Research Of Extraction Method Of Winter Wheat Planting Area Based On Landsat8 Remote-sensing Image

Ghina is a large agricultural country in the traditional sense.Since a long time ago,a large population,farmland,has been troubled by problems of China's agricultural development.In this context,it becomes very important for food security.The long-term stability of the country,the healthy and stable development of social economy,are founded upon the basis of food security.Wheat is one of the three major cereal crops grown widely around the world,there are also widely cultivated in our vast land,timely,prepared wheat crop acreage information for maintaining national food security is concerned,its importance is self-evident.In the past,crop planting area of access to information normally rely on artificial statistics escalation form to complete,this mode of operation is not only time-consuming and laborious,and the end result is often large deviations with the actual results,accordingly to the current form of economic development is concerned,this method is not appropriate.Remote sensing technology is developed in recent decades,an Earth observation integrated technology,has many advantages such as economic,macro,timeliness,especially suitable for agriculture.Remote sensing in agriculture has become a very important in the development of precision agriculture technology.This study concentrates on winter wheat planting area extraction method,which are designed to more quickly and accurately extract the planting area of winter wheat,and to ensure a certain degree of accuracy,for large area remote sensing extraction of crop acreage offers a new solution.This research area:Zaoyang city,Hubei Province,Zhongxiang city,experimental video select the Landsat8 image is March 6,2015,by analyzing image features the main feature,based on the radiometric resolution of imaging key features category on pixel brightness value for statistical analysis,looking for differences between the objects,conditional on this construction of decision trees,extracting winter wheat acreage.Through this research was supported by the following:(1)by analyzing the Landsat8 on the band the correlation between the pixel gray value and the band's own statistical information,determine the experimental area the best band program,video is the 5th,4th,3rd band in turn give the red,green,blue,color composite.The color synthesis programme can better of highlight winter wheat,and rape of features,easy crop recognition and planting area extraction,and precision validation,work;(2)for winter wheat planting area larger of zaoyang,used based on like Yuan gray value building of decision tree model extraction get of winter wheat planting area data,its extraction precision for 92.00%,and for winter wheat,and rape are has planting of Zhongxiang,due to winter wheat,and rape both of spectrum features similar,So using based on like Yuan gray value building of decision tree model extraction has winter wheat,and rape total area;(3)using non-supervision classification on after gray stretch processing of Zhongxiang winter wheat,and rape total planting district image for again classification,get of winter wheat planting area extraction precision for 92.67%,and just with non-supervision classification get of classification results compared,by decision tree,and gray stretch processing zhihou of classification precision improve has 18.00%,precision upgrade more obviously.Thus,according to Landsat8 image-rich spectral information build a decision tree model,with a planting area of winter wheat,has a better classification accuracy,and model simple and easy to operate to enhance the timeliness of statistical data has a certain positive significance.Confusing area for winter wheat,rapeseed,extraction of decision tree models have limited precision,winter wheat and oil-seed rape could not be accurately distinguished by gray stretch the image,you can on the premise of full use of the existing classification techniques and theory,significantly improve the extraction accuracy of easily-confused area of winter wheat,which is a very good complement to the decision tree model.Combination of the above two extraction methods,acreage of winter wheat had not only fast and easy operation,extract high precision results,which for a large area of crop area extraction,easily-confused distinction of crops and other related research work provides a new way to solve the problem.