基于融合和广义高斯模型的遥感影像变化检测

提出了一种基于融合和广义高斯模型的遥感影像变化检测算法。该算法融合差值影像和比值影像的结果构造差异影像，然后运用灰度形态学对差异影像进行顾及空间邻域关系的处理，再对处理后的结果运用广义高斯分布模型估计变化与非变化像元的概率密度参数，最后采用改进的KI算法计算最佳分割阈值，提取变化区域。实验结果表明，所提出的变化检测算法稳健、高效，具有较大的实用价值。

Based on Fusion and GGM Change Detection Approach of Remote Sensing Images

This paper presents a change detection approach of remote sensing images based on fusion and general Gauss model(GGM).This method constructs difference image by fusing the results of subtraction operation and division operation,then,processes difference image by gray morphology,again,estimates the probability density of change pixels and no-change pixels by GGM,finally,calculates the optimal threshold by the improved KI approach and extracts the changed regions.Experiments show that the method of fusion and GGM is solid and efficacious,which has preferable application value.