基于形态学的多尺度遥感图像分割方法  

An Algorithm of Multi-scale Image Segmentation Based on Morphology

作　　者：陈景广[1] 佘江峰[1] 黄海涛[2] 

机构地区：[1]南京大学地理信息科学系,江苏南京210093 [2]中国科学院地理科学与资源研究所,北京100101

出　　处：《地理与地理信息科学》2012年第4期22-24,37,共4页Geography and Geo-Information Science

基　　金：国土资源部公益性行业科研专项研究(201011015-1)

摘　　要：基于形态学提出了多尺度遥感图像分割方法。首先使用面积形态学算子处理原始图像,构造不同尺度下的遥感图像;然后提取多波段遥感图像梯度,充分利用遥感图像的多波段信息;再在不同尺度的图像上使用区域标记方法提取标记,作为分水岭分割算法的生长点;最后使用分水岭分割实现对遥感图像的多尺度分割。该方法在Matlab中予以编程实现,并将其应用于江苏省海安SPOT影像中,实现对图像不同尺度的分割,能够有效提取出多波段遥感图像中不同尺度下的地物。Based on morphology,an algorithm of multi-scale remote sensing image segmentation is proposed in this paper.Algorithm consists of four steps listed as below：firstly,area morphology is used on the primitive image,structuring remote sensing image under different scales;secondly,edges of multi-bands remote sensing image is detected and information in different bands is used sufficiently;thirdly,a marker-extracted approach is used on different scale images to create growing points for watershed segmentation algorithm;lastly,watershed segmentation algorithm is applied to achieve results under different scales.This algorithm gives a programming realization in Matlab.Then,the image of SPOT covering Haian County of Jiangsu Province is taken as a case study for this algorithm.The result of experiment proves that this algorithm is able to extract surface features under different scales in multi-bands remote sensing image effectively.

关 键 词：面积形态学 分水岭分割 多尺度 图像分割 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]