A multi-index learning approach for classification of high-resolution remotely sensed images over urban areas

In recent years, it has been widely agreed that spatial features derived from textural, structural, and object-based methods are important information sources to complement spectral properties for accurate urban classification of high-resolution imagery. However, the spatial features always refer to a series of parameters, such as scales, directions, and statistical measures, leading to high-dimensional feature space. The high-dimensional space is almost impractical to deal with considering the huge storage and computational cost while processing high-resolution images. To this aim, we propose a novel multi-index learning (MIL) method, where a set of low-dimensional information indices is used to represent the complex geospatial scenes in high-resolution images. Specifically, two categories of indices are proposed in the study: (1) Primitive indices (PI): High-resolution urban scenes are represented using a group of primitives (e.g., building/shadow/vegetation) that are calculated automatically and rapidly; (2) Variation indices (VI): A couple of spectral and spatial variation indices are proposed based on the 3D wavelet transformation in order to describe the local variation in the joint spectral-spatial domains. In this way, urban landscapes can be decomposed into a set of low-dimensional and semantic indices replacing the high-dimensional but low-level features (e.g., textures). The information indices are then learned via the multi-kernel support vector machines. The proposed MIL method is evaluated using various high-resolution images including GeoEye-1, QuickBird, WorldView-2, and ZY-3, as well as an elaborate comparison to the state-of-the-art image classification algorithms such as object-based analysis, and spectral-spatial approaches based on textural and morphological features. It is revealed that the MIL method is able to achieve promising results with a low-dimensional feature space, and, provide a practical strategy for processing large-scale high-resolution images.     

Spectral monitoring of moorland plant phenology to identify a temporal window for hyperspectral remote sensing of peatland

Recognising the importance of the timing of image acquisition on the spectral response in remote sensing of vegetated ecosystems is essential. This study used full wavelength, 350–2500 nm, field spectroscopy to establish a spectral library of phenological change for key moorland species, and to investigate suitable temporal windows for monitoring upland peatland systems. Spectral responses over two consecutive growing seasons were recorded at single species plots for key moorland species and species sown to restore eroding peat. This was related to phenological change using narrowband vegetation indices (Red Edge Position, Photochemical Reflectance Index, Plant Senescence Reflection Index and Cellulose Absorption Index); that capture green-up and senescence related changes in absorption features in the visible to near infrared and the shortwave infrared. The selection of indices was confirmed by identifying the regions of maximum variation in the captured reflectance across the full spectrum. The indices show change in the degree of variation between species occurring from April to September, measured for plant functional types. A discriminant function analysis between indices and plant functional types determines how well each index was able to differentiate between the plant functional groups for each month. It identifies April and July as the two months where the species are most separable. What is presented here is not one single recommendation for the optimal temporal window for operational monitoring, but a fuller understanding of how the spectral response changes with the phenological cycle, including recommendations for what indices are important throughout the year.     

高动态宽带信号码跟踪方法仿真分析

针对高动态场景,单独的码环路很难实现跟踪,由于高动态载波跟踪的算法很成熟,通常应用载波跟踪结果对码环路进行辅助,针对窄体制信号,这种方法可以帮助消除码环的动态误差,但对宽体制信号来说,辅助力度减小。从高动态宽带信号码跟踪误差门限以及跟踪精度入手,分析了单独码跟踪算法的易失锁性,理论和仿真验证应用高动态载波跟踪结果辅助码跟踪算法的有效性,且具有高的跟踪精度。这为导航接收机的跟踪算法提供了理论依据。     

福建省GPS基准站数据处理与分析

利用福建省41个GPS基准站2011—2014年的观测资料,运用GAMIT／GLOBKl0．4软件及块体整体旋转线性应变模型,得到了福建省各测站及分块的水平运动速度场和块体上的应变图像。通过分析研究2011—2013和2011-2014两个时段速度场和应变场的特征,为福建省未来地震活动判定提供形变依据。     

A Velocity Dealiasing Scheme for C-band Weather Radar Systems简

A dealiasing algorithm for radar radial velocity observed by C-band Doppler radars is presented as an extension of an existing S-band dealiasing algorithm. This has operational significance in that many portable and many commercial broadcast radars, as well as approximately one half of the Chinese weather radar network （CINRAD）, are C-band radars. With a wavelength of about 5 cm, the Nyquist interval of C-band radars is just about one half that of S-band radars （wavelength of about 10 cm） and thus has more velocity folding. The proposed algorithm includes seven modules to remove noisy data, find the starting radials, dealias velocities, and apply least squares error checking in both the radial and azimuth directions. The proposed velocity dealiasing method was applied to one widespread rain case and three strong convective cases from radars operating in China. It was found that, on average, 92.95% of the aliased radial velocity data could be correctly de-aliased by the algorithm, resulting in 96.65% of the data being valid.     

Western Arctic Cyclones and Equilibrium between the Atmospheric Boundary Layer and the Sea Surface

The data-collection campaign for the 2008 International Polar Year–Circumpolar Flaw Lead System Study saw the Canadian Coast Guard Ship (CCGS) Amundsen, a research icebreaker, overwinter in high-concentration unconsolidated sea ice in Amundsen Gulf. Environmental monitoring continued into the open-water season. During this period, the Amundsen registered five relatively deep mean sea-level pressure minima (less than 100?kPa). Three were selected for further analysis based on season and the nature of the underlying ocean or sea-ice surface: (1) a winter pressure minimum over unconsolidated sea ice, (2) a spring pressure minimum which likely contributed to the break-up of the sea-ice cover on Amundsen Gulf, and (3) a summer pressure minimum over open water. The characteristics of these pressure minima and the impact of their passage on the atmospheric boundary layer and on the sea-ice cover as they crossed Amundsen Gulf were examined. Several features were revealed by the analysis. (1) The winter and summer pressure minima were migratory cyclones accompanied by Arctic frontal waves with characteristics very similar to the polar frontal waves associated with the migratory cyclones found at more southerly latitudes, whereas the spring pressure minimum was attributed to an Arctic frontal trough of low pressure with the cyclonic centre remaining south of the Gulf. (2) The passage of the frontal-wave cyclone in winter and the frontal trough of low pressure in spring disrupted the equilibrium that had been established during more settled periods between the atmospheric boundary layer and the mosaic surface (leads, polynyas, and sea ice); however, equilibrium was quickly re-established. (3) In summer, the thermal structure of the lower atmospheric boundary layer persisted through the passage of the frontal-wave cyclone over the open-water surface. (4) The passage of the frontal-wave cyclone in winter and the frontal trough of low pressure in spring modified the mesoscale sea-icescape.     

新疆克拉玛依强下坡风暴的机理研究

利用美国中尺度数值模式 WRF 对2013年3月7—8日克拉玛依强风进行了模拟,对下坡风发生、发展和结束3个阶段的三维结构特征进行了分析,并由此提出克拉玛依强下坡风的形成机制模型：上游地区出现中高层西南风、低层西北风并伴有强冷平流的配置,当风速不断增大时,气流能够翻越加依尔山在背风坡侧形成重力波,重力波相位向气流上游方向倾斜产生非线性效应,促进了波不稳定区域的形成并导致波破碎,形成湍流活跃层,不断把上层的能量向下传播;克拉玛依中低层形成三层夹心的大气层结稳定度分布,出现明显的过渡气流带从而导致强下坡风的形成;南北风分量在低层和中层符号相反,形成了临界层,不断吸收上层波能量并向地面传送,强下坡风暴不断维持发展。最后利用2006—2012年克拉玛依33个强下坡风过程中的探空观测资料对提出的形成机制进行了验证。     

改进的HRNet应用于路面裂缝分割与检测

针对利用传统卷积神经网络进行路面裂缝分割时存在准确率低、信息丢失及边缘模糊的问题,本文提出了基于改进HRNet模型的路面裂缝分割算法。模型在原始HRNet的基础上进行改进,主干网络部分采用DUC模块代替双线性插值上采样;下采样改为passthrough layer代替原始卷积;在模型解码部分,进行逐级上采样的同时引入SE-Block,对不同特征层的融合重新标定权重。通过与原始HRNet及传统卷积神经网络U-Net对比可知,本文算法在公共数据与自制数据集上的分割精度表现优秀,F1分值分别达到了91.31%和78.69%,可以很好地满足实际工程的需求。     

基于平面波解构滤波的速度无关动校正及速度分析方法

为了避免速度分析拾取速度参数引起的繁复费时、拾取参数不准确导致的校正效果不佳和传统动校正的拉伸畸变问题,利用速度无关动校正方法进行相关处理。通过数据解构获取局部同相轴斜率,解析斜率数据与速度的关系,由局部斜率可直接获得动校正所需时差参数,其信息足以进行动校正处理。通过模拟和实际数据进行的实验计算,证明此方法优于常规动校正处理效果。