摘要
利用2006年4月6-7日在甘肃省庆阳市境内环江、柔远河和马莲河实测的水体波谱数据,以及化学需氧量COD测定数据,采用Fisher判别分析方法,建立了基于地面实测光谱数据并适用于本研究区域的水环境COD遥感识别模式;将该模式应用于2005年10月16日和2006年4月10日过境的两景Landsat/TM图像;根据遥感模式反演结果图,对该区域水环境COD的时空变化进行了分析。水环境COD污染遥感模式的建立可为从遥感影像上快速、大面积获取COD信息提供了一种技术手段。
This work was to give an identification mode of ChemicaL Oxygen Demand （COD） within water body using remotely sensing technique. For this purpose the field data, including the spectral data of water body and the concentration of COD within water , were collected at Huanjiang river, Rouyuan river and Malian river in Qingyang city , Gansu province from April 6 to April 7,2006. The 90% samples were employed to establish the identification mode of COD according to Fisher muhiclass discriminant rule. By the actual status and the national standard, the concentration of COD were classed three level. The field spectral data were processed as corresponding bands of Landsat/TM. And the concentration of COD can be identified as the following equations
class1 = -262. 466TM1 ＋ 1377. 867TM2 -741. 706TM3 ＋ 104.5TM4 （ 1 ）
class2 = - 114. 409TM1 ＋ 1151.57TM2 -658. 292TM3 ＋98. 428TM4 （2）
class3 = -216. 288TM1 ＋ 1379.91TM2 -771. 508TM3 ＋ 114. 572TM4 （3）
Where TM1, TM2, TM3 TM4 are respectively the corresponding band of Landsat/TM, and classl, class2 , class3 represent three level of COD concentration.. The accuracy reaches 75% with the validation of the rest 10% samples. Further, the model was applied to the two Landsat/TM images collected on Oct. 16, 2005 and Apr. 10, 2006 in order to obtain the distributing image of COD in the rivers. By mean of the image the temporal change and spatial distribute of the concentration of COD within the three rivers were analyzed. The result shows that the establishment of identification mode based on remotely sensing provides an effective means to obtain rapidly and low - cost the concentration of COD in water environment.
出处
《干旱区地理》
CSCD
北大核心
2006年第6期885-893,共9页
Arid Land Geography
基金
中国石油天然气股份有限公司勘探生产部"遥感技术在油田环境保护中的应用"(项目编号:040148-6)
中国石油天然气股份有限公司科技管理部"遥感技术环境监测与评价研究"(项目编号:06-06B-01-03-02)
关键词
COD
水体光谱
FISHER判别分析
遥感模式
COD, Water spectra
Fisher multiclass discriminant rule
Remotely sensing identification mode
