中文摘要：

以浙江省瓯江流域为例,基于SWBD修复的SRTM DEM数据,采用Arc Hydro Tools水文分析工具自动提取瓯江水系,并分地貌、分河流等级地定量评价水系数据精度,开展1∶250 000水系自动更新的可行性研究。结果表明：①SWBD修复的SRTM DEM的空白区域面积为54.78 km2,有效地弥补了SRTM DEM的数据缺失,进而提高了水系提取的准确度和精度;②与1∶250 000水系数据相比,基于SWBD修复后的SRTM DEM,在小起伏山、中起伏低山、低海拔丘陵上提取的水系数据精度高于其他地貌,而干流、一级支流、二级支流的精度又高于三级支流;③以资源三号卫星ZY-3遥感影像为参照,从水系上采集同名点反复比较点位精度后发现,利用SRTM DEM提取的水系符合制图规范和测绘内业规范（限差1 mm）,可以满足1∶250 000水系自动更新的要求。

英文摘要：

In recent years,free access of high- precision DEM and Arc Hydro Tools have created good conditions for the automatically update of the river based on DEM. To investigate the possibility of automatically updating 1∶ 250 000 river network,we take Oujiang watershed in Zhejiang Province as a study case,use Arc Hydro Tools to extract river network based on SRTM DEM repaired by SWBD,and analyze the data accuracy in terms of geomorphology and river level. The results show that：（ 1） SWBD repairs the blank area of 54. 78 km2,making up for the data missing of SRTM DEM,so as to improve the accuracy of river network extraction;（ 2） we divided the landforms of Oujiang watershed,and found that small rolling mountain,middle rolling low mountain account the Proportion of more than 80% of the total area,while low- altitude plains and low- altitude tableland have a small proportion,close to 3% of the total area.（ 3） compared with the 1∶ 250 000 river network,the river morph metric parameters agree well with 1∶ 250 000,and the accuracy of river network extracted from SRTM DEM is higher on small rolling mountain,middle rolling low mountain,low- lying hills than other landforms. Also,In aspects of river level,the trunk,first tributary,second tributary river are better than the third tributaries river;（ 4） taking remote sensing image of ZY- 3 as a reference,the positional accuracy of river network automatically extracted from SRTM DEM is higher than that of 1∶ 250 000 data,conform to the surveying and mapping standards（ limited error is 1 mm）,so can meet the requirements of 1∶ 250 000 river network automatically update.