基于ALBERT模型的园林植物知识实体与关系抽取方法

园林植物知识图谱可为顾及区域适应性、观赏性和生态性等因子的绿化树种的选型提供知识支持。植物描述文本的实体识别及关系抽取是知识图谱构建的关键环节。针对植物领域未有公开的标注数据集,本文阐述了园林植物数据集的构建流程,定义了园林植物的概念体系结构,完成了园林植物语料库的构建。针对现有Word2vec、ELMo和BERT等语言模型存在无法解决多义词、融合上下文能力差、运行速度慢等缺点,提出了嵌入ALBERT（A Lite BERT）预训练语言模型的实体识别和关系抽取模型。ALBERT预训练的动态词向量能够有效地表示文本特征,将其分别输入到BiGRU-CRF命名实体识别模型和BiGRU-Attention关系抽取模型中进行训练,进一步提升实体识别和关系抽取的效果。在园林植物语料库上进行方法的有效性验证,结果表明ALBERT-BiGRU-CRF命名实体识别模型的F₁值为0.9517,ALBERT-BiGRU-Attention关系抽取模型的F₁值为0.9161,相较于经典的语言模型（如Word2vec、ELMo和BERT等）性能有较为显著的提升。因此基于ALBERT模型的实体与关系抽取任务能有效提高识别分类效果,可将其应用于植物描述文本的实体关系抽取任务中,为园林植物知识图谱自动构建提供方法。     

基于路网数据和复杂图论的中国城市交通布局评价

城市化的快速发展推动了我国城市道路网的不断扩张和发展。城市道路网是一个动态、开放、自组织的空间复杂网络,构成了城市的结构框架,研究城市道路网的结构特征对路网规划和城市建设有着重要的应用价值。在国内城市道路网络结构特征的相关研究中,很少有学者基于道路走向的视角对整个城市路网结构进行研究,也缺少对全国主要城市路网的整体性评价。本文选取我国一、二线及新一线49个城市作为研究区和2020年2月的城市道路网作为实验数据,首先采用图论和玫瑰图的方法对49个城市的道路网方向特征进行可视化,定性分析城市路网的复杂结构,然后基于最值比R、道路首位度S、过阈值比T、走向顺序φ和路网密度δ等5个路网指标对49个城市的道路网进行聚类分析,从而探究我国城市路网的分布特征。结果表明,我国城市主要以南北和东西走向为城市道路的主导向,部分城市因地形的影响沿着有利于城市交通和居民生活的方向规划道路;基于路网指标的聚类得到了十字正交型、十字状向风车状扩散型、风车状向圆弧四角星扩散型和混合复杂型等4类城市道路网类型,这4类城市路网在方向分布特征上具有显著差异,突出表现在有序性和复杂性;从路网类型的空间分布来看,十字正交型和十字状向风车状扩散型路网大多分布在我国的内陆地区,而风车状向圆弧四角星扩散型和混合复杂型路网则大多分布在沿海地区。本文主要通过分析我国一、二线及新一线城市道路网的分布特征,探究了我国主要城市交通路网的布局现状,能够为城市新区道路规划和路网布局优化提供参考依据。     

渤海表层悬沙的时空分布特征及其动力成因

针对目前渤海整个海域悬浮泥沙分布全貌的研究不充分。根据2000—2004年渤海表层悬浮泥沙分布特征选取7个典型海区,通过利用长时间序列悬沙质量浓度和风场遥感反演资料,在分析悬沙质量浓度与局地风速、物质来源等关系的基础上,定量研究风浪和潮流共同作用下、随季节显著变化的沉积物再悬浮过程,从而揭示整个渤海海域代表性海区悬沙质量浓度时空分布的动力成因。渤海不同海区表层悬沙质量浓度绝对值差别很大,多年平均最高质量浓度在20~450mg/L变化,高质量浓度集中在近岸河口区及其邻近海域,如黄河口和辽河口地区,低质量浓度区位于渤海中部、渤海海峡以及秦皇岛外海(属于近岸海域却质量浓度常年偏低的特殊海区)。渤海表层悬沙质量浓度具有明显的季节变化特征,风场的季节变化是主要影响因子,各代表性海区悬沙质量浓度与风速之间具有显著正相关关系。悬沙质量浓度与风速之间存在一定时间段的滞后相关。沉积物再悬浮的定量研究表明,除渤海海峡外,渤海其它典型海区表层悬沙质量浓度及其季节变化,均与各自海区风浪和潮流共同作用产生的最大底流速及其季节变化相对应。在渤海,底层沉积物再悬浮的季节变化是影响悬沙质量浓度季节变化最关键的动力过程。     

Optical closure of parameterized bio-optical relationships

An optical closure study on bio-optical relationships was carried out using radiative transfer model matrix operator method developed by Freie Universitt Berlin.As a case study,the optical closure of bio-optical relationships empirically parameterized with in situ data for the East China Sea was examined.Remote-sensing refl ectance(R rs)was computed from the inherent optical properties predicted by these biooptical relationships and compared with published in situ data.It was found that the simulated R rs was overestimated for turbid water.To achieve optical closure,bio-optical relationships for absorption and scattering coeffi cients for suspended particulate matter were adjusted.Furthermore,the results show that the Fournier and Forand phase functions obtained from the adjusted relationships perform better than the Petzold phase function.Therefore,before bio-optical relationships are used for a local sea area,the optical closure should be examined.     

Phytoplankton community characteristics in the coastal waters of the southeastern Arabian Sea Phytoplankton community characteristics in the coastal waters of the southeastern Arabian Sea

Remote sensing applications are important in the fisheries sector and efforts were on to improve the predic-tions of potential fishing zones using ocean color. The present study was aimed to investigat...     

Seasonality of oceanic primary production and its interannual variability from 1998 to 2007

The seasonality of primary productivity plays an important role in nutrient and carbon cycling. We quantify the seasonality of satellite-derived, oceanic net primary production (NPP) and its interannual variability during the first decade of the SeaWiFS mission (1998 to 2007) using a normalized seasonality index (NSI). The NSI, which is based upon production half-time, t(1/2), generally becomes progressively more episodic with increasing latitude in open ocean waters, spanning from a relatively constant rate of primary productivity throughout the year (mean t(1/2) ~5 months) in subtropical waters to more pulsed events (mean t(1/2) ~3 months) in subpolar waters. This relatively gradual, poleward pattern in NSI differs from recent estimates of phytoplankton bloom duration, another measure of seasonality, at lower latitudes (~40°S–40°N). These differences likely reflect the temporal component of production assessed by each metric, with NSI able to more fully capture the irregular nature of production characteristic of waters in this zonal band. The interannual variability in NSI was generally low, with higher variability observed primarily in frontal and seasonal upwelling zones. The influence of the El Niño–Southern Oscillation on this variability was clearly evident, particularly in the equatorial Pacific, where primary productivity was anomalously episodic from the date line east to the coast of South America in 1998. Yearly seasonality and the magnitude of annual production were generally positively correlated at mid-latitudes and negatively correlated at tropical latitudes, particularly in a region bordering the Pacific equatorial divergence. This implies that increases of annual production in the former region are attained over the course of a year by shorter duration but higher magnitude NPP events, while in the latter areas it results from an increased frequency or duration of similar magnitude events. Statistically significant trends in the seasonality, both positive and negative, were detected in various patches. We suggest that NSI be used together with other phenomenological characteristics of phytoplankton biomass and productivity, such as the timing of bloom initiation and duration, as a means to remotely quantify phytoplankton seasonality and monitor the response of the oceanic ecosystem to environmental variability and climate change.     

Cycle graph analysis for 3D roof structure modelling: Concepts and performance

The paper presents a cycle graph analysis approach to the automatic reconstruction of 3D roof models from airborne laser scanner data. The nature of convergences of topological relations of plane adjacencies, allowing for the reconstruction of roof corner geometries with preserved topology, can be derived from cycles in roof topology graphs. The topology between roof adjacencies is defined in terms of ridge-lines and step-edges. In the proposed method, the input point cloud is first segmented and roof topology is derived while extracting roof planes from identified non-terrain segments. Orientation and placement regularities are applied on weakly defined edges using a piecewise regularization approach prior to the reconstruction, which assists in preserving symmetries in building geometry. Roof corners are geometrically modelled using the shortest closed cycles and the outermost cycle derived from roof topology graph in which external target graphs are no longer required. Based on test results, we show that the proposed approach can handle complexities with nearly 90% of the detected roof faces reconstructed correctly. The approach allows complex height jumps and various types of building roofs to be firmly reconstructed without prior knowledge of primitive building types.     

一种结合地理知识的遥感影像目标实体关联方法北大核心CSCD

遥感影像目标识别技术已广泛应用于目标动态监测与定位等领域。但影像目标识别的结果缺乏与目标属性信息的链接,导致分析人员只能依据影像特征进行分析,难以进行更复杂的目标数据关联分析与挖掘。针对遥感影像目标识别语义属性信息缺失的问题,本文利用知识图谱相关技术将影像判别的目标信息与知识语义网链接。首先,提出了一种遥感影像目标知识图谱构建框架;其次,针对遥感影像目标不同的数据类型,构建遥感影像目标知识抽取模型,提出了基于相似度目标实体识别和预定义模式的关系抽取方法;然后,基于多特征Logistic模型的影像目标实体链接方法,实现了遥感影像目标实体与百科知识库的知识关联;最后,针对预定试验区域进行试验,验证了本文方法的可行性。     

资源一号02C卫星多光谱数据真彩色模拟技术研究

资源一号02C卫星多光谱数据没有蓝色波段,给制作真彩色影像带来了一定困难。本文针对02C卫星多光谱波段的特点,分析了常用线性模拟蓝波段影像方法,选取多景不同时相的、覆盖山地或平地的真实数据进行加权法和平均值法的对比试验,定性与定量分析后的结果表明：基于加权法生成蓝波段的真彩色模拟结果,无论是从整体色调方面,还是从典型地类的颜色方面,均优于基于平均值法生成蓝波段的真彩色模拟结果,可进一步深入研究后进行规模化应用。     

Characterization of spatial and temporal variation of suspended sediments in the Yellow and East China Seas using satellite ocean color data

Total suspended sediment (TSS) data concentrations are retrieved from two sets of satellite ocean color data (the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard Aqua and the Korean Geostationary Ocean Color Imager (GOCI)) using an existing regional model to characterize spatial and temporal variation of TSS in the Yellow and East China Seas. MODIS-derived TSS maps show that TSS concentrations are, in general, high along the Korean and Chinese coasts including the Bohai Sea and the Yangtz River estuary, and lower in the middle of the Yellow Sea and the southeastern area of the East China Sea. The monthly average of 10-year MODIS data reveals that TSS values are highest during winter (January to February) and lowest in summer (July to August). Short-term TSS concentrations retrieved from GOCI data showed the dominant influence of semi-diurnal tidal changes on sediment dynamics through temporal (hourly) and spatial distribution in coastal zones of the Yellow sea. The results presented here demonstrate that the satellite-derived TSS products can be utilized as an application tool for future studies on long- and short-term sediment dynamics of turbid coastal waters. In particular, GOCI observations provide unique important capabilities to characterize and quantify the water properties at high temporal (hourly) and spatial (0.5 km) resolutions in the turbid coastal waters of the Yellow Sea and its vicinities.