“高分二号”卫星数据面向对象的海岸线提取法

Coastline Extraction Based on Object-oriented Method Using GF-2 Satellite Data

"高分二号"(GF-2)卫星的成功发射是中国遥感事业发展的重大突破,标志着中国遥感卫星进入亚米级时代。由于兼具高辐射精度、高定位精度和快速机动能力等特点,GF-2卫星数据应用于海岸线提取将极大提高提取的精度和速度。现有的遥感解译海岸线方法主要有阈值分割法、边缘检测法、区域生长提取法、神经网络法和面向对象法等,其中面向对象法是一种新兴的遥感图像解译方法,该方法通过对影像的分割,使同质像元组成大小不同的对象,从而实现较高层次的遥感图像分类和目标地物提取。文章以深圳大鹏半岛GF-2卫星数据为例,通过对影像进行正射校正、配准、融合和裁剪等预处理,再采用面向对象方法对海岸线进行了提取,并对入海河流河口处的海陆分界线划分进行了初步研究。研究结果表明,将海岸线提取结果与GF-2卫星遥感影像叠加进行验证,结果可靠,且所耗时间短,效率高。

The successful launching of GF-2 satellite of high spatial resolution, high radiation precision, high positioning accuracy and fast maneuver is a significant breakthrough in the development of remote sensing and marks the advent of the submeter era in China. The main methods of coastline extraction by remote sensing include threshold method, edge detection method, region growing method, neural network method and object-oriented method of which the last one can realize different classification and extracts target objects by segmenting image. This paper extracted the coastline by using pre-processing such as ortho-rectification, registration, fusion and clipping and object-oriented method with GF-2 data of Dapeng Peninsula, Shenzhen. In addition, the division of sea-land demarcation line is initially discussed. At last, the extraction result proves reliable by comparing coastline with GF-2 data and it can improve the precision and efficiency of coastline extraction by using GF-2 data.