Change Detection Of Urban Typical Land Coverage Based On High-resolution Satellite Remote Sensing

Research for high-resolution satellite remote sensing application is a hot spot in remote sensing field, with the high spatial resolution, short revisit cycle and agility of the high-resolution remote sensing satellite, high-resolution remote sensing image is playing an important role in land use change detection, land use survey, land spatial database updates and other tasks. Change detection based on high-resolution remote sensing has great significance for national geographical conditions monitoring and digital city.Currently, there is no optimal change detection method for high-resolution satellite remote sensing images. While using the general detection method for high-resolution remote sensing images, there will be low efficiency, fragmentation, difficult to determine the change threshold and other issues in the detection process. This paper is for urban land use dynamic monitoring needs, based on high-resolution satellite remote sensing images, improve a change detection algorithm based on canonical correlation analysis methods, use a comprehensive approach to detect the changes in images with multiple bands, the introduction of the cross window mean filtering during the detection process, is designed to reduce the fragmentation, and the dynamic threshold algorithm based on traditional mean-standard deviation fixed threshold can reduce the information loss, lower the omission. Based on the theories mentioned above, this paper did an experiment using GDAL and other open source libraries to implement the improved canonical correlation analysis change detection algorithm and selected a construction zone which has significant land coverage changes in Shenyang Hunnan as the study area, used two images of GF-1 in July 4th 2013 and August 21th 2013 for change detection, the experiment also assessed the detection accuracy to verify the validity of this algorithm. Classification experiment based on object-oriented classification method, using ENVI Feature Extraction module to classify the change detection results of study area, and statistics the area and the change rate of each land coverage types using ArcGIS, finally, produced the land coverage changes thematic map of the study area.For the large amount of data and rich minutiae of high-resolution satellite remote sensing images, this paper improved a change detection algorithm based on canonical correlation analysis methods, and verified the validity of this algorithm, used the object-oriented classification method to classify the change detection results, and used the classification results for statistical analysis and mapping. The results proved that theories in this paper have certain value in the field of change detection in urban areas based on high-resolution satellite remote sensing images, and provide a practical method and practical basis for the automatic change detection of the urban typical land coverage.