Abstract: It is a difficult task for shadow and feculent water area processing when extracting water body from remote sensing image by traditional methods.Light Laser Detection and Ranging（LiDAR） is robust to gain point cloudy data from shadow and feculent water as it has less affection from that area.However，the point cloudy data is always with low resolution which results to contour line with low accuracy.A new approach combing point clouds data with high resolution image for water body extraction is proposed to improve the accuracy.Experiments results indicate the contour lines obtained through this way are accuracy and with details.The accuracy of water body extraction can reach 98.6%.

Key words: Light Laser Detection and Ranging（LiDAR）, water body, contour line, seed filling

摘要： 从遥感影像中提取水体的传统方法在阴影处和水体浑浊处存在一定局限；机载激光雷达（LiDAR）获取点云数据时水体受阴影和浑浊的影响小，因此使用点云数据提取水体比使用影像数据更加稳健；但点云数据的分辨率低，所提取的轮廓线精度不高。为了提高水体边缘轮廓线的精确度，提出一种高分辨率影像与点云数据相结合的水体轮廓线提取方法。实验结果表明，该方法所提取的水体轮廓线定位精确、细节完好，水体提取准确率达到98.6%。

关键词: 机载激光雷达, 水体, 轮廓线, 种子填充