利用卷积神经网络的高光谱图像分类

针对高光谱图像分类中对光谱信息利用不足的问题,提出一种基于卷积神经网络在光谱域开展的分类算法。该算法通过构建五层网络结构,逐像素对光谱信息开展分析,将全光谱段集合作为输入,利用神经网络展开代价函数值的计算,实现对光谱特征的提取与分类。实验中采用三组高光谱遥感影像数据进行对比分析,以India Pines数据集为例,提出的基于卷积神经网络的分类方法的分类正确率达到90.16%,比RBF-SVM方法高出2.56%,相比三种传统的深度学习方法高出1%~3%,训练速度也较为理想。实验结果表明,本文所提出的算法充分利用了高光谱图像中逐像素点的光谱域信息,能够有效提高分类正确率。与传统学习算法相比,在较少训练样本的情况下,更能发挥其良好的分类性能。     

鄱阳湖区域夏冬季节沉积物面积变化分析研究

为更好地研究河流污染、河道阻塞、土壤沙化和湖泊面积锐减等问题,本文采用适宜的分析模型和高分辨率资源一号卫星影像,按照影像预处理、影像分类、矢量提取、特征分析的作业步骤,利用面向对象的遥感影像分类方法进行信息提取,针对夏冬两季的地物特征分别对鄱阳湖地区进行沉积物识别、地物类别判读以及湖泊的面积变化分析。研究表明:夏季鄱阳湖区域水体面积较大,泥沙较少,而冬季湖泊面积减少,泥沙等沉积物增多,并且存在相互影响的关系。     

爆发性气旋研究的回顾

爆发性气旋又称“气象炸弹”,定义为在考虑地转调整到60°N时气旋中心气压加深率大于1 hPa/h的快速发展的气旋,具有中心气压急剧降低、强度急剧增大的特点,多发于洋面上,对海上航行安全及沿岸人民生产生活具有重要的影响。近几十年来,众多学者对爆发性气旋开展了广泛而深入的研究,在爆发性气旋的气候学特征、结构特征和发展机制等方面取得了较大的进展,但同时还有许多亟待解决的问题。本文结合前人对爆发性气旋的研究工作,系统地回顾和总结了爆发性气旋的研究进展,希望能够为将来对爆发性气旋的研究工作带来一些启发和思考。     

Performance evaluation of object based greenhouse detection from Sentinel-2 MSI and Landsat 8 OLI data: A case study from Almería (Spain)

This paper shows the first comparison between data from Sentinel-2 (S2) Multi Spectral Instrument (MSI) and Landsat 8 (L8) Operational Land Imager (OLI) headed up to greenhouse detection. Two closely related in time scenes, one for each sensor, were classified by using Object Based Image Analysis and Random Forest (RF). The RF input consisted of several object-based features computed from spectral bands and including mean values, spectral indices and textural features. S2 and L8 data comparisons were also extended using a common segmentation dataset extracted form VHR World-View 2 (WV2) imagery to test differences only due to their specific spectral contribution. The best band combinations to perform segmentation were found through a modified version of the Euclidian Distance 2 index. Four different RF classifications schemes were considered achieving 89.1%, 91.3%, 90.9% and 93.4% as the best overall accuracies respectively, evaluated over the whole study area.     

基于GIS的岷江上游流域土地利用／覆被变化研究

以岷江上游流域为对象,选取3期9景TM/+ETM遥感影像,通过多步骤最大似然监督分类、变化检测,结合空间动态分析测算模型,分析近20年土地利用/覆被变化情况。结果表明:从整个流域分析,林地面积减少,主要转化为未利用地、建设用地和耕地;未利用地在前8年以每年3.7%、后8年以每年0.4%的速度增加;建设用地在1994—2002年以每年34%的速度增加,到2002—2010年增长速度减缓;耕地总面积减少54 431hm2;从县域分析,1994—2002年间,松潘和黑水县大面积林地转为未利用地;2002—2010年间,松潘县未利用地转为林地和建设用地,茂县和汶川县未利用地面积大幅增加。该研究结论不仅为国土资源管理部门优化土地利用结构提供依据,亦为当地政府实现生态资源可持续发展提供数据支撑。     

改进的邻近四点法建筑物多边形化简算法

针对应用邻近四点法化简建筑物多边形时因实际数据的复杂性及该算法的不完善出现的效率下降和准确性降低问题,该文提出了一种基于邻近四点法建筑物多边形化简的改进算法。该方法排除了冗余点的干扰,避免运算耗时;细化了基本处理单元的分类,避免遗漏特殊结构类型基本处理单元的处理;对建筑物多边形进行了分类及实时标记,以达到在同时考虑基本处理单元和建筑物多边形的情况下实现建筑物多边形化简的目的。实验结果表明:此改进算法更加实用化,且提高了建筑物多边形化简的效率和准确性。     

综合纹理与光谱的土地利用模糊分类

针对利用TM影像进行土地利用传统分类精度不高的问题,该文提出了一种综合应用影像纹理与光谱特征对TM影像进行土地利用模糊分类的方法。采用主成分分析法对研究流域TM影像的光谱及纹理特征信息进行压缩与融合,并对融合后的TM影像数据进行3个组别的多尺度分割,在影像分割对象单元的基础上应用面向对象的模糊逻辑隶属度函数法实现影像的软语义分类。相对传统分类方法而言,该方法在充分利用影像光谱信息的基础上综合了影像的纹理信息,且分类理论思想更加符合人们对于客观事物的认知规律,分类精度有了显著的提高,为TM影像分类方法的改进提供一定的参考。     

可扩展的迭代自组织分析并行处理算法

随着遥感影像数据量的增加,传统非监督分类迭代自组织分析(ISODATA)算法的运算将十分耗时,应用并行计算技术能够有效解决该性能瓶颈。针对现有基于并行计算模型MapReduce的遥感迭代自组织分析并行算法存在的局限性,提出一种可扩展的基于MapReduce的迭代自组织分析并行处理算法。该算法通过其包含的全局子采样算法、聚类中心点集合过滤算法以及聚类映射算法,有效克服了现有并行算法中存在的不足。实验结果表明,在同等规模遥感计算中,该算法效率高于现有并行处理算法,具有良好的加速比,且在处理更大的影像块时具有更高的精度。     

Satellite-Based Assessment of the Extent and Changes in the Mangrove Ecosystem of the Niger Delta

This study was carried out with the primary aim of understanding how the mangrove ecosystem in the Niger Delta has been altered recently. Specifically, we determined the spatial extent of the mangrove forest in the Niger Delta using remotely sensed satellite data and estimated changes in the spatial extent of the forest from the mid-1980s through 2003. Overall, about 21,340 hectares of Mangrove forest was lost over the study period. Fieldwork confirmed that these losses were primarily due to urbanization, dredging activities, activities of the oil and gas industries, and the spread of Nypa Palm (Nypa frutican) plant species.     

Surface geostrophic currents across the Antarctic circumpolar current in Drake Passage from 1992 to 2004

The Southern Ocean plays an important role in the global overturning circulation as a significant proportion of deep water is converted into intermediate and deeper water masses in this region. Recently, a secular trend has been reported in wind stress around the Southern Ocean and it is thought theoretically that the strength of the ACC is closely related to wind stress, so one consequence should be a corresponding increase in ACC transport and hence changes in the rate of the global overturning. There are no long-term data sets of ACC transport and so we must examine other data that may also respond to changing wind stress. Here we calculate surface currents in Drake Passage every seven days over 11.25 years from 1992 to 2004. We combine surface velocity anomalies calculated from satellite altimeter sea surface heights with measured surface currents. Since 1992, the UK has regularly occupied WOCE hydrographic section SR1b across the ACC in Drake Passage. From seven hydrographic sections surface currents are estimated by referencing relative geostrophic velocities from CTD sections with current measurements made by shipboard and lowered acoustic Doppler current profilers. Combining the seven estimates of surface currents with the altimeter data reduces bias in the estimates of average currents over time through Drake Passage and we show that surface current anomalies estimated by satellite and in situ observations are in good agreement. The strongest surface currents are found in the Subantarctic and Polar Fronts with average speeds of 50 cm/s and 35 cm/s, respectively and are inversely correlated, so that maximum westward flow in one corresponds to minimum westward flow in the other. The average cross-sectional weighted surface velocity from 1992 to 2004 is 16.7 ± 0.2 cm/s. A spectral analysis of the average surface current has only weakly increasing energy at higher frequencies and there is no dominant mode of variability. The standard deviation of the seven day currents is 0.68 cm/s and a running 12 month average has only a slightly smaller standard deviation of 0.52 ± 0.16 cm/s. The southern annular mode (SAM) measures the circumpolar average of wind stress and like the surface currents its spectrum has slightly increased energy at frequencies greater than 1 cpy. A cospectral analysis of these, averaging cospectra of five slightly overlapping 36 month segments improve statistical reliability, suggests that there is coherence between them at 1 cpy with the currents leading changes in the Southern annular mode. We conclude that the SAM and average Drake Passage surface currents are weakly correlated with no dominant co-varying modes, and hence predicting Southern Ocean transport variability from the SAM is not likely to give significant results and that secular trends in surface currents are likely to be masked by weekly and interannual variability.