Research On Extraction Technology Of Winter Wheat Planting Area In Kaifeng City Based On Sentinel-2A Satellite Remote Sensing Image

Winter wheat is one of the main food crops in China,and its planting area accounts for about 20% of the total planted area of national food crops.Kaifeng City is one of the important wheat planting and exporting bases in China.In recent years,it has been affected by the market supply and demand,the wheat planting area fluctuating greatly.Therefore,it is of great significance to obtain wheat information in a short time to guide agricultural production.Remote sensing technology has the characteristics of large-scale,fast,low-cost,short-period and massive information.It can forecast the demand of wheat area timely and accurately.It is the most effective means to extract wheat planting area at present.In this study,we take Sentinel-2A satellite remote sensing data as the data source on April 8,2018,and take the Kaifeng city in Henan Province as the research area,and use the county as the production unit to carry out the research on the extraction technology of winter wheat planting area.Firstly,we establish the interpretation signs of ground features based on the best band combination image,and through the analysis of the spectral curve characteristics of the objects,we completed the recognition of the objects.Then,we use(Support Vector Machine,SVM),object-oriented classification methods to extract the winter wheat planting area,and compare the characters of the two classification effects.Finally,in order to further improve the extraction precision of crop planting area,we combine the above two classification methods and propose a(Vector Object Oriented Area Extraction,V2OAE)to extract the winter wheat planting area of the whole study area more accurately.The main conclusions are as follows:1.Through the analysis of the multi-period vegetation time series change curve in 2018,we use April 8th as the best identification date,and through the comparative analysis of different band combination images of April 8,we use 11-8-5 and 4-3-2 band combination as the best band combination remote sensing image map.2.With the help of the field measured data,we establish the interpretation signs of ground features based on the best band combination image,Through analyze the spectral characteristic curves of ground objects,and we find that Band4 and Band8 can better realize the identification of objects in different places.3.The V2 OAE classification method is better than SVM classification and object-oriented classification in classification effect,which can better solve the phenomenon of “pepper and salt”and eliminate unnecessary influence factors,making the extraction result of wheat more accurately.4.Due to the short launch time of Sentinel-2A,there is little research on crop area extraction.we propose the technical methods can provide exploration experience for the application of this data source in crop acreage extraction,and also provide certain technical support for the agricultural departments in this area to grasp the situation of winter wheat crop cultivation and formulate agricultural policies.