Land Use Classification Of Alar City In The First Division Based On Landsat8 Remote Sensing Image

Using remote sensing image data and classification technology to extract and analyze land use is very important for many environmental and social applications,including agriculture,ecology and other social applications.At present,with the increasing number of land crop use classification algorithms,the accuracy is constantly optimized and improved.At the same time,the improvement of the availability of remote sensing data also brings the development of new digital pattern classification technology.However,most of them focus on the comparison of one or two or three kinds of surface features or the optimization of a single algorithm.There is less research on multiple complex types of surface features in the same research area,especially in southern Xinjiang,which has our unique geographical and geomorphic,climatic and hydrological characteristics.In this paper,four classification methods commonly used in supervised classification are used as theoretical and technical cases,combined with the unique geographical and hydrological characteristics of Southern Xinjiang,the land use classification of typical cities in southern Xinjiang is discussed,and the land use classification technology is verified and compared in different scenarios,so as to provide reference for the selection of land use classification extraction algorithm of typical cities in southern Xinjiang For theoretical reference.In this study,the first division of Alar city in southern Xinjiang is taken as the research area.Referring to the national land use classification standards,combined with the field survey data of the field sample points,as well as the digital elevation model,Google Earth,Baidu and other satellite reference information and auxiliary data,a more accurate interpretation mark and sample database of the region of interest are established.Based on the in-depth discussion of land use classification and extraction methods,and the support of GIS and remote sensing technology,this study uses landsat8 remote sensing image data,combined with the geographical climate and seasonal changes in southern Xinjiang,and selects alar city,with the least cloud cover,the most obvious water area change and the most effective remote sensing image in 2020 Two good satellite images of wet season and dry season.By adjusting the best eigenvalue of the classification algorithm,the remote sensing data are processed,and the land use information of the study area is obtained.On the whole,landsat8 in 2020 is the best Using OLI image as the data source,the land use of alar city in 2020 was classified by maximum likelihood,neural network,random forest and object-oriented technology.The accuracy of landsat8 ordinary multispectral(30M)and multispectral(15m)resolution of alar city after data preprocessing was verified This paper discusses the applicability of four common land use classification methods in the unique terrain area of Southern Xinjiang.The comprehensive situation shows that: compared with the other two methods,the object-oriented method and the neural network method can obtain higher classification accuracy in the case of 15 m resolution,and the object-oriented method can obtain higher classification accuracy in the case of 30 m resolution.Among them,the neural network classification method for the first division of alar city land use classification results and the actual category is the best,the overall accuracy of classification can reach 86%,of which the area of unused land and cultivated land accounts for more than 50% of the total area of the study area,the average classification accuracy of 30 m and 15 m resolution is 99.95% and 90.85% respectively.