Research On Urban Built-up Area Extraction Via Brightness Correction Indexes Based On NPP/VIIRS Images

The rapid development of urbanization has aggravated the disorderly expansion of cities and has led to a series of socio-economic and environmental problems.It is urgently necessary for a reasonable urban planning policy to ensure the orderly advancement of urbanization.Through urban built-up areas,not only can the location and scope changes of cities be monitored,but also the development dynamics and mechanisms of urbanization can be grasped.Then urban governance and planning policies can be formulated in a targeted manner.At present,effective and efficient extraction of urban built-up areas from nighttime light(NTL)remote sensing data has become a hot issue in related fields such as GIS,remote sensing,and urban planning.After combing the existing research,this research proposes a method of combining night light remote sensing images with new multi-source data to improve the extraction accuracy of urban built-up areas.Given the lack of innovation in current research data and the single observation angle,this paper introduces road network data and Point of Interest(POI)data and proposes Road Density & EVI(Enhanced Vegetation Index)Adjusted NTL Index(REANUI),POI Adjusted NTL Urban Index(PANUI),and POI and LST(Land Surface Temperature)Adjusted NTL Urban Index(PLANUI),all of which are based on Suomi National Polar-Orbiting Partnership / Visible Infrared Imaging Radiometer Suite(NPP/VIIRS)NTL remote sensing images,to extract urban built-up areas from multiple perspectives.Given the limited scope of the study area and lack of representativeness,this paper selects 9central cities in china that are more typical and regionally representative in terms of social economy,geographic location,and natural environment as the study area,and utilizes the reference comparison method,one of the threshold methods,to extract urban built-up areas from five types of data including NTL data and above indexes.Finally,this study selects the Vegetation Adjusted NTL Urban Index(VANUI)as the reference item and the high-precision urban built-up area data as the reference data source and establishes a precision verification system combining quantitative evaluation and qualitative analysis.The results are compared and analyzed by accuracy evaluation indicators and landscape pattern evaluation indicators,the method in this paper is comprehensively evaluated from the accuracy improvement and the rationality of the spatial distribution.The conclusions are as follows:(1)The method of urban built-up area extraction proposed in this paper based on multiple perspectives fully optimizes the extraction effect of urban built-up areas.This method combines NTL data with road network and POI data.NTL remote sensing data is based on the perspective of "high-altitude observation",and POI data and road network data are based on the perspective of human activities on the surface.The results prove that REANUI,PANUI,and PLANUI proposed in this paper can significantly improve the extraction accuracy of urban built-up areas based on nighttime light remote sensing.The overall effect is better than the VANUI.The average accuracy of the REANUI increased by about 4.5%,and the average accuracy of the PANUI and the PLANUI increased by about 8%.This proves that the REANUI,the PANUI,and the PLANUI can effectively compensate for the low spatial resolution of the NPP/VIIRS data.They can also suppress the light overflow effect and reduce the background noise.Thereby,the lack areas are made up,false patches are reduced,and the spatial distribution characteristic information of urban built-up areas is enriched.The final results become more accurate and more reasonable in spatial distribution.The method proposed in this paper has achieved good results in the 9 central cities,which proves the universality and superiority of the method.The 9central cities have their characteristics in terms of economy,politics,geographic location,and natural environment,and have certain typicality and regional representativeness.Therefore,the methods proposed can be further extended to research on the extraction of urban built-up areas throughout China and even the world.(2)The research results prove the applicability and advantages of road network data and POI data in the study of urban built-up area extraction and have broad application space in the field of urban research in the future.Among them,POI data has a better performance by giving play to the advantages of social perception big data.It not only performs best in terms of accuracy improvement but also can describe the spatial distribution of urban built-up areas from a finer scale.This contributes to the distribution of landscape patterns becoming more reasonable.(3)The research results confirm that the method proposed in this paper has better overall performance.It can efficiently obtain the location and scope changes of cities and can be applied to real-time dynamic monitoring of cities throughout China and even the world.Among them,the overall performance of the PANUI is the best.According to the results of urban built-up areas extracted from the PANUI,the accuracy has been significantly improved,and the spatial distribution is the most realistic.This research provides new ideas,new methods and data support for solving the problems caused by urbanization,guiding the formulation of urban planning policies and other urban research.