Hyperspectral remote sensing image(HSI)fusion with multispectral remote sensing images(MSI)improves data resolution.However,current fusion algorithms focus on local information and overlook long-range dependencies.The...Hyperspectral remote sensing image(HSI)fusion with multispectral remote sensing images(MSI)improves data resolution.However,current fusion algorithms focus on local information and overlook long-range dependencies.The parameter of network tuning prioritizes global optimization,neglecting spatial and spectral constraints,and limiting spatial and spectral reconstruction capabilities.This study introduces SwinGAN,a fusion network combining Swin Transformer,CNN,and GAN architectures.SwinGAN’s generator employs a detail injection framework to separately extract HSI and MSI features,fusing them to generate spatial residuals.These residuals are injected into the supersampled HSI to produce thefinal image,while a pure CNN architecture acts as the discriminator,enhancing the fusion quality.Additionally,we introduce a new adaptive loss function that improves image fusion accuracy.The loss function uses L1 loss as the content loss,and spatial and spectral gradient loss functions are introduced to improve the spatial representation and spectralfidelity of the fused images.Our experimental results on several datasets demonstrate that SwinGAN outperforms current popular algorithms in both spatial and spectral reconstruction capabilities.The ablation experiments also demonstrate the rationality of the various components of the proposed loss function.展开更多
Based on field runoff plots observation and sample analysis, the effect of slope gradient on soil organic carbon loss was studied under natural rainfall conditions in loess hilly region. The results showed that with s...Based on field runoff plots observation and sample analysis, the effect of slope gradient on soil organic carbon loss was studied under natural rainfall conditions in loess hilly region. The results showed that with slope gradient increasing （from10° to 30°）, the changing trend of soil erosion intensity （A） was A20°〉A30°〉A15°〉A10°〉A25°, suggesting that slope gradient between 20° and 25° was a critical degree ranger to exist. Soil organic carbon loss was mainly influenced by soil erosion intensity, appearing the same trend as did soil erosion intensity with increasing slope gradient. Soil erosion results in organic carbon enrichment in sediment. Average enrichment ratios of five runoff plots varied from 2.27 to 3.74, and decreased with increasing erosion intensity and slope gradient. The decrease of surface runoff and soil erosion is the key to reduce soil organic carbon loss.展开更多
The fracture gradient is a critical parameter for drilling mud weight design in the energy industry. A new method in fracture gradient prediction is proposed based on analyzing worldwide leak-off test(LOT) data in off...The fracture gradient is a critical parameter for drilling mud weight design in the energy industry. A new method in fracture gradient prediction is proposed based on analyzing worldwide leak-off test(LOT) data in offshore drilling. Current fracture gradient prediction methods are also reviewed and compared to the proposed method. We analyze more than 200 LOT data in several offshore petroleum basins and find that the fracture gradient depends not only on the overburden stress and pore pressure, but also on the depth. The data indicate that the effective stress coefficient is higher at a shallower depth than that at a deeper depth in the shale formations. Based on this finding,a depth-dependent effective stress coefficient is proposed and applied for fracture gradient prediction. In some petroleum basins, many wells need to be drilled through long sections of salt formations to reach hydrocarbon reservoirs.The fracture gradient in salt formations is very different from that in other sedimentary rocks. Leak-off test data in the salt formations are investigated, and a fracture gradient prediction method is proposed. Case applications are examined to compare different fracture gradient methods and validate the proposed methods. The reasons why the LOT value is higher than its overburden gradient are also explained.展开更多
Temperature and freeze-thaw events are two key factors controlling litter decomposition in cold biomes.Predicted global warming and changes in freeze-thaw cycles therefore may directly or indirectly impact litter deco...Temperature and freeze-thaw events are two key factors controlling litter decomposition in cold biomes.Predicted global warming and changes in freeze-thaw cycles therefore may directly or indirectly impact litter decomposition in those ecosystems. Here, we conducted a2-year-long litter decomposition experiment along an elevational gradient from 3000 to 3600 m to determine the potential effects of litter quality, climate warming and freeze-thaw on the mass losses of three litter types [dragon spruce（Picea asperata Mast.）, red birch（Betula albosinensis Burk.）, and minjiang fir（Abies faxoniana Rehd. et Wild）]. Marked differences in mass loss were observed among the litter types and sampling dates. Decay constant（k） values of red birch were significantly higher than those of the needle litters. However, mass losses between elevations did not differ significantly for any litter type.During the winter, lost mass contributed 18.3-28.8 % of the net loss rates of the first year. Statistical analysis showed that the relationships between mass loss and litter chemistry or their ratios varied with decomposition periods. Our results indicated that short-term field incubations could overestimate the k value of litter decomposition.Considerable mass was lost from subalpine forest litters during the wintertime. Potential future warming may not affect the litter decomposition in the subalpine forest ecosystems of eastern Tibetan Plateau.展开更多
针对柔性直流输电系统(voltage source converter based high voltage direct current transmission,VSC-HVDC)控制参数设计过程中存在的鲁棒性差、依赖已知电路参数、工程设计经验化等问题,提出一种基于马尔科夫转换场(Markov transiti...针对柔性直流输电系统(voltage source converter based high voltage direct current transmission,VSC-HVDC)控制参数设计过程中存在的鲁棒性差、依赖已知电路参数、工程设计经验化等问题,提出一种基于马尔科夫转换场(Markov transition field,MTF)与深度确定性策略梯度算法(deep deterministic policy gradient,DDPG)结合的鲁棒性强、不依赖电路参数特性以及可视化的VSC-HVDC控制参数优化设计方法。首先,采用马尔科夫转换场将电路功率、电压等一维时序波形数据转换为二维马尔科夫转换场域图像并使用马尔科夫转换场损失函数(Markov transition field loss,MTFL)判断二维转换域图的数据波动性;其次,将MTFL损失函数与DDPG算法相结合,综合利用MTFL损失函数对系统输出时序数据动态特性评价能力更强的优点和DDPG算法泛化性能优秀的特点,实现VSC-HVDC系统控制参数优化;最后,通过MATLAB模拟和实验结果验证该方法的有效性。展开更多
Rainfall intensity and slope gradient are two of the most important factors affecting the variations of runoff nitrogen(N).However,the effects of slope gradient and rainfall intensity on N loss via surface flow and in...Rainfall intensity and slope gradient are two of the most important factors affecting the variations of runoff nitrogen(N).However,the effects of slope gradient and rainfall intensity on N loss via surface flow and interflow on weathered granite slopes are poorly understood.In this study,12 artificial rainfalls(three rainfall intensities and four slope gradients)were simulated to investigate the coupling loss characteristics of surface flow–interflow–total nitrogen(TN),nitrate nitrogen(NO_3^--N)and ammonia nitrogen(NH_4^+-N)on weathered granite slopes.The results show that slope gradient has a greater impact on the surface flow when the rainfall intensity is relatively large.The effect gradually weakens with the decrement of rainfall intensity.The interflow yield increases firstly with the prolongation of rainfall duration,then tends to be stable and finally decreases.The total surface flow percentage increases with rainfall intensity while it decreases with increasing slope gradient with a range of 10.88%-71.47%.The TN loss concentration of the surface flow continually decreases with rainfall duration while that of the interflow shows different fluctuations.However,the TN loss loads of both surface flow and interflow increase with increasing rainfall intensity and slope gradient.The NO_3^--N concentration of interflow is much higher than that of the surface flow.The NH_4^+-N concentration is always less than that of NO_3^--N with no significant difference between surface flow and interflow.The percentages of the TN,NO_3^--N,and NH_4^+-N total loss load and concentration of surface flow and interflow were analyzed.The results show that N loss via both surface flow and interflow occurs mainly in the form of NO_3^--N.Most of the N loss is caused by interflow which is the preferential path of runoff nutrient loss.These findings provide data support and underlying insights for the control of runoff and N loss on the weathered granite slopes.展开更多
在基于模型的锂离子电池荷电状态(state of charge,SOC)估计中,传感器采集的数据对SOC估计至关重要。在实际应用中,由于传感器随机故障,可能导致数据缺失的问题,将严重影响SOC估计的准确性。为了解决这一问题,提出了一种基于变递推间隔...在基于模型的锂离子电池荷电状态(state of charge,SOC)估计中,传感器采集的数据对SOC估计至关重要。在实际应用中,由于传感器随机故障,可能导致数据缺失的问题,将严重影响SOC估计的准确性。为了解决这一问题,提出了一种基于变递推间隔修正辅助模型随机梯度的方法。首先,引入了正态分布参数模拟传感器随机故障的情况。然后,在数据缺失下,通过定义已知数据点的整数序列,采用基于变递推间隔的方法得到不可观测的数据,并通过修正辅助模型随机梯度迭代算法补偿未知的信息向量参数。最后,在参数辨识中引入了收敛指数和遗忘因子以提高参数的收敛速度和精度。实验结果表明,所提方法在各种工况和不同缺失率下的平均绝对误差在2.5%以内,表现出较强的鲁棒性。展开更多
目的突发感音神经性耳聋是耳科急症之一,常诱发焦虑-抑郁样情绪障碍,但是突发感音神经性耳聋如何促进情绪障碍的发生尚不清楚。本研究利用脑功能梯度技术探索突发感音神经性耳聋与情绪障碍的关系。材料与方法我们对44名突发感音神经性...目的突发感音神经性耳聋是耳科急症之一,常诱发焦虑-抑郁样情绪障碍,但是突发感音神经性耳聋如何促进情绪障碍的发生尚不清楚。本研究利用脑功能梯度技术探索突发感音神经性耳聋与情绪障碍的关系。材料与方法我们对44名突发感音神经性耳聋患者和40名健康对照组的听力、多维度神经功能及静息状态脑功能梯度进行了评估。采用脑功能连接梯度来识别双侧突发感音神经性耳聋引起的大脑皮层连接梯度的变化。采用统计参数图软件包(statistical parametric mapping 12,SPM12)处理功能磁共振成像(functional magnetic resonance imaging,fMRI)数据,采用社会科学统计软件包(statistical package for the social sciences 22.0,SPSS 22.0)计算fMRI数据与焦虑-抑郁样情绪障碍评分的皮尔森相关性。结果患者与健康对照在年龄、性别和受教育程度上匹配良好。患者组两耳的平均听阈明显高于健康对照组,且突发感音神经性耳聋患者的焦虑和抑郁评分显著增高。结果显示两组间的主梯度没有显著差异。在网络水平,突发感音神经性耳聋在初级梯度上差异无统计学意义,但在节点水平,患者左侧楔前叶的梯度值显著高于健康对照组,而左侧、右侧距状裂周围皮层,右侧海马旁回及左内侧额上回的主梯度值显著低于健康对照组(P<0.001)。突发感音神经性耳聋患者的右侧距状裂周围皮层梯度值与焦虑自评量表得分呈负相关(r=-0.413,P=0.005)。结论突发感音神经性耳聋患者脑区的功能梯度变化有助于阐明突发感音神经性耳聋患者出现情绪障碍的神经病理学基础。展开更多
基金supported by the National Key Research and Development Program of China(No.2020YFA0714103).
文摘Hyperspectral remote sensing image(HSI)fusion with multispectral remote sensing images(MSI)improves data resolution.However,current fusion algorithms focus on local information and overlook long-range dependencies.The parameter of network tuning prioritizes global optimization,neglecting spatial and spectral constraints,and limiting spatial and spectral reconstruction capabilities.This study introduces SwinGAN,a fusion network combining Swin Transformer,CNN,and GAN architectures.SwinGAN’s generator employs a detail injection framework to separately extract HSI and MSI features,fusing them to generate spatial residuals.These residuals are injected into the supersampled HSI to produce thefinal image,while a pure CNN architecture acts as the discriminator,enhancing the fusion quality.Additionally,we introduce a new adaptive loss function that improves image fusion accuracy.The loss function uses L1 loss as the content loss,and spatial and spectral gradient loss functions are introduced to improve the spatial representation and spectralfidelity of the fused images.Our experimental results on several datasets demonstrate that SwinGAN outperforms current popular algorithms in both spatial and spectral reconstruction capabilities.The ablation experiments also demonstrate the rationality of the various components of the proposed loss function.
基金the National Natural Science Foundation of China (90502007) the National Basic Research Program of China (973 Program 2007CB407206)
文摘Based on field runoff plots observation and sample analysis, the effect of slope gradient on soil organic carbon loss was studied under natural rainfall conditions in loess hilly region. The results showed that with slope gradient increasing （from10° to 30°）, the changing trend of soil erosion intensity （A） was A20°〉A30°〉A15°〉A10°〉A25°, suggesting that slope gradient between 20° and 25° was a critical degree ranger to exist. Soil organic carbon loss was mainly influenced by soil erosion intensity, appearing the same trend as did soil erosion intensity with increasing slope gradient. Soil erosion results in organic carbon enrichment in sediment. Average enrichment ratios of five runoff plots varied from 2.27 to 3.74, and decreased with increasing erosion intensity and slope gradient. The decrease of surface runoff and soil erosion is the key to reduce soil organic carbon loss.
基金partially supported by the Program for Innovative Research Team in the University sponsored by Ministry of Education of China(IRT-17R37)National Key R&D Project(2017YFC0804108)of China during the 13th Five-Year Plan PeriodNatural Science Foundation of Hebei Province of China(D2017508099)
文摘The fracture gradient is a critical parameter for drilling mud weight design in the energy industry. A new method in fracture gradient prediction is proposed based on analyzing worldwide leak-off test(LOT) data in offshore drilling. Current fracture gradient prediction methods are also reviewed and compared to the proposed method. We analyze more than 200 LOT data in several offshore petroleum basins and find that the fracture gradient depends not only on the overburden stress and pore pressure, but also on the depth. The data indicate that the effective stress coefficient is higher at a shallower depth than that at a deeper depth in the shale formations. Based on this finding,a depth-dependent effective stress coefficient is proposed and applied for fracture gradient prediction. In some petroleum basins, many wells need to be drilled through long sections of salt formations to reach hydrocarbon reservoirs.The fracture gradient in salt formations is very different from that in other sedimentary rocks. Leak-off test data in the salt formations are investigated, and a fracture gradient prediction method is proposed. Case applications are examined to compare different fracture gradient methods and validate the proposed methods. The reasons why the LOT value is higher than its overburden gradient are also explained.
基金supported by the National Natural Science Foundation of China(3157044531570601+2 种基金31500509 and31570605)Postdoctoral Science Foundation of China(2013M540714 and 2014T70880)Collaborative Innovation Center of Ecological Security in the Upper Reaches of Yangze River
文摘Temperature and freeze-thaw events are two key factors controlling litter decomposition in cold biomes.Predicted global warming and changes in freeze-thaw cycles therefore may directly or indirectly impact litter decomposition in those ecosystems. Here, we conducted a2-year-long litter decomposition experiment along an elevational gradient from 3000 to 3600 m to determine the potential effects of litter quality, climate warming and freeze-thaw on the mass losses of three litter types [dragon spruce（Picea asperata Mast.）, red birch（Betula albosinensis Burk.）, and minjiang fir（Abies faxoniana Rehd. et Wild）]. Marked differences in mass loss were observed among the litter types and sampling dates. Decay constant（k） values of red birch were significantly higher than those of the needle litters. However, mass losses between elevations did not differ significantly for any litter type.During the winter, lost mass contributed 18.3-28.8 % of the net loss rates of the first year. Statistical analysis showed that the relationships between mass loss and litter chemistry or their ratios varied with decomposition periods. Our results indicated that short-term field incubations could overestimate the k value of litter decomposition.Considerable mass was lost from subalpine forest litters during the wintertime. Potential future warming may not affect the litter decomposition in the subalpine forest ecosystems of eastern Tibetan Plateau.
文摘针对柔性直流输电系统(voltage source converter based high voltage direct current transmission,VSC-HVDC)控制参数设计过程中存在的鲁棒性差、依赖已知电路参数、工程设计经验化等问题,提出一种基于马尔科夫转换场(Markov transition field,MTF)与深度确定性策略梯度算法(deep deterministic policy gradient,DDPG)结合的鲁棒性强、不依赖电路参数特性以及可视化的VSC-HVDC控制参数优化设计方法。首先,采用马尔科夫转换场将电路功率、电压等一维时序波形数据转换为二维马尔科夫转换场域图像并使用马尔科夫转换场损失函数(Markov transition field loss,MTFL)判断二维转换域图的数据波动性;其次,将MTFL损失函数与DDPG算法相结合,综合利用MTFL损失函数对系统输出时序数据动态特性评价能力更强的优点和DDPG算法泛化性能优秀的特点,实现VSC-HVDC系统控制参数优化;最后,通过MATLAB模拟和实验结果验证该方法的有效性。
基金supported by the National Natural Science Foundation of China (4187706541471221)
文摘Rainfall intensity and slope gradient are two of the most important factors affecting the variations of runoff nitrogen(N).However,the effects of slope gradient and rainfall intensity on N loss via surface flow and interflow on weathered granite slopes are poorly understood.In this study,12 artificial rainfalls(three rainfall intensities and four slope gradients)were simulated to investigate the coupling loss characteristics of surface flow–interflow–total nitrogen(TN),nitrate nitrogen(NO_3^--N)and ammonia nitrogen(NH_4^+-N)on weathered granite slopes.The results show that slope gradient has a greater impact on the surface flow when the rainfall intensity is relatively large.The effect gradually weakens with the decrement of rainfall intensity.The interflow yield increases firstly with the prolongation of rainfall duration,then tends to be stable and finally decreases.The total surface flow percentage increases with rainfall intensity while it decreases with increasing slope gradient with a range of 10.88%-71.47%.The TN loss concentration of the surface flow continually decreases with rainfall duration while that of the interflow shows different fluctuations.However,the TN loss loads of both surface flow and interflow increase with increasing rainfall intensity and slope gradient.The NO_3^--N concentration of interflow is much higher than that of the surface flow.The NH_4^+-N concentration is always less than that of NO_3^--N with no significant difference between surface flow and interflow.The percentages of the TN,NO_3^--N,and NH_4^+-N total loss load and concentration of surface flow and interflow were analyzed.The results show that N loss via both surface flow and interflow occurs mainly in the form of NO_3^--N.Most of the N loss is caused by interflow which is the preferential path of runoff nutrient loss.These findings provide data support and underlying insights for the control of runoff and N loss on the weathered granite slopes.
文摘现有的基于双向长短时记忆(BiLSTM)网络的命名实体识别(NER)模型难以全面理解文本的整体语义以及捕捉复杂的实体关系。因此,提出一种基于全域信息融合和多维关系感知的NER模型。首先,通过BERT(Bidirectional Encoder Representations from Transformers)获取输入序列的向量表示,并结合BiLSTM进一步学习输入序列的上下文信息。其次,提出由梯度稳定层和特征融合模块组成的全域信息融合机制:前者使模型保持稳定的梯度传播并更新优化输入序列的表示,后者则融合BiLSTM的前后向表示获取更全面的特征表示。接着,构建多维关系感知结构学习不同子空间单词的关联性,以捕获文档中复杂的实体关系。此外,使用自适应焦点损失函数动态调整不同类别实体的权重,提高模型对少数类实体的识别性能。最后,在7个公开数据集上将所提模型和11个基线模型进行对比,实验结果表明所提模型的F1值均优于对比模型,可见该模型的综合性较优。
文摘在基于模型的锂离子电池荷电状态(state of charge,SOC)估计中,传感器采集的数据对SOC估计至关重要。在实际应用中,由于传感器随机故障,可能导致数据缺失的问题,将严重影响SOC估计的准确性。为了解决这一问题,提出了一种基于变递推间隔修正辅助模型随机梯度的方法。首先,引入了正态分布参数模拟传感器随机故障的情况。然后,在数据缺失下,通过定义已知数据点的整数序列,采用基于变递推间隔的方法得到不可观测的数据,并通过修正辅助模型随机梯度迭代算法补偿未知的信息向量参数。最后,在参数辨识中引入了收敛指数和遗忘因子以提高参数的收敛速度和精度。实验结果表明,所提方法在各种工况和不同缺失率下的平均绝对误差在2.5%以内,表现出较强的鲁棒性。
文摘目的突发感音神经性耳聋是耳科急症之一,常诱发焦虑-抑郁样情绪障碍,但是突发感音神经性耳聋如何促进情绪障碍的发生尚不清楚。本研究利用脑功能梯度技术探索突发感音神经性耳聋与情绪障碍的关系。材料与方法我们对44名突发感音神经性耳聋患者和40名健康对照组的听力、多维度神经功能及静息状态脑功能梯度进行了评估。采用脑功能连接梯度来识别双侧突发感音神经性耳聋引起的大脑皮层连接梯度的变化。采用统计参数图软件包(statistical parametric mapping 12,SPM12)处理功能磁共振成像(functional magnetic resonance imaging,fMRI)数据,采用社会科学统计软件包(statistical package for the social sciences 22.0,SPSS 22.0)计算fMRI数据与焦虑-抑郁样情绪障碍评分的皮尔森相关性。结果患者与健康对照在年龄、性别和受教育程度上匹配良好。患者组两耳的平均听阈明显高于健康对照组,且突发感音神经性耳聋患者的焦虑和抑郁评分显著增高。结果显示两组间的主梯度没有显著差异。在网络水平,突发感音神经性耳聋在初级梯度上差异无统计学意义,但在节点水平,患者左侧楔前叶的梯度值显著高于健康对照组,而左侧、右侧距状裂周围皮层,右侧海马旁回及左内侧额上回的主梯度值显著低于健康对照组(P<0.001)。突发感音神经性耳聋患者的右侧距状裂周围皮层梯度值与焦虑自评量表得分呈负相关(r=-0.413,P=0.005)。结论突发感音神经性耳聋患者脑区的功能梯度变化有助于阐明突发感音神经性耳聋患者出现情绪障碍的神经病理学基础。
