An efficient nonlinear approach for removing fixed-value impulse noise from grayscale images

The latest-generation earth observation instruments on airborne and satellite platforms are currently producing an almost continuous high-dimensional data stream. This exponentially growing data poses a new challenge for real-time image processing and recognition. Making full and effective use of the spectral information and spatial structure information of high-resolution remote sensing image is the key to the processing and recognition of high-resolution remote sensing data. In this paper, the adaptive multipoint moment estimation (AMME) stochastic optimization algorithm is proposed for the first time by using the finite lower-order moments and adding the estimating points. This algorithm not only reduces the probability of local optimum in the learning process, but also improves the convergence rate of the convolutional neural network (Lee Cun et al. in Advances in neural information processing systems, 1990). Second, according to the remote sensing image with characteristics of complex background and small sensitive targets, and by automatic discovery, locating small targets, and giving high weights, we proposed a feature extraction method named weighted pooling to further improve the performance of real-time image recognition. We combine the AMME and weighted pooling with the spatial pyramid representation (Harada et al. in Comput Vis Pattern Recognit 1617–1624, 2011) algorithm to form a new, multiscale, and multilevel real-time image recognition model and name it weighted spatial pyramid networks (WspNet). At the end, we use the MNIST, ImageNet, and natural disasters under remote sensing data sets to test WspNet. Compared with other real-time image recognition models, WspNet achieve a new state of the art in terms of convergence rate and image feature extraction compared with conventional stochastic gradient descent method [like AdaGrad, AdaDelta and Adam (Zeiler in Comput Sci, 2012; Kingma and Ba in Comput Sci, 2014; Duchi et al. in J Mach Learn Res 12(7):2121–2159, 2011] and pooling method [like max-pooling, avg-pooling and stochastic-pooling (Zeiler and Fergus in stochastic-pooling for regularization of deep convolutional neural networks, 2013)].     

基于深度学习的高分辨率卫星遥感影像围填海检测识别

基于高分辨率卫星遥感影像自动、准确提取围填海土地利用现状,是实现围填海集约使用的重要技术手段。针对高分辨率卫星遥感影像地物特征复杂,依赖人工提取特征的传统方法较难满足业务部门实际需求的问题,提出了基于深度学习的围填海检测识别技术框架,该框架使用UNet网络的多约束变体结构,并针对高分辨率遥感影像地物特征复杂导致地物分类不一致的问题,引入全连接条件随机场和图像腐蚀运算对分割结果进行后处理。以天津市滨海新区2016年和2020年高分辨卫星遥感影像为数据源进行了验证,实验表明围填海地物分割整体准确率、F1-score、Kappa系数以及mIoU分别达到96.73%、92.87%、90.28%、86.82%。在此基础上,分析提取了该围填海区域土地利用动态变化特征,为围填海集约使用管理提供了有效技术支撑。     

A new shadow detection index for 8-band very high-resolution satellite images

ABSTRACT

The new generation of remote sensing satellite with very high-resolution images has provided a high level of details, which make them a reliable source of information. Presence of shadow can reduce the amount of information that can be extracted from these images. Shadow can be confused with dark objects such as water and dark vegetation. The main aim of this research is to develop a new index to detect shadow in the presence of dark objects using the capabilities of the new remote sensing satellite images. For this study, WorldView-2 (WV-2) remote sensing satellite images with eight spectral bands were used. A spectral reflectance analysis for the main ground features has been studied along the eight spectral bands to determine the most effective bands for shadow detection. These bands are employed with the Hue-Saturation-Intensity colour model for producing the new proposed Saturation Intensity Shadow Detection Index (SISDI). The proposed index is applied to four study areas and compared with two state-of-the-art indices of shadow detection. Results of this comparison demonstrate the more accuracy effectiveness and feasibility of that proposed index. The proposed index achieves the highest overall accuracy (average of 97.8%) and has the ability for detecting small shadow areas.     

Extraction of bridges over water from IKONOS panchromatic data

Compared to remote sensing images of medium or low spatial resolution, high‐resolution remote sensing images can provide observation data containing more detailed information for georesearch. Accordingly, an important issue for current computer and geoscience experts is to develop useful methods or technology to extract information from these high‐resolution satellite images. As part of a series of research into object extraction, this paper focuses mainly on the extraction of bridges over water from high‐resolution panchromatic satellite images. Since bridges over water are obviously adjacent to water in remote sensing images, this paper proposes a practical knowledge‐based bridge extraction method for remote sensing images of high spatial resolution. The steps involved are: water extraction based on Gauss Markov Random Field (GMRF)‐Support Vector Machine (SVM) classification methods which use a SVM to classify the image based on textural features expressed by a GMRF; image thinning and removal of fragmented lines; main trunk detection by width; vectorization; and feature expression. Finally, tests are described for two pieces of panchromatic IKONOS satellite images with a 1 m resolution. The experimental results show that the proposed method is suitable for images with a single‐peak histogram (contrast between water and land is sharp) or a multi‐peak histogram (greyscale value of water is close to that of land).     

基于对象关系特征的高分辨率光学卫星影像水上桥梁目标识别方法

利用卫星遥感技术对大中型桥梁进行识别定位,在民用上和军事上都具有很重要的意义。本研究提出了一套利用基元对象关系特征提取高分辨率卫星影像中水上桥梁的技术方法。首先利用多尺度分割算法对高分辨率卫星影像进行分割,利用水体指数或GLCM同质性纹理特征区分河水和陆地;其次,利用对象形状特征和相邻的关系特征提取桥梁潜在区;将河流片段和桥梁潜在区专题二值化,利用数学形态学算子实现河流水面的连续化;最后利用叠加分析的方法获得最终的桥梁目标。本方法充分利用了桥梁与河流相邻和相交的空间关系特征,利用QuickBird和IKONOS高分辨率卫星影像进行实验,证明所提出的方法可以高精度的实现大中型水上桥梁的识别定位。     

基于WPS快速构建面向专业应用的WEB遥感影像处理应用系统试验研究

卫星遥感技术的飞速发展,促进了遥感数据的应用从专业领域向更加广阔的领域发展,这就需要提供应用简单、使用方便、不依赖于专业遥感影像处理系统的应用工具为各领域的专业人员、公众使用,OGC的WPS服务为基于通用浏览器的遥感影像处理应用提供了可能。本文尝试了以ERDAS APOLLO SERVER为服务器基于WPS服务快速构建网络遥感影像处理应用的一种新框架。基于该框架开发的系统很好地解决了海量影像数据的管理、网络影像应用功能的提供、网络影像处理功能使用数据的输入/输出、功能和数据的安全性等问题,并初步构建了可实际应用的通用影像处理应用系统。     

Towards an early‐warning system for global landslides triggered by rainfall and earthquake

Drawing upon the recent advances of satellite remote‐sensing technology and landslide modelling techniques, a framework is proposed to attempt an early‐warning system for landslide hazards after heavy rainfall and/or earthquake, the two major triggers for landslides. This framework includes three major components: (1) a landslide susceptibility information database, including geology, elevation, topography, soil, and land‐cover types; (2) a real‐time space‐borne precipitation estimation system (http://trmm.gsfc.nasa.gov); and (3) a near‐real‐time ground‐shaking prediction system after earthquakes (http://earthquake.usgs.gov/eqcenter/shakemap/). The ultimate goal of this framework is to rapidly predict landslide potential after large earthquakes and/or heavy rainfall by combing the dynamic triggers with landslide susceptibility information derived from high‐resolution geospatial datasets. However, the challenge for integrating these real‐time systems into an operational landslide prediction network and quickly disseminating alerts around the world is tremendous. It requires continued efforts and interdisciplinary collaboration in the next 2–5 years in order to realize such a system, providing early warning for landslides around the globe in a day‐to‐day decision‐making operation.     

一种卫星遥感影像村庄区域提取算法

卫星遥感影像提取村庄区域在地理和气象领域均有十分重要的意义.针对卫星遥感影像的特点,提出了一种村庄区域提取方法.利用改进的去雾算法对卫星遥感影像进行预处理,通过遥感卫星影像的颜色特征实现分割,结合村庄区域分布特点进行去噪处理,实现卫星遥感影像村庄区域的提取.实验结果表明:该算法能够对卫星遥感图像中不同类型村庄区域进行提取,且提取准确率高,可以应用于地理以及气象等领域.     

Quantitative monitoring of inland water using remote sensing of the upper reaches of the Huangpu River,China

In this paper, a quantitative framework using common and readily available remote sensing data, including ground hyperspectral data, multispectral remote sensing images and a regular in situ water quality monitoring programme, is proposed to monitor inland water quality. The entire framework has three steps: (1) collecting and processing basic data, including remote sensing data and water quality data; (2) examining the relationships between water quality parameters and water reflectance from both remote sensing images and in situ measurement data. According to their relationships with ground hyperspectral reflectance, the water quality parameters are classified into three categories, and the corresponding monitoring models using remote sensing data are presented for these three categories; and (3) analysing the spatial distribution by using water quality concentration maps generated with the monitoring models. The upper reaches of the Huangpu River were chosen as our study area to test this framework. The results show that the concentration maps inverted by the proposed models are in accordance with the actual situation. Therefore, we can conclude that the proposed framework for quantifying water quality based on multisource remote sensing data and regular in situ measurement data is an effective and economic tool for the rapid detection of changes in inland water quality and subsequent management.     

基于深度学习的卫星遥感图像边缘检测方法

为解决卫星遥感图像边缘模糊噪点过多,导致图像清晰度过低的问题,提出基于深度学习的卫星遥感图像边缘检测方法。利用Softmax分类器结构,提取边缘图像节点处的数据信息参量,遵循深度学习算法,完成对图像信息的卷积与池化处理,实现基于深度学习的卫星遥感图像识别。根据尺度空间定义原则,确定边缘检测特征点所处位置,再联合梯度信息熵计算结果,完成对卫星遥感图像的拼接处理。分别计算一阶微分边缘算子、二阶微分边缘算子的具体数值,确定梯度幅值的取值区间,总结已知数值参量,建立完整的双阈值表达式,完成基于深度学习的卫星遥感图像边缘检测方法的设计。实验结果表明,应用所提方法后卫星遥感图像边缘节点处信噪比指标增大,可有效控制模糊噪点对图像清晰度的影响,在卫星遥感图像边缘精准检测方面具有较强的实用性。     

遥感图像星上智能处理地面仿真模拟系统设计与实现

传统卫星遥感应用模式复杂繁长,无法满足用户越来越关注的实时化遥感服务需求,为卫星配备智能化大脑,一方面可以降低数据传输带宽,另一方面可以提高数据获取的时效性,因此,星上智能处理已经成为遥感卫星发展的必然选择。但星上处理在轨调试困难,现有遥感卫星星上处理平台的地面测试系统都是卫星实验室测试时,针对不同的卫星载荷临时组建,缺乏通用性且并未形成集成化的装置,导致现有遥感图像星上智能处理的地面测试效率偏低。尤其是面对目前星上处理智能化的新需求,缺乏一套高性能、低功耗、全流程的星上处理地面仿真系统。针对遥感数据处理自动化与智能化发展的新特点,提出了一套基于FPGA与GPU相结合的遥感图像星上处理地面仿真模拟系统。该系统能够在地面模拟实现多种载荷的0到1级数据预处理,在预处理的基础上实现智能遥感影像的加速识别,其关键难点在于遥感图像智能处理算法的高计算复杂度和嵌入式计算机有限计算力之间的平衡;遥感图像处理领域的AI专用算法固化和硬件加速之间的平衡;不同卫星平台测试需求和系统架构通用性之间的平衡。本文阐述了仿真平台设计的方法,构建了基本原型并对其进行了验证。测试结果表明：该系统可以较好地完成星上智能处理典型算法地面全流程测试,所有硬件可以直接上星组装,完备度高,对优化和指导卫星地面仿真系统运行管理体系具有一定的参考价值。     

基于卫星遥感数据的恶劣气象监测预警系统设计

针对传统恶劣气象监测预警系统的检测准确率低、预警时效性较差的问题,提出基于卫星遥感数据的恶劣气象监测预警系统。利用GDAL软件自带的读写功能,处理已读取的遥感影像基本信息,再根据辐射定标与亮温计算数值,完成卫星遥感影像的预处理模块设计。连接插件模块的逻辑结构,利用待集成的平台插件,控制下级遥感信息显示与预警模块,实现恶劣气象监测预警系统的设计与应用。按照系统多线程的同步与通信关系,完成遥感数据的云识别处理,实现卫星遥感程序与恶劣气象监测预警模块之间的实时交互。选取风向指标作为实验对象,分析对比实验数据可知,在不同风向条件下,卫星遥感型监测系统所检测出的风向指标均呈现显著性状态,能够及时响应系统主机的预警指令。     

Image-Based Attitude Control of a Remote Sensing Satellite

This paper deals with the image-based control of a satellite for remote sensing. Approach is demonstrated by simulation where the position of the satellite is obtained with the Simplified General Perturbations Version 3 model and its orientation by simulating its dynamic and kinematic models. For a known position and orientation of the satellite the images are obtained using the satellite’s onboard camera, simulated by the Google Earth application. The orientation of the satellite is governed by reaction wheels, which produce the required moments to the satellite. The image-based control law using SIFT image features is applied to achieve an automatic reference-point observation on the Earth’s surface. Main contributions of the paper are the following: use of the same sensor for Earth observation and attitude control, simplicity of the approach, no need for explicit calibration of camera parameters and good tracking accuracy. Demonstrated simulation results and performance analysis confirm the approach applicability.     

一种遥感卫星动量轮容错控制的方法研究

针对低轨遥感卫星动量轮故障频发影响姿态稳定的情况,对以动量轮为执行器的低轨遥感卫星进行容错控制研究,提出一种结合卡尔曼滤波及模糊PID控制率的容错控制方法。首先建立低轨遥感卫星动量轮故障模型,应用卡尔曼滤波算法进行噪声滤除,然后根据卡尔曼滤波的结果,设计模糊PID容错控制器,并与传统PID容错控制器进行仿真比对实验,在闭环姿态控制系统的容错控制领域验证了该方法的较强自适应性和较强鲁棒性的优点。     

HJ-1A/1B卫星远程播报系统的实现

基于环境与灾害监测预报小卫星星座中的HJ-1A、1B卫星遥感数据,提出一种卫星影像数据与矢量数据在远程播报系统中快速叠加匹配的方法与技术流程。在此基础上,实现了HJ-1A、1B卫星宽覆盖CCD相机的数据远程播报。系统测试表明,远程播报系统对HJ-1A、1B卫星数据空间定位准确,数据处理速度快,满足远程播报实时显示数据的要求。     

基于流式计算的遥感卫星数据快视处理方法

随着高分辨率遥感卫星数据获取能力和地面数传接收能力的提高，现有遥感卫星快视处理系统的处理负载增大，实时性要求越来越难以满足。针对这些问题，采用流式计算思想提出了一种新的遥感卫星数据快视处理系统设计方法。在分析遥感卫星数据快视处理数据流特点的基础上，应用Storm框架对现有系统进行并行优化，设计遥感数据流处理任务拓扑结构，同时利用消息队列中间件Kafka改进处理单元间数据交换和数据缓存方式。实验表明，该系统在数据吞吐率和可靠性方面测试效果良好。     

基于元数据和快视图的遥感图像库管理与发布

从遥感多光谱图像数据库的应用角度,讨论了如何构建一个基于元数据和快视图的,能够精确定位和查看光谱的遥感图像数据库管理系统,并详细说明了其设计开发的方法和步骤。     

基于稠密残差网络的多序列卫星图像去云

遥感影像中最常见的问题是云层污染,它会导致图像信息缺失,降低遥感数据的可用性。针对该问题,提出了一种基于稠密残差网络的多序列卫星图像去云算法。首先,该网络使用多序列的有云卫星图像作为输入,能为网络提供更多的时序特征信息,提升去云效果;其次,在网络中段使用稠密残差层,以保证卷积层之间最大程度地传递和使用特征信息,让生成的修复图像整体结构合理、边缘细节更加清晰;最后,使用像素上采样来增强空间信息的利用,提升修复效果。该方法在欧洲"哨兵-2"遥感卫星图像数据集上进行验证,峰值信噪比和结构相似度指标为27.59和0.854 0,两项指标均超过了该数据集的原处理方法STGAN,提升了遥感图像去云的效果。     

遥感视频实时云检测的FPGA高效实现方法

针对目前遥感图像云检测算法及算法运行所需硬件平台复杂度高,无法进行在轨实时检测的问题,提出了一种基于FPGA的面向卫星在轨实时运行的遥感视频云检测方法.首先根据不同的遥感视频输入格式对其自适应降采样处理;其次对顺序流入的图像自适应阈值分割,然后对分割后的图像进行聚类获取云区域,进而提取每一块云区域的特征向量;最后计算整幅图像的云覆盖率和可用度,以此判断是否将图像下传.实验结果表明,在60 MHz的时钟下,且Camera Link接口每个时钟周期同时输入两个像素时,822×1096大小的遥感视频云检测速度可达132 fps,相对于传统的嵌入式双核CPU,速度提升了6～7倍.该方法可实现卫星在轨实时云检测,极大地缓解了有限的星地数传带宽和巨大的遥感数据量之间的矛盾,大幅提升遥感卫星系统应用效能,具有很强的实用价值.