Research On Method And Application Of Urban Built-up Area Information Extraction Based On Spectral Index

This thesis aims to study the extraction of built-up information based on the remote sensing spectral index.Firstly,the performance difference of the typical built-up index inversion of the remote sensing image under Digital Number(DN)and Surface Reflectance(SR)is analyzed.Secondly,the Urban Index(UI)was improved,and an algorithm model for extracting built-up information to suppress bare land was proposed;finally,a long-term sequence of built-up in the main urban area of Chongqing from 2001 to 2019 was constructed the information data set.The specific research content and main conclusions are as follows:1.In order to study the impact of atmospheric correction on the performance difference of the built-up index,seven typical built-up indexes were selected to invert the built-up index before and after the atmospheric correction,experimental results show that:(1)After the atmospheric correction,the minimum value of the built-up index becomes smaller,the maximum value becomes larger,and the range of values becomes wider.The standard deviation of the built-up index after atmospheric correction is approximately doubled compared to the uncorrected built-up index.The larger the standard deviation of the built-up index,the greater the degree of dispersion.From the perspective of extracting built-up information,atmospheric correction removes the atmospheric influence of the image can better reflect the true situation of surface radiation.(2)There is a strong positive correlation between the corrected built-up index and the uncorrected inversion index,but there are obvious differences.The overall accuracy of various built-up indexes before and after atmospheric correction is different from the Kappa coefficient.Compared with the index of uncorrected image inversion,the overall accuracy and Kappa value of indexed built-up information after atmospheric correction are correspondingly improved,so atmospheric correction is beneficial to improve built-up index performance.2.In order to improve the extraction accuracy of the built-up index,and aim at the shortcomings of bare land and urban built-up areas that are easy to confuse and extract built-up information mixed with vegetation,this study analyzed the first article based on the spectral spectrum curves of built-up,bare land,water and vegetation.With referenceto the idea of building an index-based built-up index(IBI),combined with the optimal threshold algorithm,a new model method,and finally,an algorithm model for suppressing bare ground built-up information extraction is established.The experimental results show that:(1)The proposed built-up information extraction algorithm model for suppressing bare land has better effect and higher accuracy.The overall accuracy of this method is85.6%,which is higher than other built-up indexes,and can effectively suppress the bare land and sparse vegetation extract.The built-up information extracted at the same time has good spatial continuity,which is helpful for the structural analysis of built-up information.(2)In the current built-up index,the bare ground spectral features are avoided when constructing the index,and the overall accuracy of the composite band,component component,basic index and built-up index model constructed through multiple processing of remote sensing images is higher.3.Taking the main urban area of Chongqing as the research area,a long-term series of building information data sets of remote sensing images from 2001 to 2019 were constructed and the dynamic changes of the built-up landscape pattern were analyzed.From 2001 to 2019,the built-up expansion of Chongqing's main urban area was more obvious,with a total increase of 18.03%.The coverage area increased from 217.28 km2 to1204.06 km2,an increase of nearly 6 times,and the average annual growth was up to about 50 km2;the rate is lower,the growth rate around the city center is higher,and there is a significant expansion from the city center to the surroundings.The farther away from the city center,the faster the construction area growth rate;the landscape pattern of the main urban area changes greatly,the patch density increases,and the patch area decreased,the Shannon's diversity index increased significantly,and the landscape structure was diversified.The fragmentation degree of the architectural landscape pattern continued to increase,the dominance of the built-up landscape pattern continued to increase,and the shape of the built-up landscape became more complicated.