| The use of remote sensing technology for land-use change detection has become a hot in domestic and international land-use and cover change research field. Based on the analysis of traditional land-use change detection method, this article proposed an object-oriented land-use change detection method, and developed an object-oriented land-use change detection system. The main research contents are as follows:(1) Different resolution remote sensing image segmented in multi-scale. Studied the different resolutions remote sensing image segmented results in multi-scale by using of Mean-Shift.(2) Feature selection and extraction of image object. According to different land- use types, quantitatively described the image object's characteristics on spectrum, texture and shape, and the feature information knowledge base was established hierarchically after image segmentation.(3) Establishing change detection rules. Combining the extracted image object feature information, building an effective change detection rule and method to achieve the type of land-use change information extracting and determining.(4) Object-oriented land-use change detection. Extracting feature information in an image segmentation object, and hierarchically building feature information knowledge base with different land-use types and extracting the changed polygon object combining of change detection rules. Then segmented the image in the changed polygon object, using the Mean-Shift algorithm, and identified the land-use type. Accordingly, achieved the land-use change detection in the changed polygon objects.(5) Object-oriented land-use change detection system. In the VC++6.0 with MapObjects2.2 and GDAL, designed and developed object-oriented land-use change detection system. While, took advantage of this system for land-use change detection tests. The results showed that object-oriented change detection method had obvious advantages compared with the traditional pixel-based change detection method. |
PDF Full Text Request