Classification Of Remote Sensing Imagery With Texture Information

With more and more explore and application of High Resolution Satellite, the texture information in Remote Sensing Imagery becomes much more abundant. High Resolution Remote Sensing Imagery is playing an important role in numerous areas such as land use and investigation, surveying and mapping, urban planning, voyage, national defense military, etc. However, for the technology of remote sensing image spectral pattern recognition, improvement of spatial resolution causes descent of classification accuracy instead of improvement. Depending only on the spatial information leads to much declassification. This is because as the improvement of spatial resolution the internal structure of objects has become increasingly clear that this is texture structural features of objects become increasingly evident in remote sensing image, leading to the same type of features is expressed not the same kind of spectral information, but has expressed a wide range of spectral information. Depending only on the spatial information leads to much declassification. Therefore, the traditional pixel-based classification method can not give a satisfied performance for High Resolution Imagery. Its accuracy of classification is far from productive demand.At the same time, with the improved of the spatial resolution, the visual interpretation accuracy is getting higher and higher, this is because people can be accurately determine the properties of the objects by features of the internal spatial structure that features of internal texture information, because the internal spatial structure features of the same object is similarity, and the features of the different objects have obvious differences. Texture and spatial structure play an important part in High-resolution image and the weighting of information very informative. Therefore, We can starting from the different objects have different texture, in combination with other spatial spectrum information and structure knowledge to solve the technical difficulties of the remote-sensing image classification accuracy is lower by computer. At present, texture and spatial structure information accounts for a considerable proportion of High Resolution Imagery. Object-oriented Classification Method can make good use of them during classification process.First of all, the paper describe the status and trends of remote sensing image classification, research objectives and theoretical significance as well as the thesis content and organizational structure. Then, some pre-processing work such as geometrically corrected, texture enhanced, gray-scale compression were applied to the image data. The Gray Level Co-occurrence matrix method was applied to extract the image texture information. Finally, in detailed analysis the multi-scale segmentation techniques and fuzzy classification techniques that as the goal of analysis to remote sensing image. The texture image calculate by the object-oriented classification method will applies to the computer automatic classification. The classification result shows that key information combined with proper methods leads to a satisfied result. Each land type in the study area can be discriminated. The classification accuracy is 86%, which is much better than methods based on pixel and spectrum.