结合MNF变换和Canny算子的遥感影像变化检测

针对大面积区域的多时相遥感影像变化检测的需求，提出了一种基于最小噪声分离（MNF）的Canny边缘检测提取影像变化信息的检测方法。对多时相影像采用多种变换组合成具有多维波段信息的影像，采用最小噪声分离法分离噪声并得到单波段差异图，通过Canny边缘检测法计算梯度幅值，采用高低双阈值法细化边缘，从而提取差异图变化边缘，有效突出了变化信息。以1995年和2003年加扎勒河的两期遥感影像为例，利用两时相影像进行土地覆被变化检测。实验结果表明，该方法适用于监测大面积区域内地物的突变情况。在数据基础上进行最小噪声分离可以有效解决传统Canny边缘检测提取边缘时造成的伪边缘现象，同时采用高低双阈值法有效去除伪边缘点，从而获得更加精确、直观的变化检测效果，在自然地理变化监测、地理国情灾害监测等有很好的应用价值。

Remote Sensing Image Change Detection Based on MNF Transform and Canny Operator

In view of the requirement of change detection for multi-temporal remote sensing images in large area, a method of Canny edge detection based on Minimum Noise Fraction（MNF） is proposed to extract image change information. Firstly, multi-temporal images use a variety of transformation combined with multidimensional information band images, and then use the MNF of noise and single band difference map by Canny edge detection method to calculate the gradient amplitude, the level of the dual threshold method to extract edge thinning, the difference map change edge, highlight the change information effectively. The remote sensing images in 1995 and 2003 Nianjiazhale River as an example, it uses two images for land cover change detection. The experimental results show that this method is suitable for monitoring the mutation of large area terrain, the minimum noise based on data separation can effectively solve the pseudo edge phenomenon caused by the edge detection of the traditional Canny edge detection, and the level of the dual threshold method to remove false edge points, so as to obtain more accurate detection results, intuitive changes. It has good application value in monitoring the change of geographical conditions, natural disaster monitoring etc.