期刊名称：International Journal on Computer Science and Engineering
印刷版ISSN：2229-5631
电子版ISSN：0975-3397
出版年度：2011
卷号：3
期号：04
页码：1638-1644
出版社：Engg Journals Publications
摘要：Image classification is perhaps the most important part of digital image analysis. With supervised classification, the information classes of interest like land cover type image. These are called �training sites�. The image processing software system is then used to develop a statistical characterization of the reflectance for each information class. This stage is often called � Signature analysis� . Unsupervised classification is a method which examines a large number of unknown pixels and divides into a number of classes based on natural groupings present in the image values. Unsupervised classification is becoming increasingly popular in agencies involved in long term GIS database maintenance. The reason is that there are now systems that use clustering procedures that are extremely fast and require little in the nature of operational parameters. Thus it is becoming possible to train GIS analysis with only a general familiarity with remote sensing to undertake classification that meet typical map accuracy standards. With suitable ground truth maccuracy assessment procedures , this tool can provide a remarkably rapid means of producing quality land cover data on a continuing basis. The profusion information of the earth surface offered by the high resolution satellite images for remote sensing applications. Using change detection methodologies to extract the target changes in the areas from high resolution images and rapidly updates geodatabase information processing. However, the traditional method of change detection are not suitable for high resolution remote sensing images. To overcome the limitations of traditional pixel-level change detection of high resolution remotem sensing images, based on georeferencing and analysis method, this paper presents a clean way of multi-scale amalgamation for the high resolution remote sensing images change detection. Experiment shows that this method has a stronger advantage than the traditional pixel-level method of high resolution remote sensing image change detection.
