Impact of Grazing Exclusion on the Surface Heat Balance in North Tibet

The grazing exclusion program used by the Tibetan government to protect the ecological environment has changed the vegetation and impacted the surface heat balance in North Tibet. However, little information is available to describe the in?uences of the current grazing exclusion program on local surface heat balance. This study uses the records of fenced grassland patch locations to identify the impact of grazing exclusion on surface heat balance in North Tibet. The records of fenced grassland patch locations, including the longitude, latitude, and elevation of the vertices of each fenced patch (polygon shapes), were provided by the agriculture and animal husbandry bureaus of the counties where the patches were located. ArcGIS 10.2 was used to create polygon shapes based on patch location records. Based on satellite data and the surface heat balance system determined by the model, values for changes in land surface temperature (LST), albedo and evapotranspiration (ET) induced by grazing exclusion were obtained. All of these can influence surface heat balance and alter the fluctuation of LST in the northern Tibetan Plateau. The LST trends for day and night showed an asymmetric diurnal variation, with a larger magnitude of warming in the day than cooling at night. The maximum decrease in absorbed shortwave of LST (-0.5 - -0.4 ℃ per decade) occurred in the central region, while the minimum decrease (-0.2 - -0.1 ℃ per decade) occurred in the eastern region. The decreased latent heat lead to the LST increased maximum (>1 ℃ per decade) occurred in the central region, The eastern region increased at a rate of 0.2-0.5 ℃ per decade, while the minimum increase (0-0.1 ℃ per decade) occurred in the northwestern region.     

结合BM3D和多级非线性加权平均中值滤波的遥感影像混合噪声去噪方法

提出了一种基于三维块匹配（BM3D）和多级非线性加权平均中值滤波的遥感影像混合噪声去噪方法,使用coief3小波和sym2小波替换传统三维块匹配算法中的bior1.5和hear参数,将三维块匹配算法和多级非线性加权平均中值滤波相结合以提高算法对混合噪声的去噪能力。进行一个仿真对比试验和一个真实遥感影像去噪试验,试验表明,本文提出的方法在遥感影像混合噪声去噪方面效果要优于传统混合噪声去噪方法。     

基于GIS的石家庄市雾霾分析

为了得出雾霾气象成因机制、影响因素和时空分布特征,以石家庄市区与郊县为研究区域,将2013年9月至2016年12月石家庄市各市区、郊县的PM2.5历史监测数据中的有效数据进行了数据分析处理得到雾霾浓度数据,还有温度、降水、风速、地形和人口密度等数据,运用GIS分析的方法,模拟绘制石家庄雾霾的时空分布图、雾霾与各影响因子的专题对比图,得到雾霾形成机制的因子、雾霾的时空分布规律、雾霾季节变化特征、雾霾与地形间的关系等;运用数据分析软件OriginPro8.SR3分析雾霾浓度与风速数据、降雨数据、温度数据间的相关性。     

Small‐area patch‐merging method accounting for both local constraints and the overall area balance

Small‐area patch merging is a common operation in land use data generalization. However, existing research on small‐area patch merging has mainly focused on local compatibility measures, which often lead to area imbalances among land use types from a global perspective. To address the shortcomings of previous studies by resolving local and global concerns simultaneously, this article proposes a merging method that considers both local constraints and the overall area balance. First, a local optimization model that considers three constraints—namely, the areas of neighboring patches, the lengths of shared arcs, and semantic similarity—is established. The areas of small patches are first pre‐allocated. Subsequently, in accordance with an area change threshold for individual land use types, land use types with area changes that exceed this threshold are identified. The patches corresponding to these land use types are subjected to iterative adjustments while considering the overall area balance. Based on their area splitting abilities, the split lines for small‐area patches are determined, and small‐area patches are merged. Finally, actual data from Guangdong Province are used for validation. The experimental results demonstrate that the proposed method is capable of preserving the local compatibility of patches while balancing the overall area associated with each land use type.     

图库一体化软件系统研究与实现

为了提高DLG和DRG数据生产效率,改变传统作业模式,解决图库数据分离造成的数据不同步、数据维护复杂等问题,笔者引入"图库一体化"解决方案,通过建立图库一体化数据模型,利用制图表达原理研究算法实现复杂符号效果,采用智能化标注、地图整饰自动化等关键性技术,基于Arc GIS Engine实现Umap图库一体化系统。     

开源地理空间情报的搜集方法

在过去的20年,地理空间情报已经由不存在而一跃成为全球情报活动的主要内容主体。基于开源的地理空间情报搜集作为一种低风险、高收益的情报搜集手段,成为各国开展地理空间情报工作的一项重要内容。本文根据开源情报源的主要渠道,分析总结了几种主要的开源地理空间情报搜集方法,重点阐述基于互联网的搜集方法,并给出了具体的分析实践,最后提出开源情报搜集的几点建议。     

FY-3D/MERSI-II全球火点监测产品及其应用

FY-3D/MERSI-II全球火点监测产品主要包括全球范围内的火点位置、亚像元火点面积和火点强度等信息,可用于实时监测全球范围的森林草原火灾、秸秆焚烧等生物质燃烧状况。火点判识算法主要根据中红外通道对高温热源的敏感特性,即含有火点的中红外通道像元辐亮度和亮温较远红外通道的辐亮度和亮温偏高,同时较周边非火点的中红外像元偏高,建立合适的阈值可探测含有火点的像元。亚像元火点面积估算主要使用中红外单通道估算,根据亚像元火点面积估算结果对火点强度进行分级,不同的级别表示不同程度的火点辐射强度。基于全球火点自动判识结果,每日生成0.01°分辨率的卫星遥感日全球火点产品,每月生产0.25°×0.25°格点的全球月火点密度图。在利用FY-3D/MERSI-II火点产品开展的全球火点监测应用中,对多起全球重大野火事件进行了监测,为防灾减灾、全球气候变化研究、生态环境保护等方面提供卫星遥感信息支持。     

复杂卫星图像中的小目标船舶识别

船舶作为海上的重要目标,实现对船舶自动识别有重要的意义。针对卫星图像中云雾、海岸背景等复杂海情对船舶识别带来的干扰,以及小目标船舶高漏检率问题,本文提出一种多尺度深度学习模型训练策略,在此基础上构建了一种船舶识别的深度学习网络,该网络可分为多尺度训练、特征提取、生成目标建议区域、船舶分类这4个部分。首先,采用多尺度的训练策略,将多尺度的船舶样本送入网络中进行训练,这样在训练样本中加入了大量小目标船舶的样本,使网络充分提取到小目标船舶的特征;其次,通过卷积神经网络对目标船舶进行特征自适应提取;然后,目标区域建议网络可依据卷积神经网络提取到的特征,在图像中找到感兴趣目标区域,即框定船舶的位置;最后,通过多个全连接层的组合,将高维特征映射到一个4元组中,再运用分类函数输出每一类船舶的概率值,概率值最大的则为该船舶的类别。同时为解决云雾遮挡和海岸背景的干扰,采用了一种负样本增强学习的方法,在样本数据集中加入了大量只含有云雾和海岸背景的图片,进行负样本扩充,增强网络模型对云雾及海岸背景的特征学习能力,以此解决复杂海情的影响。实验结果表明,所提方法有效解决了复杂海情条件下的船舶识别难,以及小目标船舶识别难的问题,实现了复杂海情条件下的船舶识别。同时,与现有成熟的深度学习目标识别算法相比,本文算法的精确度和召回率分别提升了6.98%和18.17%,所训练的模型具有良好的泛化能力和鲁棒性。     

高分三号SAR影像双阈值变化检测

双阈值合成孔径雷达SAR(Synthetic Aperture Radar)变化检测算法具有在发现变化区域的同时还能确定地表发生后向散射变化类型的优点。针对广义高斯双阈值最小误差法D-GKIT(Dual Generalized Kittler and Illingworth Thresholding)在进行阈值选取时直方图中不同类别像素灰度级重叠严重时,分割结果容易在尖峰单侧选取出双阈值而导致无法正确分割差异图的问题,本文提出一种结合归一化最大类间方差和广义高斯最小误差法GKIT(Generalized Kittler and Illingworth Thresholding)的双阈值SAR变化检测方法。首先,提出以归一化最大类间方差值作为灰度级重叠程度的判别参数,确定阈值的选取顺序及两个候选区间;然后,利用GKIT在候选区间内进行分割,获取单侧阈值及非变化类拟合函数;最后,提出利用非变化类拟合函数更新后的直方图作为另一侧阈值选取基础进行分割,得到对应分割阈值。以宁波地区高分三号(GF-3)SAR卫星影像作为试验研究数据,结果表明:本文方法能较好地解决灰度级重叠时D-GKIT无法进行正确分割的问题,具有良好的变化检测效果和更强的鲁棒性且达到了利用研究区数据验证利用GF-3号SAR卫星影像进行变化检测研究可行性的目的。