摘要
遥感影像云检测是遥感影像处理中非常关键的环节,准确识别影像含云区域能够提升影像的利用价值。根据遥感影像的成像特点,将阈值法和纹理特征结合实现云和下垫面的分割。首先将影像从RGB(red-green-blue)空间转化为HSI(hue-saturation-intensity)空间,进而构建影像的显著性图像,利用Otsu法对显著性图像进行粗分割,再基于灰度共生矩阵分析云和下垫面的纹理特征,进一步提取出准确的云区。实验表明,该算法复杂度较低,提取效果良好。
Cloud detection of remote sensing image is a key point in remote sensing image processing,accurate identification of cloud areas can improve the value of image utilization.According to the imaging characteristics of remote sensing images,this paper combines threshold method and texture features to realize the segmentation of cloud and underlying surface.Firstly,the image is transformed from RGB space to HSI space,and then the saliency image is constructed.Secondly the saliency image is roughly segmented by Otsu method.Then,the texture features of cloud and underlying surface are analyzed based on gray level co-occurrence matrix,and the accurate cloud area is further extracted.Experiments show that the algorithm has low complexity and fine extraction effect.
作者
黄宇
潘励
HUANG Yu;PAN Li(School of Remote Sensing and Information Engineering,Wuhan University,Wuhan 430079,China)
出处
《测绘地理信息》
CSCD
2021年第2期16-19,共4页
Journal of Geomatics
基金
国家自然科学基金(41771363)。
关键词
云检测
显著性图像
灰度共生矩阵
cloud detection
saliency image
gray level co-occurrence matrix
