甘肃民勤县荒漠化区未利用地的遥感分类研究

以荒漠化土地典型分布区甘肃省民勤县为研究对象,以30 m的Landsat TM5与TM7遥感影像为主要数据源,在分析不同典型地物光谱特征的基础上,建立基于专家知识的决策树分类模型,利用该模型对荒漠化区的戈壁、沙地、沙漠、风蚀劣地等未利用土地进行细分,总体分类精度达到87.06%。决策树分类法总体效果较好,为荒漠化区土地管理以及再利用提供了技术支撑。

Unused Land RS Classification Study of Desertification Region in Minqin County, Gansu Province

Take Minqin County, a typical desertified land distribution area in China, as the research object, the decision tree model based on expert knowledge and the analysis of typical features of different spectral features was established by usin 30 m Landsat TM5 and TM7 remote sensing image as the main data source. The unused lands such as gobi, sandy land, desert, and aeolian barren land were subdivided by using this model, and the overall classification accuracy reached 87.06%. The results of the study show that the decision tree classification method is more effective, and can provide technical support for managing and reusing the land in desertification regions.