摘要
提出了一种基于Retinex和闭合边缘提取的新方法.这一方法对于彩色遥感图像中存在的阴影区域得到了增强.首先将彩色遥感图像利用小尺度Retinex进行增强,再利用边缘检测的结果,编制程序提取原图像中阴影的区域,将用小尺度Retinex增强的图像中相应的像素替代原图像中阴影的区域.在实施闭合边缘提取过程中,提出了真断点与准断点判别法、“蝌蚪”状尾巴回退消除法等,确保所需要的闭合区域的提取.这一方法对于遥感图像中地物的半自动提取方面有着重要意义.
The research work has brought a new method, which is based on small-scale Retinex and closed regions extraction, to enhance color remote sensing image. The information in dark regions is highlighted using Retinex, and the closed regions of the remote sensing image are extracted after edge detection. The fused image is that the shadow regions in original image are replaced by the same regions of the handled image using Retinex. The new fused image has more information than which of original image. While the closed regions is being extracted, new programs to judge the real-broken points and the quasi-broken points were developed, to eliminate the tails of ＂polliwog＂ by the backspacing method, and so on. These methods play an important role for the objects of remote sensing images by the method of semi automatic extraction.
出处
《光子学报》
EI
CAS
CSCD
北大核心
2007年第3期565-567,共3页
Acta Photonica Sinica
