不同空间分辨率遥感数据识别湖泊的误差分析

湖泊面积及空间分布是表征水情变化的重要指示因子,对研究气候变化和人类活动对水资源的影响具有重要作用。如何利用遥感影像进行湖泊制图和面积估算是当前遥感应用研究中的难点问题。从实际应用需求出发,采用GF-1/WFV(Wide Field View)、Landsat-8 OLI(Operational Land Imager)、MODIS(Moderate-Resolution Image Spectroradiometer)遥感数据,选用安徽省受人类活动影响较大的54个湖泊为研究区域,采用NDWI二值分割法得到湖泊水体信息,使用30,500 m和1 000 m 3种空间分辨率的遥感影像进行湖泊制图,探讨制图精度与湖泊形态之间的关系。结果表明:湖泊分形维数值越大,湖泊形态越复杂, Kappa系数越高,面积误差越小;遥感影像分辨率越高,湖泊制图结果的Kappa系数和总体精度越高。

Error analysis on lakes mapping accuracy by remote sensing data of different spatial resolution

Lake area and corresponding spatial distributions are important indicators for monitoring environmental changes, which plays an important role in studying the impact of climate change and human activities on water resource. Accurate lake mapping and area estimation from existing remote sensing data still remain unsettled. To quantitatively investigate the association between the morphological complexity and lake mapping according to practical application, 18 m GF-1/WFV (Wide Field View), 30 m LandSat-8/OLI (Operational Land Imager), 500 m and 1000 m EOS/MODIS (Moderate-Resolution Imaging Spectroradiometer) data of 54 human-impacted lakes in Anhui Province were selected for map drawing. For lake mapping, water surface of lakes were extracted from remote sensing data by combining NDWI (Normalized Difference Water Index) and thresholding segmentation method. The main conclusions are as follows: (1) The larger the lake fractal dimension value is, the more complex the lake’s shape is; the higher the Kappa coefficient is, the smaller the lake area error is；(2) The spatial resolution of remote sensing images can affect the Kappa coefficient and overall accuracy of the lake mapping results. The higher the spatial resolution is, the higher the lake’s Kappa coefficient and overall accuracy in lake mapping are.