Font Size: a A A

Information Extraction And Pattern Analysis Of Rural Settlements Based On GF-2 Satellite Imagery

Posted on:2020-12-25Degree:MasterType:Thesis
Country:ChinaCandidate:G S LinFull Text:PDF
GTID:2370330572496620Subject:Cartography and Geographic Information System
Abstract/Summary:PDF Full Text Request
Conventional medium-and low-resolution satellite images and classification methods was difficult to achieve settlement information at the rural scale.With the vigorous development of high-resolution remote sensing technology in China,the quality of homebred high-resolution remote sensing products was getting higher,and it had been widely used in various fields related to production and life.In order to explore the applicability of the settlement information extraction of the domestic high-resolution image products at the rural scale,and provide reference for the work of rural settlement planning and rural land resource efficiency improvement during the construction of beautiful country.This paper took the Youxin town of Jintang County as the research area,and used the object-oriented method to extract the rural settlement information based on the homebred high-resolution remote sensing image,and analyzed the spatial pattern of the rural settlement in the study area.After the research in this paper,the following conclusions were obtained:(1)By using the Deviation test method,ESP2 and the RMAS,it was possible to effectively determine the optimal segmentation scale of objects based on GF-2 at the rural scale.The optimal segmentation scales of objects in the rural area including buildings,roads,bare land,water,Non-cultivated vegetation,cultivated land and shadows were 100,200,60,200,150,80 and 80.(2)By building the multilevel object information extraction model,the accuracy of object information extraction was good.On the basis of the optimal segmentation of various objects,the remote sensing knowledge and object information extraction model was constructed from the analysis of features such as the spectrum,texture and shape of the object.And the object information was extracted by the membership function method whose overall accuracy had reached 92.02%.(3)The rural settlement information was obtained by component information polymerization,and showed the high precision.The rural settlement components were aggregated by the methods of spatial location feature selection and map aggregation.The positional accuracy and area accuracy of the rural settlement map was 91.44% and 82.07%,and the comprehensive precision reached 86.75%.Compared with the official data survey work,the rural settlement information extraction method used in this study has certain efficiency and economy.(4)According to the research of rural settlement pattern in the Youxin town,it showed that rural settlements had a small scale in regional landuse pattern,high building density,complex geometric shapes and broken structures.At the same time,it also showed regional “large dispersion” and local “small concentration” characteristics of the rural settlement in the study area.The scale agglomeration of rural settlements had occurred in the vicinity of urban settlements and industrial parks.The scale ratio,density ratio,MSI,MPFD,Kernel Density and Hotspot Detection methods were introduced to quantitatively evaluate the spatial pattern of rural settlements in the town.In terms of quantity structure,the area ratio of rural settlements was only 3.15%.Moreover,the rual settlements had a large difference in scale.Among them,the rural settlements with building density greater than 50% accounted for 89.71% of the total number of settlements.At the same time,in terms of spatial structure,the geometric complexity and structural fragmentation of rural settlements on the east side was significantly higher than those on the west side.In terms of spatial distribution,the core of the desity showed the multiple groups and small scale characteristics,which mainly appeared in the shallow gully.And the scale agglomeration of rural settlements had emerged around urban settlements and industrial parks that indicated both urban residential areas and industrial parks can promote the concentration of local rural settlements.
Keywords/Search Tags:rural settlement, GF-2, object-oriented, image segmentation
PDF Full Text Request
Related items