基于MTFC的图像复原结果对比分析

地物在成像过程中,由于受到光学成像系统作用、地物周边环境等影响,会使得到的地物影像存在噪声、模糊等各种影像质量下降的问题,这种现象就是图像退化。根据影像退化机制,利用调制传递函数补偿(Modulation Transfer Function Compensation,MTFC)原理复原图像,提高航空面阵成像系统的成像质量,使航空面阵影像更便于判读解译。采用调制传递函数补偿(MTFC)原理进行图像复原。复原图像与原始影像相比在一定程度上边缘更加清晰,更利于判读解译,但在复原过程中也不可避免的添加了噪声。通过复原图像评价指标对比发现,利用航线方向MTF得到的复原图像效果较差,这与成像系统航线方向成像性能偏低有关。     

结合深度学习和图割法的遥感影像建筑物检测

提出一种基于卷积神经网络和图割法的自动提取高分影像建筑物的方法。首先,通过卷积神经网络定位与检测建筑物的位置,逐一提取单个建筑物轮廓,利用检测结果分别建立建筑物和非建筑物的高斯混合模型（GMM）,然后结合最大流最小割的图像分割方式实现全局优化,完成建筑物初步提取,最后用形态学进行优化。通过试验证明了该方法的可行性。     

SAR影像中叠掩与阴影区域的识别 ——以湖北巴东为例

SAR影像中叠掩和阴影区域的识别是滑坡监测中的首要工作,本文针对这一工作利用雷达卫星成像时的几何模型与图像处理中形态学方法综合识别叠掩与阴影区域,以ALOS-2数据为例,对比该地区雷达强度图像,验证了本文试验方法的可靠性。     

一种高光谱遥感图像快速谱聚类算法

为获得分类效果更优良的遥感图像分类方式并解决高光谱遥感图像分类运算速度缓慢的问题,集成Lanczos算法与谱聚类算法,探讨了高光谱遥感图像谱聚类算法应用于遥感图像分类的可行性,提出了一种面向高光谱遥感图像的快速谱聚类算法;通过对比美国圣地亚哥机场高光谱遥感图像K-均值算法与谱聚类算法的分类结果,发现面向高光谱遥感图像的谱聚类算法易于识别线性地物,且分类的速度能得到较大提升。     

高分二号卫星影像自适应模糊阈值法小波去噪

高分二号卫星影像提供了丰富的图像信息,高分二号影像数据的发布打破了我国高分辨率对地观测数据长期依赖进口的局面。但是图像在传输和保存过程中会有噪声干扰,如果感兴趣区域受到污染,则会导致该区域内的影像信息不能被充分利用。为了解决高分二号遥感影像去噪这一难题,本文采用自适应模糊阈值法去噪方式,该方法根据各个尺度下噪声方差建立的自适应模糊阈值函数非线性处理后,重新构造作为新的小波系数,经小波逆变换后得到去噪图像。通过与均值滤波器滤波、高斯平滑滤波、中值滤波器滤波、小波全局阈值去噪和Birge-Massart策略阈值法去噪比较,结果表明,自适应模糊阈值去噪法充分结合软硬阈值处理方式的优点,既保留图像细节又使图像更加平滑,图像整体信息完好,去噪效果更为理想。     

Semantic segmentation of high spatial resolution images with deep neural networks

Availability of reliable delineation of urban lands is fundamental to applications such as infrastructure management and urban planning. An accurate semantic segmentation approach can assign each pixel of remotely sensed imagery a reliable ground object class. In this paper, we propose an end-to-end deep learning architecture to perform the pixel-level understanding of high spatial resolution remote sensing images. Both local and global contextual information are considered. The local contexts are learned by the deep residual net, and the multi-scale global contexts are extracted by a pyramid pooling module. These contextual features are concatenated to predict labels for each pixel. In addition, multiple additional losses are proposed to enhance our deep learning network to optimize multi-level features from different resolution images simultaneously. Two public datasets, including Vaihingen and Potsdam datasets, are used to assess the performance of the proposed deep neural network. Comparison with the results from the published state-of-the-art algorithms demonstrates the effectiveness of our approach.     

沈阳市建成区扩展动态监测

以辽宁省沈阳市为研究对象,以该地区1995年、2000年、2005年以及2016年不同时相的遥感影像为信息源。首先,利用ENVI影像处理软件进行数据预处理;然后,对沈阳地区土地类型进行MLC监督分类并评定分类精度;最后,采用分类后对比的方法并通过Arc GIS软件对沈阳建成区扩展情况进行研究。     

Detection of spatio-temporal evolutions on multi-annual satellite image time series: A clustering based approach

The expansion of satellite technologies makes remote sensing data abundantly available. While the access to such data is no longer an issue, the analysis of this kind of data is still challenging and time consuming. In this paper, we present an object-oriented methodology designed to handle multi-annual Satellite Image Time Series (SITS). This method has the objective to automatically analyse a SITS to depict and characterize the dynamic of the areas (the way that the land cover of the areas evolve over time). First, it identifies the spatio-temporal entities (reference objects) to be tracked. Second, the evolution of such entities is described by means of a graph structure and finally it groups together spatio-temporal entities that evolve similarly. The analysis were performed on three study areas to highlight inter (among the study areas) and intra (inside a study area) similarity by following the evolution of the underlying phenomena. The analysis demonstrate the benefits of our methodology. Moreover, we also stress how an expert can exploit the extracted knowledge to pinpoint relevant landscape evolutions in the multi-annual time series and how to make connections among different study areas.     

深度卷积神经网络支持下的遥感影像飞机检测

针对YOLOv3算法对小目标检测较差及出现较多漏检的问题,本文提出了一种优化的YOLOv3算法.首先使用K-means算法计算出与数据集相适用的锚框;其次将扩张卷积引入到YOLOv3网络,用来增强网络高层的感受野,改善小目标的检测效果;然后使用深度可分离卷积取代YOLOv3网络残差模块中的普通卷积,可减少计算量,从而得到一种新型卷积神经网络结构;最后在数据集上进行对比试验.结果表明,优化的YOLOv3算法能够检测出更多目标,降低漏检率,相比于YOLOv3算法,其召回率提高11.86%,F1-score提高2.99%.     

顾及局部与结构特征的稀疏多项式逻辑回归高光谱图像分类方法

稀疏多项式逻辑回归在分类中仅利用图像光谱信息,导致分类效果不太理想.本文提出了一种顾及局部与结构特征的稀疏多项式逻辑回归高光谱图像分类方法.首先利用加权均值滤波与拓展形态学多属性剖面对原始高光谱图像进行局部与结构特征提取;然后对二者进行加权平均特征级融合以获取更具唯一性的像元特征;最后由稀疏多项式逻辑回归分类器对融合结果进行分类.结果表明,本文方法能有效地提高分类精度,而且具有较强的稳健性.