基于图形形态学的海岸线自动提取算法研究  

Fast Automatic Coastline Extraction Algorithm Based on Graphics Morphology

作　　者：万剑华[1] 王珂[1] 赵世来[2] 张纪松 

机构地区：[1]中国石油大学(华东)地球科学与技术学院,山东青岛266555 [2]山东黄河河务局,山东济南250011

出　　处：《人民黄河》2011年第10期126-127,130,共3页Yellow River

基　　金：水利部公益性行业科研专项(200701030)

摘　　要：基于图形形态学的海岸线自动提取算法,应解决好自动选取初始计算像素、减少参与计算的像素个数和计算次数、充分利用波谱信息,以及去除河流、湖泊、岛屿和云的影响等问题。根据算法设计,计算分为预处理、海陆位置确定、海岸线提取三个步骤,即快速、自动从金字塔图像数据中选定一个初始种子点,根据水体光谱特性,逐层计算出所有连通的海洋区域点,从而可确定出海岸线位置。将该算法应用于TM的图像试验,结果表明:影像虽然存在一些噪声,但是入海河流、潮沟等都没有对海岸线提取的结果造成影响。In graphics morphology based coastline automatic extraction algorithm,the following problems should be well settled,including automatically select the initial seed point from a pyramid image data,reduce the pixel number introduced the calculation,use the spectral characteristics fully and cut the effects of rivers,lakes,islands and clouds.According to the algorithm design,the calculation procedure includes three parts,pre-procession of data,determination of location between land and sea and the extraction of coastline,which means we can locate the locations of coastline quickly by selecting an initial seed from a pyramid image data combined with the spectral characteristics of water in calculating all points of connecting marine areas layer by layer.The algorithm applied in coastline extraction based on TM images gets excellent results in spite the image noises such as the rivers and tide channels.

关 键 词：图形形态学 海岸线 自动提取算法 遥感 

分 类 号：P407.8[天文地球—大气科学及气象学]