Comparison Of Remote Sensing Extraction Methods For Impervious Surface In County Level And Its Spatiotemporal Dynamics Of Urbanization

With the rapid development of urbanization in China,impervious surface is a good indicator for urban expansion from a spatial perspective.At present,the information extractions of impervious surface on most researches are only from a single spatial resolution image.However,there is a lack of accuracy comparison between different resolution remote sensing images for impervious surface extraction.This paper takes Yucheng city,one of the typical county-level towns as the case study and uses different methods to extract impervious surface information from different resolution remote sensing images.The extraction accuracies of all results were compared and analyzed,so as the suitability.We analyzed the temporal and spatial dynamic changes of the impervious surface of the city from 2005 to 2015,with variables of local policies and economic factors.The results of the study are as follows:(1)The linear spectral mixture model is not applicable to GF-1 and other high-resolution remote sensing imagery because of the “pepper” phenomena which led to an undesirable requirement.Whereas,the linear spectral mixture model is suitable for Landsat images because the overall accuracy of the extraction result is 0.84 and the Kappa coefficient is 0.79 of which the results meets the accuracy standards when using this method to impervious surface extraction of Landsat image.In this study,the end-member selection compared to the typical urban research area in the use of this method adds the end member of steel plate house,which is unique to urban areas.(2)For the object-based image analysis method,the overall accuracy of extraction of impervious surface from GF-1 image and Landsat image is 0.94 and 0.9 respectively,and the Kappa coefficient is 0.92 and 0.87 respectively,both of which meet the accuracy requirements.The results of GF-1 image extraction are generally higher than those of Landsat image.The advantage of high resolution is that the GF-1 image possesses higher extraction accuracy whether in the region with high fragmentation or small patch area.(3)In typical urban areas represented by Yucheng city,the impervious surface space is concentrated on the urban area and the surrounding area,with many patches with similar square but irregular shapes.In terms of time,the impervious surface area of Yucheng city showed a trend of firstly increase and then decrease in 2000-2015.The increased area of the impervious surface is mainly caused by the change of vegetation,while the reduction area mainly becomes vegetation and soil.(4)During the process of urbanization,Yucheng City experienced three stages: population and spatial expansion,spatial expansion,and intensive use of space.The main factors affecting the impervious surface expansion include policy and economy factors.The rapid economic development of Yucheng City and the constant upgrading of its industrial structure have caused the impervious surface to expand rapidly.At the same time,differences in policies in different regions have resulted in spatial differences in the expansion of impervious surfaces.