A modified stochastic neighbor embedding for multi-feature dimension reduction of remote sensing images

In automated remote sensing based image analysis, it is important to consider the multiple features of a certain pixel, such as the spectral signature, morphological property, and shape feature, in both the spatial and spectral domains, to improve the classification accuracy. Therefore, it is essential to consider the complementary properties of the different features and combine them in order to obtain an accurate classification rate. In this paper, we introduce a modified stochastic neighbor embedding (MSNE) algorithm for multiple features dimension reduction (DR) under a probability preserving projection framework. For each feature, a probability distribution is constructed based on t-distributed stochastic neighbor embedding (t-SNE), and we then alternately solve t-SNE and learn the optimal combination coefficients for different features in the proposed multiple features DR optimization. Compared with conventional remote sensing image DR strategies, the suggested algorithm utilizes both the spatial and spectral features of a pixel to achieve a physically meaningful low-dimensional feature representation for the subsequent classification, by automatically learning a combination coefficient for each feature. The classification results using hyperspectral remote sensing images (HSI) show that MSNE can effectively improve RS image classification performance.     

Clustering based on eigenspace transformation – CBEST for efficient classification

Large remote sensing datasets, that either cover large areas or have high spatial resolution, are often a burden of information mining for scientific studies. Here, we present an approach that conducts clustering after gray-level vector reduction. In this manner, the speed of clustering can be considerably improved. The approach features applying eigenspace transformation to the dataset followed by compressing the data in the eigenspace and storing them in coded matrices and vectors. The clustering process takes the advantage of the reduced size of the compressed data and thus reduces computational complexity. We name this approach Clustering Based on Eigen-space Transformation (CBEST). In our experiment with a subscene of Landsat Thematic Mapper (TM) imagery, CBEST was found to be able to improve speed considerably over conventional K-means as the volume of data to be clustered increases. We assessed information loss and several other factors. In addition, we evaluated the effectiveness of CBEST in mapping land cover/use with the same image that was acquired over Guangzhou City, South China and an AVIRIS hyperspectral image over Cappocanoe County, Indiana. Using reference data we assessed the accuracies for both CBEST and conventional K-means and we found that the CBEST was not negatively affected by information loss during compression in practice. We discussed potential applications of the fast clustering algorithm in dealing with large datasets in remote sensing studies.     

基于高分辨率SAR图像成像机理的震害信息分析

与中低分辨率SAR图像相比,高分辨率SAR图像受目标复杂性和成像噪声等因素的干扰更为严重,致使应用高分辨率SAR图像检测震害目标变得更加困难.传统的适用于中低分辨率SAR图像的震害信息提取方法对于高分辨率SAR图像有一定的局限性.为了从高分辨率SAR图像中更加准确地提取震害信息,从SAR图像成像机理和目标的后向散射特征出发,对建筑物和道路、桥梁的震害特征进行了细致分析,为寻求合适的目标检测和提取方法提供了思路.     

机载LiDAR和高光谱数据融合提取冰川雪线

以西藏那曲县境内的“中习一号”冰川为研究区,对2011年8月获取的机载激光雷达点云进行预处理和滤波分类,提取研究区数字高程模型(digital elevation model,DEM);将DEM数据分别与同期获取的机载高光谱栅格数据和提取出的冰川矢量数据进行三维地形模拟,利用DEM数据对高光谱最大似然法分类结果进行正射纠正,从而获取研究区的数字正射影像(digital orthophoto map,DOM);最后结合研究区DOM和机载点云数据提取“中习一号”冰川的雪线.结果表明:融合机载高光谱和机载激光雷达2种数据的优势,能更方便地提取出冰川雪线,而且能很好地显示雪线的高度.     

基于暗原色的航空遥感影像薄雾去除与改进

本文对航空遥感影像的薄雾进行了研究,根据雾图像成像模型和基于暗原色先验知识,结合图像对比度的自适应调整,提出了一种对航空遥感影像薄雾去除的改进算法。采用数字航空相机DSS439获取的不同地物特征类型的应急遥感影像进行去雾实验,并对实验结果进行分析、比较与评价,验证了该方法的正确性和有效性。     

基于改进SIFT算法的视频图像序列自动拼接

本文提出了一种自适应帧采样和限定特征提取区域的拼接方法,根据帧间重叠率和帧间隔建立线性模型,并把各帧图像对准到其前后的关键帧上。在特征点提取方面,提出了一种改进的SIFT算法进行特征点提取,并采用随机采用一致性(RANSAC)方法来更新匹配点,在图像融合中采用线性加权渐入渐出的自然融合算法。实验结果表明:该方法对一般场景能稳定的抽取到关键帧,并进行拼接,取得了较好的拼接效果。     

顾及地物空间关系的卫星影像水系信息提取

为了更好地提取影像阴影部分的信息,本文采用一种顾及地物间空间关系的水系提取方法,将地物之间的空间关系与面向对象的方法相结合。通过多尺度分割算法生成同质的影像对象,利用决策树、隶属度函数和阴影与建筑物的邻接关系对水体进行提取。以昆明某区域影像为实验数据进行了试验,试验结果表明,所提出的方法可以有效解决阴影对水系提取造成的影响,大大提高了水系提取的精度。     

基于AdaBoost算法的遥感影像水体信息提取

AdaBoost算法利用每个特征构造一个简单分类器,然后将简单分类器进行训练组合成一个强分类器。算法能够充分利用每个分类器的优势并避免其劣势,得到一个最佳判别,达到提高分类精度的目的。本文利用TM影像,将影像各波段灰度、水体指数和谱间关系特征相结合,构成提取水体的强分类器,实现水体提取。实验结果表明,算法能够非常有效地、高精度地提取水体信息。     

GIS服务在测绘应急保障体系中的应用

本文研究基于GIS服务,建立一套地理空间数据快速提取、遥感影像快速纠正和遥感影像快速制图的软件服务平台,实现成果数据由人工被动提供到计算机主动提供的转变.通过四川汶川地震灾后恢复重建测绘专项建设工程的实际应用证明,平台可以实现基于GIS服务的控制点选取、DEM提取、遥感纠正及多源异构数据的智能提取;利用该平台完成了阿坝州测绘保障与援助项目、“8.13泥石流”抢险救灾的遥感影像纠正处理及影像挂图的制作.     

Feasibility of estimating heavy metal contaminations in floodplain soils using laboratory-based hyperspectral data—A case study along Le’an River,China

It is necessary to estimate heavy metal concentrations within soils for understanding heavy metal contaminations and for keeping the sustainable developments of ecosystems. This study, with the floodplain along Le’an River and its two branches in Jiangxi Province of China as a case study, aimed to explore the feasibility of estimating concentrations of heavy metal lead (Pb), copper (Cu) and zinc (Zn) within soils using laboratory-based hyperspectral data. Thirty soil samples were collected, and their hyperspectral data, soil organic matters and Pb, Cu and Zn concentrations were measured in the laboratory. The potential relations among hyperspectral data, soil organic matter and Pb, Cu and Zn concentrations were explored and further used to estimate Pb, Cu and Zn concentrations from hyperspectral data with soil organic matter as a bridge. The results showed that the ratio of the first-order derivatives of spectral absorbance at wavelengths 624 and 564 nm could explain 52% of the variation of soil organic matter; the soil organic matter could explain 59%, 51% and 50% of the variation of Pb, Cu and Zn concentrations with estimated standard errors of 1.41, 48.27 and 45.15 mg·kg?; and the absolute estimation errors were 8%–56%, 12%–118% and 2%–22%, and 50%, 67% and 100% of them were less than 25% for Pb, Cu and Zn concentration estimations. We concluded that the laboratory-based hyperspectral data hold potentials in estimating concentrations of heavy metal Pb, Cu and Zn in soils. More sampling points or other potential linear and non-linear regression methods should be used for improving the stabilities and accuracies of the estimation models.