卫星图像采集任务调度系统的设计与实现

随着卫星遥感技术的高速发展 ,卫星图像在众多领域有了广泛的应用。文章给出了一个卫星图像采集任务调度系统的总体结构及其功能 ,讨论了该系统实现过程中的卫星资源描述、卫星图像采集任务描述、调度引擎、卫星时间窗口的仿真计算、调度结果的可视化展现等关键技术 ,归纳了系统的特点 ,并总结了今后系统的研究方向     

基于人类视觉特性的遥感图像优化显示及应用

在分析遥感图像主观质量（Subjectivequality）评价的基础上,提出基于人类视觉特性的优化显示技术．并应用于遥感图像压缩、相对辐射校正评价和噪声估计。结果表明：当图像以50％比例放大时,处理前后的效果不明显：以100％比例放大时,差异性开始出现;当显示比例提高到200％时,图像的噪声、条纹及退化现象比较容易发现与识别。因此,当需要比较两幅图像差异及监测图像质量变化时,建议显示比例设为200％。上述成果已成功应用于CBERS-02B卫星HR相机及后续高分辨率相机压缩方案的评价,也用于中国卫星遥感地面处理系统的图像质量监测。     

结合邻域聚类分割的高光谱图像异常检测支持向量数据描述方法

支持向量数据描述方法在高光谱图像小异常目标检测中具有较好的检测性能,但是待检异常的几何形状受到约束和背景的选择具有盲目性影响检测效果,且检测需要对整幅图像进行遍历导致计算量大。提出邻域聚类分割和支持向量数据描述相结合的异常检测方法,首先利用邻域聚类方法分割图像,将几何尺寸小的分割块作为潜在异常目标;其次选择与潜在异常的形状和大小相适应的背景窗进行背景像元收集;最后采用SVDD方法从潜在异常中快速且准确地检测出异常目标。对HYMAP图像的实验结果表明,该算法提高了复杂地物背景下异常的检测性能,降低了SVDD用于高光谱图像异常检测的计算量。     

利用机载GNSS-R的有效波高反演技术

采用自主研发的多源多宿GNSS-R(Global Navigation Satellite System-Reflections,全球导航卫星系统反射信号)海洋环境遥感探测系统,在天津渤海湾特定区域进行了3次机载校飞试验,获得了大量有效GNSS-R信号数据,对自然条件下的海面有效波高进行了反演处理。该系统通过对接收的GNSS直达信号和海面反射信号进行相关运算处理,可以获得DDM(Delay-Doppler Mapping,时延—多普勒映射)图谱,再利用DCF(Derivative of the Correlation Function,相关函数的导数)在DDM图谱基础上进行相关功率波形的计算,进而通过计算DCF的函数波形宽度得到海面有效波高,验证了该方法的可行性。该方法具有全天时、全天候、高分辨率、低成本等优点,对全球范围内中小尺度海洋状态和物理现象的监测具有重要意义。     

Expert Algorithmic and Imagistic Problem Solving Strategies in Advanced Chemistry

Abstract

Visualization and imagistic reasoning appear central to expert practice in science; however, expert use of these strategies on authentic tasks has not been examined in detail. This study documents how science experts use both algorithms and imagistic reasoning to solve problems. Using protocol analysis, we report expert chemists' preferential use of algorithms for solving spatial problems and imagistic reasoning for deducing spatial transformations. We observed experts employ algorithms to solve the majority of spatial tasks while reserving imagistic strategies to solve a class of tasks that required translating between representations. Strategy used varied widely among experts and tasks.     

Human Four-Dimensional Spatial Judgments of Hyper-Volume

The dimensionality limitation of human spatial representations has been a long-lasting, unsolved issue in psychology, mathematics, and philosophy. The present study examined the possibility of human four-dimensional spatial representations using a spatial judgment task on hyper-volume, a novel property unique to higher dimensional space. Observers studied visual simulations of random wireframe hyper-tetrahedrons (4-simplexes) rotating around the y-z plane and judged their hyper-volume by adjusting the size of a 4-D block. Multiple regression analyses showed a significant correlation between the responses and the actual hyper-volume but not the definition-based, lower-dimensional cues such as the mean 3-D volume, providing empirical evidence of human 4-D spatial representations that can support judgments of certain novel, high-dimensional properties.     

用TH神经网络方法外推数据的超分辨雷达成象

研究用Ｔａｎｋ－Ｈｏｐｆｉｅｌｄ神经网络（ＴＨＮＮ）求解ＡＲ模型参数作数据外推的超分辨雷达成象,并用微波暗室实测数据对ＴＨＮＮ方法和Ｂｕｒｇ方法作了验证,结果表明,两种方法均能在较低的信噪比条件下实现超分辨成象,且随着ＶＬＳＩ技术的发展,神经网络方法将是一种很有希望的超分辨成象方法。     

超分辨成象信号处理的硬件实现

简要介绍线性预测数据外推的超分辨距离－多普勒成象方法,并以一片高速数字信号处理（ＤＳＰ）芯片—ＴＭＳ３２０Ｃ２５为核心设计了一个专用信号处理器,能完成数据外推、加窗和１２８复数点ＦＦＴ,实现超分辨二维成象。文中给出了对Ｂ－５２模型飞机微波暗室转台实测数据的成象结果,并作了比较。     

一种高光谱图像的双压缩感知模型

高光谱图像因其海量数据性,给存储、传输及后续分析处理带来了挑战。压缩感知理论提供了一种全新的信号采集框架。针对高光谱数据的三维特性,提出一种双压缩感知的采样与重构模型。该模型在采样阶段兼顾高光谱数据的空间和谱间稀疏特性,构造了能同时实现空间和谱间压缩采样的感知矩阵;重构阶段不同于传统的压缩感知重构方法直接重构高光谱数据,而是将高光谱数据分离成端元和丰度分别进行重构,然后利用重构的端元和丰度信息合成高光谱数据。实验结果表明,所提双压缩感知在低采样率下重构精度较三维压缩采样提高了10 dB以上,更为显著的是运算速度提升了3个数量级,同时该方法还便于获得端元和丰度信息。     

Aerial-BiSeNet: A real-time semantic segmentation network for high resolution aerial imagery

The aircraft system has recently gained its reputation as a reliable and efficient tool for sensing and parsing aerial scenes. However, accurate and fast semantic segmentation of high-resolution aerial images for remote sensing applications is still facing three challenges: the requirements for limited processing resources and low-latency operations based on aerial platforms, the balance between high accuracy and real-time efficiency for model performance, and the confusing objects with large intra-class variations and small inter-class differences in high-resolution aerial images. To address these issues, a lightweight and dual-path deep convolutional architecture, namely Aerial Bilateral Segmentation Network (Aerial-BiSeNet), is proposed to perform real-time segmentation on high-resolution aerial images with favorable accuracy. Specifically, inspired by the receptive field concept in human visual systems, Receptive Field Module (RFM) is proposed to encode rich multi-scale contextual information. Based on channel attention mechanism, two novel modules, called Feature Attention Module (FAM) and Channel Attention based Feature Fusion Module (CAFFM) respectively, are proposed to refine and combine features effectively to boost the model performance. Aerial-BiSeNet is evaluated on the Potsdam and Vaihingen datasets, where leading performance is reported compared with other state-of-the-art models, in terms of both accuracy and efficiency.