energy increase.

A Novel Clutter Suppression Algorithm for Low-Slow-Small Targets Detecting Based on Sparse Adaptive Filtering

Passive detection of low-slow-small(LSS)targets is easily interfered by direct signal and multipath clutter,and the traditional clutter suppression method has the contradiction between step size and convergence rate.In this paper,a frequency domain clutter suppression algorithm based on sparse adaptive filtering is proposed.The pulse compression operation between the error signal and the input reference signal is added to the cost function as a sparsity constraint,and the criterion for filter weight updating is improved to obtain a purer echo signal.At the same time,the step size and penalty factor are brought into the adaptive iteration process,and the input data is used to drive the adaptive changes of parameters such as step size.The proposed algorithm has a small amount of calculation,which improves the robustness to parameters such as step size,reduces the weight error of the filter and has a good clutter suppression performance.

Improved Variable Forgetting Factor Proportionate RLS Algorithm with Sparse Penalty and Fast Implementation Using DCD Iterations

The proportionate recursive least squares(PRLS)algorithm has shown faster convergence and better performance than both proportionate updating(PU)mechanism based least mean squares(LMS)algorithms and RLS algorithms with a sparse regularization term.In this paper,we propose a variable forgetting factor(VFF)PRLS algorithm with a sparse penalty,e.g.,l_(1)-norm,for sparse identification.To reduce the computation complexity of the proposed algorithm,a fast implementation method based on dichotomous coordinate descent(DCD)algorithm is also derived.Simulation results indicate superior performance of the proposed algorithm.

SRMD:Sparse Random Mode Decomposition

Signal decomposition and multiscale signal analysis provide many useful tools for timefrequency analysis.We proposed a random feature method for analyzing time-series data by constructing a sparse approximation to the spectrogram.The randomization is both in the time window locations and the frequency sampling,which lowers the overall sampling and computational cost.The sparsification of the spectrogram leads to a sharp separation between time-frequency clusters which makes it easier to identify intrinsic modes,and thus leads to a new data-driven mode decomposition.The applications include signal representation,outlier removal,and mode decomposition.On benchmark tests,we show that our approach outperforms other state-of-the-art decomposition methods.

A World of Ice and Snow Being also Invaluable Asset:A Case Study of the"Snow Village"

Based on field visit and interview,the current situation of snow village in China is summarized from four aspects:core scenic spots in snow village,skiing industry in snow village,film and television industry in snow village,and ice and snow agritainment.The investigation found that there are still significant problems in homogenization,scenic area infrastructure,and government regulation in snow village.Targeted solutions are proposed from four aspects:tapping internal advantages,strengthening top-level design and infrastructure construction,promoting tourism industry upgrading,and collaborating to innovate the ice and snow tourism supply chain,in order to further promote the economic development of snow village.

Cross-calibration between MWRI and AMSR2 to improve consistency of snow depth products

The Microwave Radiation Imager (MWRI), boarded on the FY-3 series satellites: FY-3B, FY-3C, and FY-3D, is thefirst satellite-based microwave radiometer in China, commencing passive microwave brightness temperature dataacquisition since 2010. The Advanced Microwave Scanning Radiometer 2 (AMSR2) boarded on the Global ChangeObservation Mission 1st-Water (GCOM-W1), has been operational since 2012. Despite the FY-3 series satellitesare equipped with the same MWRI and all MWRIs sharing comparable parameters and configurations as AMSR2,disparities in observation times and satellite platforms result in inconsistencies in the data obtained by differentsatellites, which further impacting the consistency of retrieved geophysical parameters. To improve the consistency of brightness temperatures from FY-3B, FY-3C, FY-3D/MWRI, and GCOM-W1/AMSR2, cross-calibrationswere conducted among brightness temperatures at ten-channel from above four platforms. The consistency ofderived snow depth from MWRIs and AMSR2 in China before and after the calibration were also analyzed. Theresults show that the correlation coefficients of brightness temperatures at all channels between sensors exceed0.98. After cross-calibration, the RMSEs and biases of brightness temperatures at all frequencies and snow depthin China derived from them reduce to varying degrees. The consistencies in both brightness temperatures andsnow depth of FY-3B/MWRI, FY-3D/MWRI, and AMSR2 are higher than those of FY-3C and others. These findingsadvocate for the utilization of cross-calibrated brightness temperatures from FY-3B/MWRI, FY-3D/MWRI, andAMSR2, which share similar satellite overpass time, to derived a long-term snow depth dataset.