测绘通报 ›› 2017, Vol. 0 ›› Issue (9): 28-31.doi: 10.13474/j.cnki.11-2246.2017.0281

• 学术研究 • 上一篇    下一篇

基于CRF模型的高分辨率遥感影像变化检测

魏立飞1,2, 牟紫微1,2, 王晓燕3, 李丹丹4   

  1. 1. 湖北大学资源环境学院, 湖北 武汉 430062;
    2. 区域开发与环境响应湖北省重点实验室, 湖北 武汉 430062;
    3. 华中师范大学城市与环境科学学院, 湖北 武汉 430079;
    4. 农业部农业信息技术重点实验室, 北京 100081
  • 收稿日期:2017-01-05 修回日期:2017-04-25 出版日期:2017-09-25 发布日期:2017-10-12
  • 作者简介:魏立飞(1979-),男,博士,讲师,主要研究方向为城市遥感及遥感影像智能化处理.E-mail:weilifeihb@163.com
  • 基金资助:
    国家自然科学基金(61201341);干旱气象科学研究基金(IAM201512);数字制图与国土信息应用工程国家测绘地理信息局重点实验室开放研究基金(GCWD201407);安徽省智慧城市与地理国情监测重点实验室开放基金(2016-K-02Z)。

Change Detection of High Resolution Remote Sensing Image Based on CRF Model

WEI Lifei1,2, MOU Ziwei1,2, WANG Xiaoyan3, LI Dandan4   

  1. 1. Faculty of Resources and Environmental Science, Hubei University, Wuhan 430062, China;
    2. Hubei Key Laboratory of Regional Development and Environmental Response, Wuhan 430062, China;
    3. The College of Urban & Environmental Sciences, Central China Normal University, Wuhan 430079, China;
    4. Key Laboratory of Agri-informatics, Ministry of Agriculture, P.R.China, Beijing 100081, China
  • Received:2017-01-05 Revised:2017-04-25 Online:2017-09-25 Published:2017-10-12

摘要: 针对传统高分辨率遥感影像变化检测方法大多直接利用光谱信息进行计算,导致检测精度不高的缺陷,提出了一种基于条件随机场模型(CRF)的高分辨率遥感影像变化检测方法。该方法利用CRF模型融合差值影像的光谱项和空间项,在融合过程中,引入EM迭代策略不断更新影像检测结果,提高变化检测精度。试验结果表明,本文方法的分类精度好于传统方法,并且稳定性良好。

关键词: 高分辨率遥感影像, 变化检测, CRF模型, 迭代计算

Abstract: For the traditional high resolution remote sensing image change detection methods mostly using spectral information to calculate directly,thus the accuracy of the detection result is not high,this paper proposes a high resolution remote sensing image change detection method based on conditional random field model(CRF).The method fuses the spectrum and space of difference image with CRF model,in the process of fusion,introducing the EM iterative strategy constantly updated the detection result,improve the change detection accuracy.The experimental results show that the classification accuracy of this method proposed in this paper is better than the traditional method,and having good stability.

Key words: high resolution remote sensing image, change detection, CRF model, Iterative calculation

中图分类号: