Change detection in SAR images based on iterative Otsu

Abstract

This work addresses the problem of detecting the changes between the two SAR images from the same region with two different phases. Because of the advantages of fast computing speed and high efficiency, the development of efficient and accurate algorithm for spatial change detection is always an important research issue in radar field. Unfortunately, the conventional spatial change detection method has low performance when the change area occupies only a small part. To solve this problem, an iterative Otsu method is proposed in this paper. First, a difference image is generated by using an improved neighborhood-based ratio approach that combines the neighborhood information. Then, we propose the iterative Otsu method to obtain the segmentation threshold of the difference image, in order to solve the segmentation problem when the proportion of the changing region is very small. The experimental results show that the detection accuracy of this algorithm is obviously improved compared with the existing methods and it has better robustness for the present of coherent speckle noise.

Keywords

• sar image
• change detection
• neighborhood-based ratio approach
• threshold segmentation
• iterative otsu