中文摘要：

澙湖是中国北方重要的滨海生态系统之一,水深较浅、透明度较低、底质类型复杂、人类干扰频繁等特点导致澙湖水深资料长期缺乏或存在较大误差,极大地制约了澙湖生态系统的研究和保护。本研究在现场实测水深的基础上,使用IKONOS高分辨率卫星影像,将山东荣成月湖水域划分为浅水区、植被区和深水区三类分别进行水深反演。研究表明红光波段（Band3）对水深的敏感性最高,分区建立线性回归模型反演水深的精度高于不分区的精度,多波段组合模型的反演精度最高,植被对水深反演的精度影响较大。反演结果表明,月湖水深最深处达271.23cm,水深反演的平均相对误差为13.16%。

英文摘要：

Lagoons in North China form an important coastal ecosystem; however, low transparency, shallow and complex bottom, and vulnerability to human activities results in poor precision of depth data, which is difficult for lagoon study and protection. We applied a high-resolution image of satellite IKONOS to inverse the depth of Yuehu Lagoon in Rongcheng, Shandong Province. The entire lagoon was divided three zones, i.e., shallow, plant-covered, and deep zones. Result shows that Band 3（red light） was the most sensitive band to depth for the inversion, while multiple bands inversion yielded the highest precision. In addition, regression on separate zones was more precise than that on the entire lake, and the precision was affected by vegetation. The deepest inversed depth of Yuehu Lagoon was –271.23 cm with a relative error of 13.16%.