SRS-Net: Training object detectors from scratch for remote sensing images without pretraining

Most of the current object detection algorithms use pretrained models that are trained on ImageNet and then fine-tuned in the network, which can achieve good performance in terms of general object detectors. However, in the field of remote sensing image object detection, as pretrained models are significantly different from remote sensing data, it is meaningful to explore a train-from-scratch technique for remote sensing images. This paper proposes an object detection framework trained from scratch, SRS-Net, and describes the design of a densely connected backbone network to provide integrated hidden layer supervision for the convolution module. Then, two necessary improvement principles are proposed: studying the role of normalization in the network structure, and improving data augmentation methods for remote sensing images. To evaluate the proposed framework, we performed many ablation experiments on the DIOR, DOTA, and AS datasets. The results show that whether using the improved backbone network, the normalization method or training data enhancement strategy, the performance of the object detection network trained from scratch increased. These principles compensate for the lack of pretrained models. Furthermore, we found that SRS-Net could achieve similar to or slightly better performance than baseline methods, and surpassed most advanced general detectors.     

Ship detection and classification from optical remote sensing images: A survey

Considering the important applications in the military and the civilian domain, ship detection and classification based on optical remote sensing images raise considerable attention in the sea surface remote sensing filed. This article collects the methods of ship detection and classification for practically testing in optical remote sensing images, and provides their corresponding feature extraction strategies and statistical data. Basic feature extraction strategies and algorithms are analyzed associated with their performance and application in ship detection and classification. Furthermore, publicly available datasets that can be applied as the benchmarks to verify the effectiveness and the objectiveness of ship detection and classification methods are summarized in this paper. Based on the analysis, the remaining problems and future development trends are provided for ship detection and classification methods based on optical remote sensing images.     

Spacecraft design considerations for small satellite remote sensing

The idea of adapting existing small satellite technology for remote sensing purposes is discussed. The major design problems and constraints influencing the design of a small low-cost remote sensing satellite bus are identified using the subsystem approach. Key design areas include the improvement of battery technology and the development of a deployable solar array, attitude control assemblies, on-board data processing/storage, and ground station data acquisition. Although the eventual satellite would also have to be somewhat larger, more powerful and, above all, more sophisticated than the previous small satellites, this is considered to be a natural progression of research in this area     

遥感图像的亚像素配准技术研究

高分辨遥感技术在军民领域应用广泛,图像配准技术是高分辨遥感图像的预处理技术。实际的机载和星载遥感成像仪,由于自身成像原理和成像仪平台抖动原因,出现了遥感图像亚像素偏移误差。对亚像素图像配准技术进行系统研究,对高精度亚像素配准算法进行研究分析,并对其算法进行仿真、处理,对仿真的结果进行系统分析,通过误差分析提出配准误差修正算法,修正算法对亚像素配准进行了很好的改进,达到遥感图像预处理的目的。     

非光学遥感数据的可视化处理

对非光学遥感数据进行可视化处理有利于信息的提取和识别。伪彩色处理是非光学遥感图像增强和显示的常用方法,调色板的编码技术是非光学遥感图像伪彩色显示的关键。对照灰度编码方法,给出了非线性伪彩色编码方法和类JET伪彩色编码方法,并以干涉合成孔径声纳湖试数据处理为例,验证了这些方法的有效性。     

基于条件随机场的遥感图像语义标注

遥感图像包含的信息丰富,纹理复杂,而遥感图像语义标注又为后续的目标识别、检测、场景分析及高层语义的提取提供了重要信息和线索,这使其成为遥感图像理解领域中一个关键且极具挑战性的任务。首先针对遥感图像语义标注问题,提出采用条件随机场(CRF)框架对遥感图像的底层特征和上下文信息建模的方法,将Texton纹理特征与CRF中的自相关势能结合来捕捉遥感图像中的纹理信息及其上下文分布,采用组合Boosting算法进行Texton纹理特征选择和参数学习;然后将Lab空间中的颜色信息与CRF中的互相关势能结合来描述颜色上下文;最后用Graph Cut算法对CRF进行推导求解,得到图像自动语义标注结果。同时,建立了可见光遥感图像数据库Google-4,并对全部图像进行了人工标注。Google-4上的实验结果表明:采用CRF框架与Texton纹理特征和颜色特征相结合对遥感图像建模的方法与基于支持向量机(SVM)的方法相比较,能够取得更准确的语义标注结果。     

Target localization based on cross-view matching between UAV and satellite

Matching remote sensing images taken by an unmanned aerial vehicle (UAV) with satellite remote sensing images with geolocation information. Thus, the specific geographic location of the target object captured by the UAV is determined. Its main challenge is the considerable differences in the visual content of remote sensing images acquired by satellites and UAVs, such as dramatic changes in viewpoint, unknown orientations, etc. Much of the previous work has focused on image matching of homologous data. To overcome the difficulties caused by the difference between these two data modes and maintain robustness in visual positioning, a quality-aware template matching method based on scale-adaptive deep convolutional features is proposed by deeply mining their common features. The template size feature map and the reference image feature map are first obtained. The two feature maps obtained are used to measure the similarity. Finally, a heat map representing the probability of matching is generated to determine the best match in the reference image. The method is applied to the latest UAV-based geolocation dataset (University-1652 dataset) and the real-scene campus data we collected with UAVs. The experimental results demonstrate the effectiveness and superiority of the method.     

Three-Dimensional Solar Wind Modeling Using Remote-Sensing Data

We have developed a computer-assisted tomography (CAT) technique that iteratively modifies a kinematic solar wind model to least-squares fit heliospheric remote sensing observations (interplanetary scintillation and Thomson-scattering observations). These remote sensing data cover a large range of solar elongations, and access high-latitude regions over the solar poles. The technique can be applied to a time-independent solar wind model, assuming strict co-rotation, or, when sufficient remote sensing observations are available, to a time-dependent model. For the time-dependent case the technique depends primarily on outward motion of structures in the solar wind to provide the perspective views required for a tomographic reconstruction. We show results of corotating tomographic reconstructions primarily using IPS velocity observations from the Solar-Terrestrial Environment Laboratory (STELab, Nagoya, Japan), and include comparisons with in situ velocity data out of the ecliptic (Ulysses) and in the ecliptic (ACE). This revised version was published online in August 2006 with corrections to the Cover Date.     

基于TV-L1模型的多尺度遥感影像分解方法

TV—L^2模型是一种有效的遥感影像分解方法,但采用TV—L^2模型进行遥感影像分解会出现分解不彻底的情况。基于TV—L^1模型的多尺度遥感影像分解方法是一种将TV—L^1模型和多尽度方法结合起来对遥感影像进行分解的遥感影像分解方法。实验证明,基于TV—L^1模型的多尺度遥感影像分解方法继承了Tv—^L2模型利用全变分分解的优点,解决了TV—L^2模型分解不彻底的问题,并且通过尺度调节实现了多尺度分解。实验结果表明,该方法在遥感影像分解中取得了理想的效果,可以进一步地扩展应用。     

Sustained data recording system based on disk array

Most sustained data recording systems in many real-time fields are based on tape equipments. We have developed an innovative sustained data recording system based on disk array which consists of Array Controller Module (ACM), String Controller Module (SCM) and Main Controller Module (MCM). This system has a better performance and higher reliability than the traditional tape recorder and can be used conveniently in many areas of data recording, storing, playback, and remote backup. This novel recording system has been used in radar data recorder for spacecraft orbit tracking and satellite remote sensing data recording successfully.     

Virtis: An Imaging Spectrometer for the Rosetta Mission

The VIRTIS (Visual IR Thermal Imaging Spectrometer) experiment has been one of the most successful experiments built in Europe for Planetary Exploration. VIRTIS, developed in cooperation among Italy, France and Germany, has been already selected as a key experiment for 3 planetary missions: the ESA-Rosetta and Venus Express and NASA-Dawn. VIRTIS on board Rosetta and Venus Express are already producing high quality data: as far as Rosetta is concerned, the Earth-Moon system has been successfully observed during the Earth Swing-By manouver (March 2005) and furthermore, VIRTIS will collect data when Rosetta flies by Mars in February 2007 at a distance of about 200 kilometres from the planet. Data from the Rosetta mission will result in a comparison – using the same combination of sophisticated experiments – of targets that are poorly differentiated and are representative of the composition of different environment of the primordial solar system. Comets and asteroids, in fact, are in close relationship with the planetesimals, which formed from the solar nebula 4.6 billion years ago. The Rosetta mission payload is designed to obtain this information combining in situ analysis of comet material, obtained by the small lander Philae, and by a long lasting and detailed remote sensing of the comet, obtained by instrument on board the orbiting Spacecraft. The combination of remote sensing and in situ measurements will increase the scientific return of the mission. In fact, the “in situ” measurements will provide “ground-truth” for the remote sensing information, and, in turn, the locally collected data will be interpreted in the appropriate context provided by the remote sensing investigation. VIRTIS is part of the scientific payload of the Rosetta Orbiter and will detect and characterise the evolution of specific signatures – such as the typical spectral bands of minerals and molecules – arising from surface components and from materials dispersed in the coma. The identification of spectral features is a primary goal of the Rosetta mission as it will allow identification of the nature of the main constituent of the comets. Moreover, the surface thermal evolution during comet approach to sun will be also studied.     

Validation of techniques for space based remote sensing of auroral precipitation and its ionospheric effects

Knowledge of the spatial distribution of auroral precipitation and its associated ionospheric effects is important both to scientific studies of the Earth's environment and successful operation of defense and communication systems. Observations with the best spatial and temporal coverage are obtained through remote sensing from space-based platforms. Various techniques have been used, including the detection of visible, ultraviolet and X-ray emissions produced by the precipitating particles. Interpretation of the measurements is enabled through theoretical modeling of the interaction of precipitating particles with atmospheric constituents. A great variety of auroral precipitation exists, with each kind differing in the type and energy distribution of the particles, as well as in its spatial and temporal behavior. Viable remote sensing techniques must be able to distinguish at least the species of particle, the total energy flux, and the average energy. Methods based on visible, ultraviolet and X-ray emissions meet these requirements to varying degrees. These techniques and the associated space instrumentation have evolved in parallel over the last two decades. Each of the methods has been tested using simultaneous measurements made by space-based imaging systems and ground-based measurements made by radars and optical instruments. These experiments have been extremely helpful in evaluating the performance and practicality of the instruments and the results have been crucial in improving instrument design for future remote sensing platforms. The next decade will see continued development and test of remote sensing instruments and the measurements, in addition to providing important operational data, will be increasingly more critical in addressing a number of scientific problems in auroral and atmospheric physics.     

基于遥感图像分块直线特征检测的机场跑道检测方法

针对遥感图像机场跑道检测问题,提出了一种基于图像分块直线特征检测的机场跑道检测方法。首先,针对遥感图像数据量大带来的计算处理问题,设计了基于直线分割检测子（LSD）的遥感图像分块直线特征检测环节;然后,在总结归纳机场跑道数学特性的基础上,对提取的直线特征进行平行线分组、直线生长、平行线合并,并以Radon变换为基础,找出候选机场跑道区域;最后,使用灰度统计信息并结合梯度方向直方图对候选区域进行处理,筛选出最终的机场道路区域。实验结果表明,在能够提取出有效直线特征的情况下,该方法可以对多类机场跑道进行有效定位。     

基于Rough/SVM的飞机目标识别模型

研究了基于粗糙集（Rough）和支持向量机（SVM）的模式分类技术,结合遥感图像中军用飞机目标识别进行理论分析和试验研究。基于粗糙集和支持向量机算法的优势,设计了基于粗糙集和决策有向无环图的支持向量机模式识别分类器,并对分类的性能进行了分析研究。利用粗糙集理论中属性约简方法去除冗余属性,降低飞机特征维数,提出了使用决策...     

A multi-index assessment method for evaluating coverage effectiveness of remote sensing satellite

This paper deals with the multi-index assessment method for evaluating coverage effectiveness of remote sensing satellite. Because a series of satellite activities such as imaging, moving target observation, and environment investigation need to know the coverage characteristics, the coverage capability of the remote sensing satellite is the most important index. Thus, it is very important to establish the method of effectiveness evaluation of coverage characteristics. This paper focuses on the assessment of coverage effectiveness of remote sensing satellite, and proposes and designs a multi-index evaluation method based on index weight using entropy weight method and analytic hierarchy process. With a simulation case, the effectiveness evaluation results of single satellite coverage and multi-satellite coverage performance are given for the proposed assessment. The experimental results show that the established coverage characteristic model and the proposed assessment method are effective and right.     

基于LSTM的岛礁战场气象环境远程感知技术

复杂的地理环境和海洋环境导致岛礁战场气象环境瞬息万变,直接影响岛礁作战指挥决策和武器装备的作战效能。为了实时准确感知岛礁战场各个点位的气象环境信息,实现指挥平台对岛礁战场气象环境的实时监控和有效预测,文章对远程感知技术和温度预测方法进行了研究,并依托 Labwindows/CVI软件平台开发了一套岛礁战场气象环境远程感知软件,软件可实时感知岛礁战场的多种气象环境数据,并根据 LSTM模型进行温度参数的短时预测。多次实地运行测试证明,软件的设计满足预定需求,温度预测结果与实测结果误差在要求范围内。     

FTIR发射光谱遥测飞机发动机排放指数

根据飞机发动机排放测量的要求,针对传统测试系统复杂、技术难度大、成本高和实时性差的缺点,引入傅里叶变换红外（FTIR）被动遥测技术,使用FTIR发射光谱对飞机发动机排放气体组分进行定量分析,计算出排放指数。介绍了遥感测量原理和大气辐射的一个3层计算模型。采用Tensor27在无人机测试台获得某型发动机的一系列单通道光谱,计算出在14%,26%,50%,69%最大推力下,对应CO的排放指数分别为40.00,30.70,6.67,2.56 g/kg,对应NO的排放指数分别为0.79,2.09,8.24,38.70 g/kg。结果表明,FTIR发射光谱技术在飞行中的飞机发动机排放测试中具有巨大的应用前景。     

一种基于 KFCM的 SVM遥感图像机场目标分类算法

在遥感图像机场目标分类方面,支持向量机(SVM)有着广泛的应用,但由于样本不平衡问题以及不确定性数据的存在,传统SVM算法的分类精度与效果还无法令人满意。为提高传统SVM分类器的性能,文章将建立在模糊理论基础上的模糊核C-均值聚类算法(KFCM)用于处理遥感数据的不确定性问题,并通过聚类分析后的目标子图,剔除非目标样本的同时保留了目标样本,较好地解决了样本不平衡问题。将基于KFCM的SVM分类算法用于遥感图像机场目标的分类,实验结果和性能分析表明该算法分类性能优于传统SVM算法。     

高光谱图像压缩感知投影与复合正则重构

压缩感知理论提供了一种新的数据获取和压缩思路,能有效地把计算负担从编码端转移到解码端。高光谱数据具备数据稀疏性、空间相关性和谱间相关性,结合这3类先验知识,提出了一种基于复合正则化的高光谱图像压缩感知投影与重构方法。该方法的编码端只需要一个简单的投影操作;在重构算法实现中,基于变量分裂的思想,把具备多个正则项的优化问题转化成多个简单的优化问题,并用迭代方式求解。实验结果表明,本文算法在高光谱图像重构上能获得更高的峰值信噪比和更好的重构效果。该方法具备极低的编码复杂度,适用于资源受限的机载和星载高光谱成像平台。     

The Indian Space Program

The Indian Space Program is described. The main objectives of the program are to provide operation space services to the nation, especially in the fields of communications and remote sensing, and to use modern space technology for the benefit of the Indian people. Some applications of the Indian Space Program are remote sensing, imagery, communications, broadcasting, and surveys of natural resources for water, crop, forest, land, minerals, and ocean. The emphasis is on the development and operation of indigenous satellites and launch vehicles for providing these space services