Research On Building Area Extraction And Dynamic Change Based On Remote Sensing Data

This paper aims to the study of building area extraction.The methods of building area extraction were compared and improved.On this basis,the Binhai New Area of Tianjin City was taken as an example,the change of building area was analyzed and the change rule model was establish.First of all,five methods including the Support Vector Machine(SVM),neural network,CART(Classification and Regression Tree)decision tree,building index,and Rule-based object-oriented method were used to extract building area based on the SPOT 6 and Landsat 8 data.Different building area extraction rules were established for SPOT 6 and Landsat 8 images respectively,and the SBEI(SPOT 6 Building Extraction Index)for SPOT 6 image was created.The extraction accuracy of different methods was verified and compared based on the verification samples and measured data.Then,five remote sensing images,in the year of 1996,2001,2006,2011,and 2016,were used to analyze the area change,focus transfer,morphological change,and driving factors of building area from 1996 to 2016 in Binhai New Area.Finally,a combination model of building area change was established,and the change law and trend of building area in Binhai New Area were obtained.This study can provide certain technical support for building area extraction and other feature information extraction of different data sources,and also provide an assist for planning and making decision in Binhai New Area in the future.The main conclusions of this paper include:(1)For the SPOT 6 images,rule-based method can make full use of the features of the spatial features and texture features,the extraction precision was higher than other methods.The SBEI index method can be used to extract the building area on SPOT 6 images quickly.(2)For Landsat 8 images,the rule-based object-oriented method still has special advantages,but the extraction accuracy of SVM method and CART decision tree method is relatively higher.(3)From 1996 to 2016,the area of building area of Binhai New Area was lasting increased,the expansion rate and expansion intensity were also increased,among which the Tanggu District with the largest contribution rate changed the fastest.The development focus transferred to the northeast,the building area morphology became more complex and the aggregation degree became higher.(4)The main factors affecting the building area change in Binhai New Area were population growth,economic development,traffic development and policy orientation.The registered population,urban employment population,GDP(Gross Domestic Product),GDP of the second industry and the GDP of the third industry had a high positive correlation with building area change.The spatial distribution of the changed building area mainly related to the policy factors.(5)Compared with the single model,the combination model constituted by the autocorrelation model,shape index model and the population economy model,which considered the influence of various factors on the building area change,can better reflect the variety regulation and the change trend of building area.