基于中值滤波和小波变换的遥感影像去云  

Remote Sensing Image Cloud Removing Based on Median Filter and Wavelet Transform

作　　者：李海巍[1] 杨敏华[1] 刘益世[1] 

机构地区：[1]中南大学地球科学与信息物理学院,湖南长沙410083

出　　处：《测绘与空间地理信息》2012年第7期49-51,共3页Geomatics & Spatial Information Technology

基　　金：国家自然科学基金项目(30570279);中南大学研究生创新项目(1343-74334000022)资助

摘　　要：去云处理是遥感图像处理以及大气纠正的重要步骤,目前常用的遥感薄云去除法一般只能去除低频区域的云,而对高频区域的云则没有处理效果。因此,在研究传统小波算法对图像进行处理的基础上,引入了中值滤波,它将每一像素点的灰度值设置为该点邻域窗口内的所有像素点灰度值的中值,旨在对高频突变云区进行处理。而后结合主观视觉和客观统计分析两方面对结果进行评价,并且与传统小波去云比较,经分析表明,高频突变云有所减少,达到了预期目的。Cloud removing is an important step in remote sensing image processing and atmospheric correction,the most commonly used remote sensing thin cloud removal method is generally used to remove low-frequency region of the cloud,while the high-frequency region of the cloud is not treated effectively.Therefore,in the study of the traditional wavelet image processing algorithm based on the introduction of a median filter,gray value of each neighborhood window pixel is set to the point of all the pixels within the median gray value in order to process high-frequency mutations in the cloud area.Then,combined with statistical analysis of the subjective visual and objective evaluation of the results of both and compared with the traditional wavelet cloud removal,the analysis shows that high-frequency mutations decrease to achieve the intended purpose.

关 键 词：图像处理 薄云 小波变换 中值滤波 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]