节点文献

战场环境建模与态势生成关键技术研究

Research on Technologies of Battlefield Environment Modeling and Situation Generation

【作者】 陈鸿

【导师】 孙茂印; 汤晓安;

【作者基本信息】 国防科学技术大学 , 信息与通信工程, 2010, 博士

【摘要】 综合世界新军事变革的发展方向可以看出,未来的军队作战将充分依托信息优势来获取强大的战斗力。而传统的战场信息来源众多,展现方式多样,且相互独立,难以满足现代战争对战场环境全面实时感知的需求。必须构建数字化的战场环境,将多种类型的基础地理空间数据以及地理空间情报信息进行有效的提取、处理、整合、显示,形成统一的作战态势图,以便指挥员做出科学决策。论文在国家级科技攻关项目支持下,围绕战场环境保障技术中的战场地理环境数字化、地理与态势可视化等方面展开研究,旨在突破关键技术,设计并实现战场环境与态势生成原型系统。论文研究工作如下:针对彩色地形图上各种图形要素相互交叠以及扫描引起的颜色混淆及假彩色的问题,借鉴面向对象思想,提出了一种地形图自动分层算法。首先将地形图线划要素与普染区背景分离;然后把具有相同特征的像素分割成线段对象与块对象,使之成为最基本的分类单元;通过提取对象的颜色、形状和近邻关系等特征,建立分类标准,实现地形图的自动分层,达到地形图矢量化的目的。该算法能有效克服扫描引起的颜色混淆和假彩色的影响,分层结果完整连通,从而在很大程度上抑制了“椒盐”现象。提出了一种基于空间索引的快速DEM生成算法。首先建立四叉树索引对三角形数据集进行层次化分块管理,来快速排除无关区域,缩小检索范围;然后通过三角网拓扑关系,采用线性步进搜索策略快速定位三角形和确定约束线段影响域,有效地提高了由等高线构建约束TIN的效率;最后利用DEM格网的规则结构建立索引,快速搜索所有包含在三角形内的内插点,较大程度上提高了从TIN内插DEM的速度。提出了一种基于视点相关透视纹理的矢量数据绘制算法。首先通过场景视点得到能充分逼近当前地形可视范围的透视投影,然后用该透视投影实时地将二维矢量数据绘制在与地形绑定的纹理上。与传统的基于纹理的绘制方法相比,视点相关透视投影能够提高纹理的像素利用率,从而减轻传统方法所产生的走样问题,提高矢量数据的绘制质量,达到将矢量数据精确、高效地叠加在三维多分辨率地形上的目的;且该算法不受地形几何数据的约束,能够与目前大多数多分辨率地形模型以及影像金字塔很好地结合在一起使用。提出了一种基于位移映射的军标生成与动态标绘算法。首先在GPU片元着色器中对地形进行平滑过滤,滤除局部细节而保留地形起伏的主体趋势,将处理结果实时绘制成位移映射纹理;然后由非规则军标的控制点集生成军标二维外轮廓,用之裁剪一次NURBS表面生成平面的军标网格;最后在GPU顶点着色器中依据关联的位移映射纹理,位移军标网格上各顶点以达到地形匹配的目的。实验结果表明,该算法能够实现非规则军标的三维地形跟随,且运算效率高,能满足交互编辑以及动态推演的实时要求。设计并实现了一个战场环境与态势生成原型系统—BEMSG,构建了系统总体框架、软件子系统。通过在某次演习中的应用,验证了论文所提出的理论、方法及所构建系统的可行性、有效性与正确性。

【Abstract】 Information advantage has been considered as a crucial element to enhance battle effectiveness. There are lots of traditionary battlefield environment data. They are from different sources and have diverse formats. These environment data are separate and hard to meet the requirement of all-around real-time battlefield environment awareness in the modern war. So it is necessary to process these environment datas, extract basic spatial information, then integrate geographical spatial information and spatial intelligence together to provide a unitive battlefield situation in support of combat decision-making. With the support of national administration’s research project, this dissertation studys three aspects of digital battlefield environment supporting technology: battlefield geographical environment digitization, battlefield geographical environment visualization and battlefield situation visualization. The aim of this study is to break through the key technologies, design and realize a battlefield environment modeling and situation generation prototype system– BEMSG.The main contents and accomplishments are as follows:The vectorization of scanned maps is a difficult process due to aliasing and false colors induced by the scanning process and due to closely spaced and intersecting/overlapping features inherent to the maps. In this dissertation, an object-oriented approach is introduced into scanned map segmentation. The procedure is described as follows. First, linear elements and area elements are separated. Then they are partitioned into image objects which are basic cells to classify map features, not single pixels. And a set of features combining color, shape and topology are extracted from each object. Based on the features, a classification criterion is employed to perform the map segmentation. The proposed method overcomes aliasing and false colors induced by the scanning process. The results seem coherent, with none of the "salt and pepper" noise.A kind of DEM constructing method based on spatial index is proposed. First, quad-tree index is created to manage triangles, so as to reduce searching area. Then, the topology of triangles is established. According to it, a gradual strategy is used to locate the goal triangle and ascertain influenced area of constrainded boundary quickly. The experiment results show that the method can effectively improve the speed of constructing TIN from contours. Finally, the regular structure of grid DEM is used to create spatial index directly, which can index all grid points in each triangle rapidly. The experiment results show that the spatial index can accelerate the interpolation of grid DEM from TIN remarkably.A kind of vector data displaying algorithm based on view-dependent projective texture-mapping is proposed. First, depending on the scene view, a perspective projection is created, which can match the currently terrain visible range well. Then it is used to generate the texture on-the-fly and calculate the texture coordinates in GPU. Quality superior to standard texture mapping is achieved by using the view-dependent perspective projection, which can improve the availability ratio of texture pixels. In addition, since the algorithm is independent of the underlying terrain geometry, it is suited to work with terrain LOD and image pyramid algorithms.A kind of military symbols rendering algorithm based on displacement mapping is proposed. First, terrain is smoothed using the programmable fragment processor on GPU, to wipe off local details and reserve the main hypsography. The result is rendered to a displacement texture on the fly. Then, 2D outlines of irregular military symbols are generated, used to trim NURBS surfaces to create military symbol grids. Finally, according to the displacement texture, all vertices of the military symbol grids are displaced using the programmable vertex processor on GPU. The algorithm allows high-quality overlay of military symbols on virtual landscapes, and is GPU-based, can meet the real-time requirement of interactive edit and dynamic plotting.A prototype system– BEMSG is designed and implemented. The general framework, software subsystems of the system are constructed. The application in a manoeuvre validates the feasibility, effectiveness and correctness of the proposed theories, algorithms and the system constructed in the dissertation.

节点文献中: 

本文链接的文献网络图示:

本文的引文网络