一种遥感图像建筑物检测新方法

遥感图像中的建筑物检测对于土地规划和地图绘制等具有重要意义。文章针对高分辨率全色遥感图像植被覆盖中隐蔽建筑物检测问题,提出了一种结合屋顶结构信息和纹理信息的快速房屋检测方法。首先利用屋顶边缘的几何关系寻找具有矩形屋顶的建筑物;然后使用形态学方法提取屋顶较亮的建筑物;最后利用局部二值模式（Local Binary Pattern,LBP）描述屋顶的纹理特征,并利用支持向量机（Support Vector Machine,SVM）排除误检区域。通过在全色遥感图像集上的试验证明,所提方法对高分辨率遥感图像植被覆盖区域中的建筑物目标具有较高检测率和较低误检率。     

SAR图像中道路检测方法研究

提出一种两步算法用于合成孔径雷达（SAR）图像中无监督地提取线性特征，特别是公路网的提取。第一步利用局部线特征检测算子提取线基元，为此本文提出一种改进的均值比率（RoA）算子，应用SAR图像的统计特性，保证较高的准确率；第二步引入公路结构特征的先验信息，构造方向势能函数（DPF）。通过DPF最小化，来对提取的线基元进行组织，获得宏观的线性结构，文中方法通过ERS－1图像加以验证。     

基于U形结构提取的港口目标检测研究

遥感图像中港口目标的检测是目标识别研究的重要方面,准确、高效地自动提取港口目标对于提高遥感图像自动解译能力具有重大意义。文章提出一种基于U形结构单元提取的遥感图像港口目标检测的方法。首先利用梯度算子和灰度相似性聚类分割海洋和陆地区域并提取粗略海岸线,再根据内港区域的封闭特性提取疑似区域,在分析内港区域结构的基础上,引入U形结构元的概念以及提取算法来对疑似区域进行确认。试验表明,该方法检测的港口目标准确率较高,速度较快,能够满足民用和军事识别的需求。     

遥感图像中基于Hough变换的直线提取算法

针对遥感图像中大型目标的直线特征提取问题,设计了一种基于类直线提取的改进霍夫(Hough)变换算法。类直线的提取相当于Hough变换的预处理,既约束了Hough变换所使用的直线上的特征点,改进了直接进行Hough变换时不同直线的特征点互相干扰的缺点,又降低了计算量。对图像空间的每条类直线分别进行Hough变换,改进投票过程的映射方式,寻找类直线中最多特征点所在的那条直线,只取一个投票峰值,在减小了计算复杂度的同时又去掉了虚假峰值的影响。实验结果表明,改进Hough变换直线特征提取算法性能好、效率高,可用于遥感图像处理领域。     

舰空导弹武器系统战术技术指标结构研究

通过对航空导弹武器系统使命，任务，结构和功能的分析，提出了航空导弹武器系统技术指标的基本结构和体系，研究了各结构包含的主要战术技术指标并给出定义，分析了影响和决定武器系统战术技术指标的主要因素和影响效果，给出了航空导弹是武器系统战术技术指标体系图，为全国描述航空导弹武器系统的战术技术性能提供了依据。     

一种用于无人机航空相机标定的角点检测算法

在实现无人机航空相机的高精度标定中,模板图像中角点提取的精度是至关重要的因素。为了实现高精度角点的提取,以及亚像素角点像素的修正,提出了基于Hessian矩阵的X型角点形状算子的亚像素角点提取方法。实验分析验证,该方法提取精度较高,提取过程简单,计算量小,大大简化了标定过程。     

色彩直方图混合特征提取在海量图像数据检索中的应用

随着多媒体技术的发展,图像的应用越来越广泛,如何有效利用图像信息具有越来越重要的意义。颜色特征不受图像大小和旋转角度的影响,鲁棒性较好,在图像检索中具有较强的应用。本文在对图像特征进行提取的基础上,提出了一种新的图像检索数据方法。该方法引入了色彩直方图、颜色中心矩、灰度直方图等颜色特征,使用色彩直方图对整幅图像的色彩信息进行统计;使用分块颜色中心矩对色彩的空间分布信息进行统计;对由于色彩浓度和对比度发生变化而对检索效果的影响,采用了灰度直方图进行淡化。通过对这三种特征距离进行加权和计算,进行相似度匹配,检索出目标图像。本文针对性地设计了基于色彩直方图混合特征的图像检索系统,实验证明色彩直方图混合特征检索的结果与示例图像在颜色与空间分布上基本相似,具有较强的实用性。     

Hough变换在图像匹配中的应用

针对图像匹配中实时图存在旋转失真的问题,提出一种利用Hough变换提取图像中的直线特征及该直线的倾斜角,对实时图中旋转失真进行校正的方法。实验结果证明,该方法切实可行,能将图像匹配速度提高一个数量级,便于实时处理,同时其匹配精度依然较高。     

基于标准固体火箭发动机CT结构图像配比的缺陷自动提取方法

针对固体火箭发动机CT检测图像因存在算法上的伪影和噪声影响缺陷识别这一问题,提出基于标准固体火箭发动机CT结构图像配比的缺陷自动提取方法。通过缺陷对CT成像影响的理论分析和固体火箭发动机缺陷配比自动提取的仿真和实验验证,得出该方法中配比图像分割最佳阈值为30,以此阈值能有效、准确地提取固体火箭发动机CT图像缺陷。     

基于数据扩充和特征增强的雷达回波无人机检测

针对基于雷达无人机目标识别难度高、精确度低、适应性差等问题，本文提出了利用深度学习的方法，对雷达回波进行无人机检测。首先，利用相参累积的方法生成雷达回波的距离-多普勒图像，增强目标特征并提高信噪比;其次，采用生成对抗模型对距离-多普勒图像进行数据扩充，以获得充足的图像数据减小网络的过拟合并提高网络鲁棒性;最后，使用基于位置感知的卷积神经网络增强特征，通过构建基于距离-多普勒图像的感知模块，实现对目标距离和运动速度的检测。通过在雷达回波序列中弱小飞机目标检测跟踪数据集上验证的结果表明:最终检测结果在召回率89%的情况下达到了91%的准确率。相比于基准方法，本文提出的方案具有更高的检测精度和更好的网络运行效率。     

Damage assessment of large space structures through the variational approach

The present investigation is focused on the solution of a dynamic inverse problem which is concerned with the assessment of damage in large space structures by means of measured vibration data. This inverse problem has been presented as an optimization problem and has been solved through the use of the conjugate gradient method with the adjoint equation, also called the variational approach. When a high number of damaged elements has to be found and these elements are also severely damaged, it is shown that the use of an additional method is necessary in order to provide a better initial guess for the conjugate gradient method. A stochastic method, represented by the genetic algorithm method, has been chosen because it provides robust search in complex spaces and also reduces the chance of converging to local optima. The application of this hybrid approach showed that better results can be achieved, although the computational time for the application analyzed here could increase. The damage estimation has been evaluated using noiseless and noisy synthetic experimental data, and the reported results are concerned with a space truss structure.     

基于多结构元素的遥感图像去噪及边缘检测方法

边缘检测是遥感图像处理的一个重要方面。利用数学形态学原理，提出了一种基于多结构元素的遥感图像边缘检测方法。首先采用基于视觉模型的边缘阈值得到灰度突变像素点，然后利用几种具有代表性的结构元素对这些像素点进行形态学腐蚀操作以区分真实像点和噪声点，将噪声点去除并通过伪中值滤波填补像素值，最后得到图像边缘。实验结果表明，该算法能够在保持图像边缘细节的前提下，很好地滤除图像中的孤立噪声点，在检测精度和抗噪声性能方面都优于传统的形态学算法。     

基于梯度特征的图像自动分割方法

利用遥感图像进行海上目标检测、监视的关键技术是实现海陆分离，从而为目标检测、监视减少不必要的处理工作。传统的采用阈值分割方法分离陆地，但是这种方法对于陆地上灰度级较低区域容易造成误分割。而且，很显然的不能够充分利用图像中的灰度变化信息，即梯度。为此提出了一种基于梯度特征的海陆分割算法。首先利用边缘检测提取图像的梯度特征，梯度特征对灰度级并不敏感而对梯度的变化敏感，对于陆地灰度级较低的区域也能够提取边缘。对得到的边缘进行形态学处理，利用设定的面积阈值将图像分为主体陆地区域和孤立陆地区域，进一步可以屏蔽陆地。实验结果表明，该方法可以较好的分割陆地中阴暗区域，对于未经辐射校正图像有较好的适用性，并且有较快的处理速度。     

Experimentally tracing the key steps in the origin of life: The aromatic world

Life is generally believed to emerge on Earth, to be at least functionally similar to life as we know it today, and to be much simpler than modern life. Although minimal life is notoriously difficult to define, a molecular system can be considered alive if it turns resources into building blocks, replicates, and evolves. Primitive life may have consisted of a compartmentalized genetic system coupled with an energy-harvesting mechanism. How prebiotic building blocks self-assemble and transform themselves into a minimal living system can be broken into two questions: (1) How can prebiotic building blocks form containers, metabolic networks, and informational polymers? (2) How can these three components cooperatively organize to form a protocell that satisfies the minimal requirements for a living system? The functional integration of these components is a difficult puzzle that requires cooperation among all the aspects of protocell assembly: starting material, reaction mechanisms, thermodynamics, and the integration of the inheritance, metabolism, and container functionalities. Protocells may have been self-assembled from components different from those used in modern biochemistry. We propose that assemblies based on aromatic hydrocarbons may have been the most abundant flexible and stable organic materials on the primitive Earth and discuss their possible integration into a minimal life form. In this paper we attempt to combine current knowledge of the composition of prebiotic organic material of extraterrestrial and terrestrial origin, and put these in the context of possible prebiotic scenarios. We also describe laboratory experiments that might help clarify the transition from nonliving to living matter using aromatic material. This paper presents an interdisciplinary approach to interface state of the art knowledge in astrochemistry, prebiotic chemistry, and artificial life research.     

地基光学图像中本体-帆板结构航天器姿态估计方法

提出一种利用地基光学图像估计空间目标姿态的方法。针对本体-帆板结构航天器外形结构特点，定义了一种不易遮挡的两轴结构特征，设计了地基光学图像中该特征的提取方法；根据航天器成像过程建立了特征点二三维几何投影方程，实现了航天器三维姿态的解算。利用单站序列图像，可获得航天器姿态变化角速度信息。在轨航天器的试验结果表明，本方法能有效地从地基光学图像估计出航天器姿态。     

关于太阳耀斑观测的地图匹配算法实现

利用海洋的镜面反射进行太阳耀斑观测,要求以星下点为中心的卫星视场必须落在海洋区内。针对如何判定卫星视场和海岸线地图的位置关系问题,文章将卫星视场构成的矩形区域与海岸线地图构成的不规则多边形区域等效为多条线段集合,并提出一种把判断矩形与不规则多边形位置关系转化成判断等效线段之间位置关系的算法。该算法与传统的线段裁剪法或多边形裁剪法比较,具有无需求取线段交点、判断次数大幅减小等优点,能够达到快速判断能否进行耀斑观测的目的。     

复合材料层合锥壳非线性稳定性的有限元分析

本文根据稳定性分析的前屈曲完全理论和修正型的Ｈａｈｎ－Ｔｓａｉ非线性本构关系，采用有限元法分析了复合材料层合回转壳的非线性稳定性，讨论了屈曲产壳体法向与面内支承条件和复合材料的横向和剪切非线性性能对壳体结构总体稳定性的影响，为工程结构设计者们提供了一些有益的参考意见。     

航天器污染检测新技术的研究

根据飞行时间二次离子质谱学的新进展,用当代最先进的仪器首次对三种典型的我国地面空间环境模拟污染样品进行了探索性实验研究。结果表明：由于极高的检测灵敏度,具备分析微量航天污染物成分的能力;由于很高的质量分辨本领,具备检测航天污染物包含的所有元素、同位素和各种化合物的指纹鉴别能力;成像分析可获得航天污染物各成分面分布及一定的分层信息,从而可望揭示出各类污染物的形成过程。与国内外现有的各类检测手段相比,飞行时间二次离子质谱是航天污染物检测强有力的一种新手段。讨论了需要进一步研究解决的主要问题,如果能开发出可避免二次污染并分阶段取样、保存样品的装置与之配合,该技术有望在迅速崛起的航天污染系统工程中发展成为一种独具特色的航天污染检测新技术。     

Numerical investigation on the shock wave transition in a three-dimensional scramjet isolator

The scramjet isolator, which is used to prevent the hypersonic inlet from disturbances that arise from the pressure rise in the scramjet combustor due to the intense turbulent combustion, is one of the most critical components in hypersonic airbreathing propulsion systems. Any engineering error that is possible in the design and manufacturing procedure of the experimental model, and the intense heat release in the scramjet combustor, may cause the performance of the isolator to decrease, leading to its lack of capability in supporting the back pressure. The coupled implicit Reynolds Averaged Navier–Stokes (RANS) equations and the two-equation standard k?ε turbulent model have been employed to numerically simulate the flow fields in a three-dimensional scramjet isolator. The effects of the divergent angle and the back pressure on the shock wave transition and the location of the leading edge of the shock wave train have been estimated and discussed. The obtained results show that the present numerical results are in very good agreement with the available experimental shadow-pictures, and the numerical method is more suitable for capturing the shock wave train and predicting the location of the leading edge of the shock wave train in the scramjet isolator than the present two-dimensional numerical methods. This is due to the small width-to-height ratio of the isolator and the intense three-dimensional flow structures. On increasing the divergent angle of the scramjet isolator, the static pressure along the central symmetrical line of the isolator decreases sharply. This is due to the strong expansion wave generated at the entrance of the isolator, and when the divergent angle of the isolator is sufficiently large, namely 1.5°, a zone of negative pressure is formed just ahead of the leading edge of the shock wave train. At the same time, the shock wave train varies from being oblique to being normal, and then back to oblique. With an increase in the prescribed back pressure at the exit of the scramjet isolator, the leading edge of the shock wave train moves forward towards the entrance of the isolator, and when the back pressure is sufficiently large, unstart conditions in the hypersonic inlet can take place if the shock train reaches the inlet.