【作者】 张素莉； 潘欣； 李晓峰； 孟祥萍；

【Author】 Suli ZHANG1,Xin PAN1,Xiaofeng LI2,Xiangping MENG1 1School of Electrical & Information Technology,Changchun Institute of Technology,Changchun,China 2Northeast Institute of Geography and Agricultural Ecology,CAS,Changchun,China

【机构】 长春工程学院电气与信息工程学院； 中国科学院东北地理与农业生态研究所；

【摘要】 在高分辨率遥感影像上进行道路信息的准确提取对于提高规划、维护、导航及交通分析等信息化管理技术具有重要意义。以往的研究多集中于通过数学手段提取道路的特征信息尤其是边缘信息,但是如何有效利用已经获取的边缘信息并进一步提取出道路仍然是需要研究的课题。本文将多智能体技术引入到平行线段检测和道路提取中来,构造了一种带"传感器"的自主智能体,它是一种可以在Canny算子计算的边缘信息上检测平行线段并沿其方向移动的自主智能体。通过构造多智能体系统,构造大量Agent并使其在影像上的运动,实现平行线的识别与道路的提取。实验结果表明,该方法不但能适应包含不确定、噪声和断路的平行线,而且在道路的提取能力方面有着很好的表现。更多还原

【Abstract】 Road extraction from high-resolution remote sensing images is very important for improving the level of information administration including planning,road network maintaining and navigation and traffic analyzing. Previous studies usually focus on extracting road feature information through edge detection from high-resolution remote sensing images with mathematical methods,but problems of over-and underetection impede the usefulness of the extracted results. Further treatments are therefore necessary. In this paper,we introduced Multi-Agent-System into parallel lines detection and road extraction. An agent which could detect edge information obtained by canny operator,moving along the parallel lines(i.e. edges of the road) ,was introduced. By constructing a multi-agent system,a large number of Agents was running on the remote sensing image,realized the parallel line’s detection and road extraction. The experimental results showed that this method had a good performance in road extraction,adapting to uncertainties caused by noise and broken lines.更多还原