基于QuickBird数据的信息提取方法研究——以耕地提取为例

A study of information extraction based on QuickBird data——taking the extraction of cultivated land for an example

以成都西部某地区为研究对象,应用QuickBird数据、实测和统计等辅助数据,结合Erdas Imagine、eCogntion和Ar-cView等软件平台,进行耕地信息的提取研究。分别采用非监督分类法、最大似然分类法和面向对象分类法提取研究区的耕地范围,并进行误差评价与精度分析。结果表明,研究区统计结果与实地调查结果吻合较好,证明了Quickbird数据以及面向对象分类法在耕地信息提取中的有效性,为耕地信息的遥感提取研究提供参考价值。

This paper selected a region study in the west of Chengdu,applied to QuickBird data,measurement data and statistics data and other auxiliary data,combined with Erdas Imagine,eCogntion and ArcView and other software platform for information extraction of the cultivated land.The paper used non-supervised classification method,maximum likelihood classification method and object-oriented classification method to extract the cultivated land area,and put up the error evaluation and accuracy analysis.The results showed that the statistical results of the study area was agreement with the field survey results,proved the QuickBird data and object-oriented classification method to be effective in the cultivated land information extraction,provided a certain reference value for cultivated land remote sensing information extraction research.