Fusion Of Multi-source Remote Sensing Information For The Extraction Of Impervious Surface Area

The impervious surface refers to artificial surface cover,which is made up of building systems and traffic systems.Roof and square consisting of building systems,roads and parking lot composed of traffic system.On the one hand,impervious surface is the evaluation index of urbanization process,which can reflect the evolution and expansion of the city.On the other hand,the area,coverage and spatial pattern of impervious surface affect water circulation of city,whether regional change,hydrology and water quality,urban environmental pollution and other issues,therefore impervious surface is also the important reference of the environmental quality assessment.Accurate and effective spatial cartography of urban impervious layer can not only contribute to the urban environment management,but also have strong guidance to urban planning.There are many kinds of data sources and a variety of Remote sensing methods are used to extract the impervious surface.But the low precision of information extraction limited practical applications.It is an important research topic that how to improve the precision of impervious surface information extraction.Multi-source remote sensing images fusion can make full use of the complementary advantages between them and improve the accuracy of information extraction.In addition,research shows the classification error between different classifier is not exactly the same and the multiple classifiers combination can get higher accuracy than single classifier.From both multi-source remote sensing data fusion and multiple classifiers fusion considerations,this paper discusses how to combine the advantages of multi-source information and obtain more accurate thematic information extraction.First,the basic knowledge about the method,level,key technology of multi-source remote sensing data fusion and the based classifier composition,combination level,fusion algorithm of multi-classifier combination is described.Secondly,Some experiments were carried out to analysis the performance of all kinds of multiple classifier combination method.Finally,DS evidence theory is applied to multiple classifier and multi-source data fusion.To label evidence conflict areas by singular value decomposition(SVD),and then use the SAR data to build rule library to self-adaptive fuse.The experiment got a good classification result?The research about The method of impervious surface information extraction based multi-source data fusion,not only can get higher accuracy of impervious surface extraction result,but also provide more accurate basic data support for the related fields,and it has practical application value.At the same time,this paper is an exploration of multi-source remote sensing data fusion method,contribute to improve the processing technology of massive amounts of remote sensing data,and promote the development of remote sensing technology.