| 张治英,徐德忠,孙志东,张波,周晓农,周云,龚自力,刘士军.遥感图像非监督分类分析江宁县江滩钉螺孳生地植被特征[J].中华流行病学杂志,2003,24(4):261-264 |
| 遥感图像非监督分类分析江宁县江滩钉螺孳生地植被特征 |
| Unsupervised classification of remote sensing image analysis of jiangning county river beach oncomelania breeding grounds of vegetation characteristics |
| 收稿日期:2002-12-23 出版日期:2014-09-15 |
| DOI： |
| 中文关键词: 血吸虫病 遥感 钉螺 非监督分类 |
| 英文关键词: schistosomiasis Remote sensing snails Unsupervised classification |
| 基金项目:全军“十五”指令性课题资助项目( 0 1L0 78) |
|
| 摘要点击次数: 2095 |
| 全文下载次数: 919 |
| 中文摘要: |
| 目的 从遥感图像中提取江宁县江滩钉螺孳生地的植被特征。方法 在ERDASIMAGINE 8.5软件支持下,根据LANDSAT7ETM图像中各波段的光谱特征,选择对植被敏感的 2、3、4波段的伪彩色复合图像ETM2 34进行非监督分类,并通过计算分离率 (TD值 )对分类效果进行评价 ;同时将分类结果图与江宁县江滩钉螺孳生地矢量图重叠,提取江宁县江滩钉螺孳生地的植被特征,并分析钉螺在各植被中的分布。结果 通过对ETM2 34进行非监督分类,可将江宁县江滩地表特征分为包括水、沙滩、裸露地表等非植被区域在内的 1 0种地表类型 ;对分类分离率评价显示,虽然总体分离率较好 (TD =1 860 ),但存在类间的混杂错分。为了进一步提高分类效率,以植被指数为参考滤掉ETM2 34图像中无植被的像素然后再进行分类,消除了非相邻类间的混杂错分。提取江宁县江滩钉螺孳生地植被类型发现,钉螺孳生地的植被主要为非监督分类的第 3(C3)、5(C5)、6(C6)三类,且钉螺密度以C3<C5<C6,实地考察显示该三类分别为植被稀疏的滩面、杂草丛及芦苇滩。b>结论 通过对遥感图像适当分析能将江滩钉螺孳生地的植被种类有效地鉴别出来,有利于对钉螺孳生地的监测,为钉螺的有效防制及血吸虫病的预防提供依据 |
| 英文摘要: |
| Purpose Extracted from remote sensing image jiangning county river beach vegetation characteristics of oncomelania breeding sites.Methods in ERDASIMAGINE 8.5 software support, according to the spectral characteristics of LANDSAT7ETM image of each band, choose sensitive to vegetation of 2, 3, 4 band of false color composite image ETM2 34 unsupervised classification, and by calculating the separation rate (TD) to evaluate classification effect;Figure at the same time the classification result with the jiangning county river beach oncomelania breeding grounds vector graphics overlap, extraction of jiangning county river beach vegetation characteristics of oncomelania breeding sites, and analyzes the oncomelania snail distribution in each vegetation.Results through the study of the unsupervised classification of ETM2 34, jiangning county river beach surface features can be divided into such as water, sand, bare surface vegetation area, 1 0 types of surface;Separative rate of classification evaluation shows that, although the overall separation rate is better (TD = 1, 860), but there is mixed fault points between the classes.In order to further improve the efficiency of classification to the vegetation index as a reference filter out ETM2 34 image pixels without vegetation then classify, eliminates the jumble of wrong points between adjacent class.Extraction of jiangning county river beach oncomelania breeding grounds of vegetation found that oncomelania breeding grounds of vegetation mainly for unsupervised classification 3 (C3), 5 (C5), 6 (C6) three categories, and the density of snail with C3C5C6, on-the-spot investigation showed that the three types of vegetation sparse beach face respectively, miscellaneous grass and reeds beach.Conclusion through the analysis of remote sensing image appropriate to river beach oncomelania breeding grounds of vegetation species identified effectively, is advantageous to the monitoring of oncomelania breeding sites, for the effective of snail control and schistosomiasis prevention provide the basis |
|
查看全文
Html全文
查看/发表评论 下载PDF阅读器 |
|
| 关闭 |
|
|
|
