不同特征遥感图像在江滩钉螺孳生地监测中的应用  

Application of digital imageries from Landsat ETM+ to the surveillance of snail habitats in marshland

作　　者：张治英 [1] 徐德忠 [1] 周云  孙志东 [1] 张波 [1] 周晓农 [3] 刘士军 [4] 龚自立 [4] 

机构地区：[1]第四军医大学流行病学教研室,陕西西安,710032 [2]江苏省江宁县疾病预防控制中心,江苏江宁,211100 [3]中国疾病预防控制中心寄生虫病防治研究所,上海,200025 [4]南京军区联勤部防疫队,江苏南京,210014

出　　处：《西安交通大学学报（医学版）》2004年第3期304-306,309,共4页Journal of Xi’an Jiaotong University（Medical Sciences）

基　　金：全军"十五"指令性课题 (No.0 1L0 78);第四军医大学创新工程 (No .CX99F0 0 9);科技部十五攻关课题 (No .2 0 0 1BA70 5B08)

摘　　要：目的　探讨由LandsatETM +遥感图像组成的特征图像在江滩钉螺孳生地监测中的应用。方法　对以LandsatETM +图像组成的 3种特征图像进行非监督分类 ,并通过现场勘察和计算类别间分离度对分类结果进行评价。结果　以ETM的第 2、3、4波段组成的ETM2 34伪彩色复合图像的分类识别效果比较好 ,能有效地将江滩芦苇、梭叶草和杂草滩等识别出来 ,有助于江滩钉螺孳生地的监测。结论　遥感图像非监督分类能有效识别江滩地表植被的分布而用于钉螺孳生地的监测 ,但应根据研究地区的实际情况选择合适的特征图像。Objective To explore the application of digital imageries from Landsat ETM+ to the surveillance of snail habitats in marshland. Methods The feature imageries from Landsat ETM+ sensor were classified in the ERDAS IMAGINE 8.5 to analyze the vegetation types in the marshland using the unsupervised classification. The vegetation types were validated by on-the-spot survey, and the discrimination of vegetation types was evaluated using the transformed divergence in ERDAS IMAGINE 8.5. Results The ETM234 false composition from the 2, 3 and 4 brands of Landsat ETM+ could effectively identify the different vegetation types in marshland. It could differentiate the vegetation such as reed, shuttle-leaf grass, ruderal, which is suitable for the survival of schistosomiasis-transmitted snails from other vegetation. Conclusion Although the images from the unsupervised classification of Landsat ETM+ are useful for the surveillance of snail habitats in marshland, the optimal feature imagery used for the classification should be selected according to the condition of the field.

关 键 词：遥感 特征图像 非监督分类 

分 类 号：R181.13[医药卫生—流行病学]