Research Of Land Use And Land Cover Change Detection Based On Object-Oriented

With the further development of remote sensing technology and geosciences analysis model, land use and land cover change detection using remote sensing techniques regular or irregular, has become a key of basic work for current land use planning, land management and environmental protection. A post-classification change detection method used to solve the problem, for poor level of automation and the low of detection accuracy and so on.The content of studied contains several respects.(1) Image segmentation is a base for Object-Oriented class. Choosing an optimal segmentation scale is important. For it directly affects the feature extraction, classification and change detection results. Therefore, researching of the optimal scale of research is a key part of this article.(2) Analysis the advantages and disadvantages of these methods of these methods in Object-oriented software, threshold classification, fuzzy classification, decision tree classification and so on. Use decision tree classification method to achieve the degree of automation for Extracts the classes from the same surface category with quite different spectral features or the more obvious surface features category by the threshold, and then, uses fuzzy classifier to classify the classification which are so more difficult to distinguish, just like "synonyms spectrum", validate the technology can effectively improve the classification accuracy.(3)After classification, images and vector object has the multi-source information (class hierarchy, attribute information, relationships, etc.). Integrated the multi-source information into the object-oriented change detection, and use overlay analysis and relationship to obtain change information of land use and land cover.In short, change detection based on object-oriented analysis technology is not limited to classification, but also to broaden the applications of the technology; it can be well applied to the land use/land cover change detection.