ETM+影像亚像元级城市土地覆盖组分丰度提取——以南京市为例

城市景观复杂,地物具有强烈的空间异质性,利用中等空间分辨率遥感影像(如ETM )进行城市土地覆盖研究会出现大量混合像元。针对该问题,以南京市为例,通过混合像元分解提取城市土地覆盖组分丰度,进行城市土地覆盖研究,并从各个波段的残差、平均残差、相对误差、均方根误差等方面分析模型误差。利用IKONOS影像对提取的组分丰度图像进行精度验证,发现不同城市土地覆盖组分丰度图像具有一定的可靠性,与城市实际情况吻合。城市土地覆盖组分丰度图像还能作为一种新的数据源用于城市功能分区、城市人口数据空间化和城市环境监测与评估等研究中。

Urban Land Cover Composition Abundance Extraction of Sub-pixel Level Based on ETM+Image: A Case Study of Nanjing

Urban landscapes are complex and the objects in urban area have strong spatial heterogeneity,which makes the mixed pixels emergence when performing urban study using moderate spatial resolution remote sensing images(e.g.ETM ).This paper selected Nanjing city as study area and performed urban land cover study through attaining urban land cover composition abundance images using mixed pixel decomposition method according to the above thing.Model error analysis was performed using residual errors of each band;mean residual errors,relative errors and RMS.The accuracy validation of composition abundance images was performed using IKONOS images of the study area.The results indicate that several urban land cover composition abundance images have a considerable reliability and they are consistent with urban actuality.Moreover,these urban land cover composition abundance images can be used as a new kind of data source in many urban study domains,e.g.urban function area classification,spatial distribution of urban population and urban environment monitoring and evaluation and so on.