Research Of Remote Sensing Image In Land Use Classification Based On The Fractal Theory

With the advancement of technology and population growth, the range of human land use continues to expand, growing strength. Governments have long been going on at different levels of resources and land use survey of working conditions.To grasp the type of land use change information is a necessary condition of land use planning, protection of basic farmland, land use and other land use control management.Currently, remote sensing technology has become an important means of monitoring changes in the type of land use. Rely on remote sensing technology to detect the type of land use change on the premise that accurate remote sensing image classification. Although people in remote sensing image classification has made a lot of research, but there is no general theory of segmentation, the proposed algorithm has now mostly on specific issues. It is because of the uncertainty of the importance of remote sensing image classification and classification tasks, people still continue to study and explore new classification theory and classification algorithm.The research on large areas of land use and land cover classification survey,fractal theory is introduced into the classification of remote sensing to explore the fractal theory in land remote sensing images to monitor changes in the type of application. Research results with traditional supervised classification and unsupervised classification results were compared to explore the advantages of this method. At the same time this method Zhanjiang Potou 2004 and 2013 ETM / TM image processing to obtain land use type classification results, monitoring the region in 2004- 2013 on land use change. The main results of this study have achieved:1. TM / ETM images for supervised classification to obtain land use types in Zhanjiang Potou supervised classification. Using Google Earth images as a reference,and to verify the accuracy of classification.2. Using unsupervised classification approach to TM / ETM remote sensing image classification again to obtain land use types unsupervised classification results using Google Earth as a reference image, and the classification accuracy verified.3. For the characteristics of fractal theory Band Math remote sensing images,remote sensing image classification based on fractal theory to obtain the land use type classification results using the same Google Earth images as a reference, and to verify the accuracy of classification.4. Comparative analysis supervised classification, unsupervised classification and the resulting classification of remote sensing image classification method based on fractal theory, based on remote sensing image classification method proved fractal theory of superiority.5. The remote sensing image classification results, Zhanjiang Potou 2004 and2013 land-use changes will be discussed.