期刊文献+

基于多种差异影像算法改进回溯土地覆被变化检测方法的探讨 预览

Discussion on Improving Backtracking Land Cover Change Detection Method Based on Multi-difference Image Algorithm
在线阅读 下载PDF
分享 导出
摘要 近年来,回溯分析方法已经越来越多地应用于自动化变化分析和分类研究中。以兰州市2008和2017年两期Landsat影像为例,分别采用基于PCA、基于CVA和基于FCM的回溯分析方法获得兰州市2008-2017年间的土地覆被分布和变化情况并对其结果进行对比,分析了不同的差异影像算法在回溯分析中的适应性。结果表明基于FCM的回溯分析方法的结果最好,分类总体精度达到88.5%,变化检测结果总体精度达到83.8%;基于CVA的回溯分析方法次之,因存在漏检现象导致分类总体精度为86%,变化检测结果总体精度不足80%;而基于PCA的回溯分析方法精度最低。通过采用FCM算法有效改善土地覆被变化检测结果中的漏检现象,从而获得精度较高的分类及变化检测结果。 In recent years,backtracking analysis methods have been more and more used in automatic change analysis and classification research.Taking the Landsat images of Lanzhou City in 2008 and 2017 as an example,the distribution and change of land cover in Lanzhou City from 2008 to 2017 were obtained by using the backtracking analysis methods based on PCA,CVA and FCM respectively,and the results were compared.The adaptability of different differential image algorithms in backtracking analysis is analyzed.The results show that the backtracking analysis method based on FCM has the best results,the overall accuracy of classification is 88.5%,and the overall accuracy of change detection results is 83.8%,followed by the backtracking analysis method based on CVA.Because of the phenomenon of missed detection,the overall accuracy of classification is 86%,the overall accuracy of change detection results is less than 80%,and the accuracy of backtracking analysis method based on PCA is the lowest.By using FCM algorithm to effectively improve the phenomenon of missed detection in the results of land cover change detection,the classification and change detection results with high accuracy can be obtained.
作者 蒋月 邢瑾 何江 刘巨峰 JIANG Yue;XING Jin;HE Jiang;LIU Jufeng(不详)
出处 《科技创新与应用》 2019年第7期1-7,共7页 Technology innovation and application
关键词 基于对象影像分析 差异影像 回溯分析 变化检测 对比研究 object-based image analysis differential image backtracking analysis change detection comparative study
作者简介 蒋月(1994-),女,河南信阳人,硕士研究生,主要从事基于对象影像分析方法的研究。
  • 相关文献

参考文献14

二级参考文献193

共引文献118

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部 意见反馈