联合概率密度空间的遥感自适应变化检测方法

多种因素引起的辐射特征变化,将造成阈值法变化检测的误检。对此,本文提出了一种联合概率密度空间的多阈值自适应变化检测方法。首先,将影像从像素空间转化到联合概率密度空间,将变化地物定义为联合概率密度空间的离群点,并采用迭代方法将其提取,然后映射回原始影像后确定变化区域。选取两种典型应用进行试验,结果表明,本文方法在正确率、误检率和漏检率方面优于传统方法,具有较好的稳健性。

Joint Probability Space Based Self-adaptive Remote Sensing Change Detection Method

A variety of factors has led to radiometric variations of the land cover, which severely limits the threshold based change detection method performance. To overcome this problem, we propose a joint probability density space based self adaptive multi-threshold change detection approach. Firstly, the two images of the same geographic area acquired at different time are transformed into the joint probability space. In which, the land cover change pixels are defined as outliers and identified by an iterative method. Then, the extracted outliers are mapped back to the original image space and determine the change area. To illustrate the performance of the proposed method, an experimental analysis on two classical applications is reported and discussed, results show that the proposed method over performed the state of art method in true rate, false alarm rate and omit alarm rate, with high stability.