高分辨率遥感影像配准与变化检测一体化处理

Unifying approach for high resolution remotely sensed imagery registration and change detection

为了克服高分辨率遥感影像配准与变化检测作为单独环节处理的局限,该文提出了一种基于变分理论的配准与变化检测一体化处理方法。该方法将配准误差作为一种光谱变化决策因子,变化信息以权值的方式迭代反馈给变分配准模型的解算过程。为了更准确地检测建筑物这个特定目标的真实变化,该文采用多尺度最大形态学轮廓建筑物检测指数的差异作为另外一个决策因子。最后将配准误差反映的变化和建筑物检测指数的差异这两个决策因子在D-S证据理论框架下建立概率模型进行融合处理,进而得到建筑物的变化检测结果。该文选取WorldView-2数据进行实验,实验结果表明,一体化处理思路可以有效地解决单独处理的局限,从根本上解决配准误差对变化检测结果的影响以及由于变化而使配准精度降低的问题,进而提高配准和变化检测的质量。

In order to overcome the limitation of high resolution remotely sensed imagery registration and change detection as separate processes,a unifying approach method for high resolution remotely sensed imagery registration and change detection based on variational theory was proposed.The registration error would be used as decision indicator of the spectral change and the change information would be fed back to the variational registration by the means of weights.In order to detect the building change more accurately,the difference of multi-scale maximum morphological profiles building index(MMMPBI)was used as another decision factor.The D-S theory fused the spectral change reflected by registration error and difference of MMMPBI to construct the probabilistic model to get the final change detection results.The paper chose WorldView-2 data to experiment and the experiment result showed that the idea of integration process could solve the limitation of separate approaches effectively,fundamentally solve the problem of the effect of registration error on the change detection result and the reduction of registration precision due to change information,and then improve the quality of registration and change detection.