结合IR-MAD与均值漂移算法的密集城区遥感影像变化检测

不同时相遥感影像变化检测已成为土地利用变更调查、城市扩张分析、自然灾害分析及其他环境问题必不可少的技术手段之一。本文提出了一种结合IR-MAD与均值漂移算法的密集城区遥感影像变化检测方法。该方法通过伪不变特征法完成两期影像的相对辐射校正，有效改善影像间的配准误差，并利用IR-MAD算法对校正后的影像进行迭代运算，采用均值漂移算法对迭代后的影像进行分割，同时运用形态学方法处理分割后的影像，最终提取变化图斑。试验结果表明，该方法可以有效检测出变化区域，可应用于城市地表覆盖的变化检测。

Change detection of remote sensing imagein dense urban area based on IR-MAD and mean shift algorithm

Change detection of different phase remote sensing image has become one of the indispensable technical means for land use change survey, urban expansion analysis, natural disaster analysis and other environmental issues.This paper proposes a method for detecting remote sensing image changes in dense urban areas based on a combination of IR-MAD and mean shift algorithm. The relative radiation correction of the two-stage image is completed based on the pseudo-invariant feature method, which effectively reduces the registration error between images. Then IR-MAD algorithm is used to perform iterative operation on the corrected image,the mean shift clustering algorithm is used to classify the iterated images, the morphological methods are used to process the segmented images, and finally the change maps are obtained.This paper shows that this method can effectively detect the change area and can be applied to urban change detection and disaster monitoring.