ABSTRACT
Image segmentation is to divide the image into a number of specific, unique characteristics of the region and to put forward the technology and process of the target of interest. It is the key step of preprocessing in the regime of image recognition and computer vision. Without correct image segmentation, there wouldn’t be correct identification. However, the only basis for segmentation is the brightness and color of pixels in the image. Therefore, when the image is automatically processed by the computer segmentation, a variety of difficulties will be encountered. Because remote sensing image has many characteristics such as high resolution and unequal noise distribution, the traditional segmentation algorithm cannot achieve satisfactory accuracy when analyzing these images. This paper proposes a new image segmentation framework based on the graph theory. During the recent years, the use of graph theory in many mature theories and mathematical tools for image segmentation has become a hot topic in the field of image segmentation research. The graph theory is a branch of applied mathematics, and have a good relationship with the mapping between the images. This paper also includes experimental result on the high resolution that is spatial resolution 0.1~0.3 m for remote sensing image.
