首页> 外文学位 >Highway extraction from high resolution aerial photography using a geometric active contour model.
【24h】

Highway extraction from high resolution aerial photography using a geometric active contour model.

机译:使用几何主动轮廓模型从高分辨率航拍中提取高速公路。

获取原文
获取原文并翻译 | 示例

摘要

Highway extraction and vehicle detection are two of the most important steps in traffic-flow analysis from multi-frame aerial photographs. The traditional method of deriving traffic flow trajectories relies on manual vehicle counting from a sequence of aerial photographs, which is tedious and time-consuming. This research presents a new framework for semi-automatic highway extraction.; The basis of the new framework is an improved geometric active contour (GAC) model. This novel model seeks to minimize an objective function that transforms a problem of propagation of regular curves into an optimization problem. The implementation of curve propagation is based on level set theory. By using an implicit representation of a two-dimensional curve, a level set approach can be used to deal with topological changes naturally, and the output is unaffected by different initial positions of the curve. However, the original GAC model, on which the new model is based, only incorporates boundary information into the curve propagation process. An error-producing phenomenon called leakage is inevitable wherever there is an uncertain weak edge. In this research, region-based information is added as a constraint into the original GAC model, thereby, giving this proposed method the ability of integrating both boundary and region-based information during the curve propagation. Adding the region-based constraint eliminates the leakage problem.; This dissertation applies the proposed augmented GAC model to the problem of highway extraction from high-resolution aerial photography. First, an optimized stopping criterion is designed and used in the implementation of the GAC model. It effectively saves processing time and computations. Second, a seed point propagation framework is designed and implemented. This framework incorporates highway extraction, tracking, and linking into one procedure. A seed point is usually placed at an end node of highway segments close to the boundary of the image or at a position where possible blocking may occur, such as at an overpass bridge or near vehicle crowds. These seed points can be automatically propagated throughout the entire highway network. During the process, road center points are also extracted, which introduces a search direction for solving possible blocking problems. This new framework has been successfully applied to highway network extraction from a large orthophoto mosaic. In the process, vehicles on the highway extracted from mosaic were detected with an 83% success rate.
机译:高速公路提取和车辆检测是多帧航拍照片交通流分析中最重要的两个步骤。推导交通流轨迹的传统方法依赖于从一系列航空照片中进行手动车辆计数,这既繁琐又费时。该研究提出了一种半自动公路提取的新框架。新框架的基础是改进的几何活动轮廓(GAC)模型。这个新颖的模型试图最小化将规则曲线传播问题转化为优化问题的目标函数。曲线传播的实现基于水平集理论。通过使用二维曲线的隐式表示,可以使用水平集方法自然地处理拓扑变化,并且输出不受曲线的不同初始位置的影响。但是,新模型所基于的原始GAC模型仅将边界信息合并到曲线传播过程中。每当存在不确定的弱边缘时,就不可避免地会产生称为泄漏的错误产生现象。在这项研究中,将基于区域的信息作为约束添加到原始GAC模型中,从而使此提议的方法能够在曲线传播期间集成边界和基于区域的信息。添加基于区域的约束消除了泄漏问题。本文将提出的增强GAC模型应用于高分辨率航空摄影中的高速公路提取问题。首先,设计了优化的停止标准,并将其用于GAC模型的实现中。它有效地节省了处理时间和计算量。其次,设计并实现了种子点传播框架。该框架将高速公路提取,跟踪和链接合并到一个过程中。通常将种子点放置在靠近图像边界的高速公路路段的末端节点上,或者放置在可能发生阻塞的位置,例如天桥或车辆拥挤的位置。这些种子点可以自动传播到整个高速公路网络。在此过程中,还将提取道路中心点,这会引入搜索方向以解决可能的阻塞问题。这个新框架已成功应用于从大型正射影像马赛克中提取高速公路网络。在此过程中,从马赛克中提取的高速公路上的车辆被检测到,成功率为83%。

著录项

  • 作者

    Niu, Xutong.;

  • 作者单位

    The Ohio State University.;

  • 授予单位 The Ohio State University.;
  • 学科 Geodesy.; Remote Sensing.
  • 学位 Ph.D.
  • 年度 2004
  • 页码 146 p.
  • 总页数 146
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 大地测量学;遥感技术;
  • 关键词

  • 入库时间 2022-08-17 11:44:23

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号