摘要
去除遥感图像上朵云及阴影的前提是快速准确地提取出它们的相应区域。本文从朵云及其阴影的基本特征出发,改变了传统意义上仅仅依靠阈值分割进行提取的方法,对图像完成预处理和自动双阈值分割后,应用基于Freeman链码的轮廓跟踪算法分别从二值和灰度图像上对它们进行了提取实验。整个处理过程不需要进行人工干预,在完成图像去噪的同时,较准确地提取出了完整的云区域,为后续的去云去阴工作打下了基础。
The premise of removing cirrus clouds and shadows in remote sensing images lies in extracting their corresponding areas rapidly and exactly.According to the basic characteristics of cirrus clouds and shadows,this paper changes the traditional way of extracting them merely depends on the threshold,and after image preprocessing and automatic double threshold segmentation,the algorithm of tracing contour based on the Freeman chain code is used to carry out experiments upon extracting them respectively in the binary images and grayscale images.Manual intervention is not needed in the whole process,and the integrated cloud areas are extracted quite accurately meanwhile the noises in the images are eliminated,and it prepares for the subsequent work of cirrus clouds and shadows removal.
出处
《测绘与空间地理信息》
2010年第5期3-7,共5页
Geomatics & Spatial Information Technology
关键词
朵云及阴影
提取
阈值分割
FREEMAN链码
轮廓跟踪
cirrus clouds and shadows
extracting
threshold segmentation
freeman chain code
trace contour
