土地利用现状更新调查中荒漠化地区绿地信息提取的方法研究

Research in the method of extracting vegetation information from the desertification area in land use renewal surveying

以内蒙古奈曼旗地区为研究区域,采用landsand5-TM数据,通过对影像进行预处理,主成分分析后进行监督分类,并运用第一主成分分量与实测数据建立植被盖度遥感信息模型,对研究地区的绿地信息进行定量化的提取。结果表明:对于该荒漠化地区,遥感影像进行主成分分析后,再进行监督分类效果较好,第一主成分分量进行线性变换后,与实测数据进行一次,二次,三次拟合,以三次拟合曲线精度最高。说明在定量提取植被盖度时,采用主成分分析后的第一主分量与实测数据建立遥感信息模型的方法是可行的。

The Naimanqi region in Inner Mongolia was selected as studied area in this research, using landsand5- TM data, the supervised clarification after principal constituent analysis was carried out through the preprocessing of image. The component of the first principal constituent and measured data were employed to establish the remote sensing information models of vegetation coverage, the vegetation coverage of the studied area was extracted in quantitative. For this desertification area, the results showed that the effect of supervised classification used by the prior three component of the principal constituent after principal constituent analysis of remote sensing image was better, the precision of synthetical classification reached 80.15 %, and the precision of cubic fitting curve was the highest when the first-order, second-order and third-order fitting were performed for the data of linear transformation of the component of first principal constituent and measured data. It demonstrated that adopting the method of remote sensing information models based on the first principal component after principal constituent analysis and measured data was feasible when the vegetation coverage was extracted in quantitative.