基于不变像素的遥感数字图像标准化

遥感数字图像直方图的标准化常用于不同时相相邻遥感图像的拼接和进行土地利用变化监测的预处理,还可以消除不同成像条件对图像分析的影响,部分纠正由于太阳/传感器位置或大气影响造成的相邻图像的辐射误差,从而达到相同成像条件的效果。由于常用的基于数值统计学的直方图的标准化方法不能达到理想效果,该文提出了基于无变化像素进行图像标准化的方法,通过多光谱图像差异直方图及每一波段黑灰白像素的相关系数计算,自动选择无变化像素,进而进行直方图的标准化。将该方法应用于2期TM图像遥感数据,明显减小了成像条件差异,提高了数字图像标准化的有效性。

Normalization of Digital Satellite Image Histogram Based on Unchanged Pixels

Normalization of histogram is often required in image preprocessing stage such as multi temporal remote sensing image matching and land-use change detection.The normalization can also minimize different imaging condition effects in analysis and rectify radiometric differences derived from solar/sensors location,atmospheric effects between adjacent images in such a way as if they have been acquired at the same imaging conditions.Relative radiometric normalization which is normally applied in image preprocess...