| The Huang-Huai-Hai Plain is one of the three great plains in China,located in North and East China.It is also one of the most important grain production areas in China.The main grain crops are winter wheat,corn,rice,etc.The types and spatial distribution of crops in a region or production unit are an important part of the spatial pattern of crops.Understanding the distribution of crops in a region is of great significance for studying agricultural development strategies,issuing agricultural support policies,and even guiding the national adjustment of economic development policies.Therefore,It is very necessary to extract the main planting areas of grain crops in the Huang-Huai-Hai Plain.The traditional methods of crop extraction are field visits and investigation and statistics,which have high requirements for human and material resources.Besides,they are time-lagged and susceptible to subjective factors of researchers,so they have certain drawbacks.Remote sensing means have the advantages of large area coverage,fast data change speed,rich and easy access to information,low cost,has been widely used in the global agricultural monitoring,land cover classification and other fields.The traditional remote sensing satellite data for crop extraction are mostly MODIS.Although it has the advantages of short revisit period and daily data acquisition,the spatial resolution is relatively coarse and the mixed pixel phenomenon will lead to low accuracy of ground object classification.The spatial resolution of Landsat satellite data is moderate,but the temporal resolution is 16 days.The growth cycle of food crops is short,and the important growth period is only a few months.If the weather environment such as cloud is disturbed,it is easy to miss the optimal identification cycle of food crops.Gaofen-1(GF-1)is the first satellite of "domestic Gaofen-1 series" launched by China in 2013.The WFV data of Gaofen-1 is free for users,and the data is highly accessible.Its 16-meter spatial resolution and 2-day revisit cycle have great advantages in the identification and classification of food crops.In this paper,the Huang-Huai-Hai Plain was taken as the study area,and the stubble and soil samples of winter wheat were collected in the typical area,and the indoor spectral experiment was carried out.Based on the Gaofen-1 image data,according to the phenological characteristics And spectral information of major grain crops,support vector machine Classification And CART(Classification And Regression Tree)decision Tree were selected to identify And estimate the planting areas of three major grain crops,namely winter wheat,corn And rice.Finally,the spatial distribution pattern of major grain crops in the Huang-Huai-Hai Plain from 2018 to2020 was analyzed.The main research contents and conclusions of this paper are as follows:(1)The indoor spectral experiment was carried out to obtain the hyperspectral data of winter wheat crop stubble and soil samples with different water contents.According to the correlation analysis,it was found that the sensitive band R2005 could accurately invert the water content of the samples.Compared with the traditional ratio water index,the sensitive band R2005 had a higher accuracy in the inversion of water content.It can effectively reduce the effect of water content on CAI of winter wheat.(2)according to the crop phenological characteristics and spectral information,in this paper,the maximum likelihood method,the minimum distance method and support vector machine(SVM)method and random forest law precision evaluation on the four methods,support vector machine(SVM)classification results show the classification accuracy is higher,so the final choice support vector machine(SVM)for the extraction of winter wheat,maize,rice crop planting area.Since soybean and maize are similar in phenology,it is difficult to distinguish them by a single spectral information.In this paper,NDVI values and NDVI differences of the two crops in important growing periods were used as features to construct CART decision trees to distinguish them.The overall accuracy and total accuracy of the final classification are about 90%.(3)The spatial distribution pattern of major grain crops in the Huang-Huai-Hai Plain was as follows: winter wheat was mainly distributed in the southeastern part of Hebei Province,the central and eastern part of Henan Province,the central and western part of Shandong Province,the northern part of Anhui Province,and a little in the eastern coastal area of Jiangsu Province.In the Beijing-Tianjin-Hebei region,Shandong Province and Henan Province,corn and winter wheat are distributed in similar areas.Henan Province is the province with the largest planting area of winter wheat and corn,and winter wheat and corn are mainly distributed in the plains.Rice is distributed in the central and southern parts of Anhui Province,the central part of Jiangsu Province,along the Yangtze River and around lakes,such as Chaohu Lake,Hongze Lake and Taihu Lake.The rice area in Anhui Province is slightly larger than that in Jiangsu Province.(4)In recent three years,the distribution of main grain crops in Huang-Huai-Hai Plain changed little.Compared with 2018,the planting area of winter wheat in the Huang-Huai-Hai Plain increased by 1.0% in 2020;Maize planting area in Huang-HuaiHai Plain increased by 0.95% year on year.Rice acreage in the Huang-Huai-Hai Plain decreased by 0.6 percent year on year.This study based on the high score 1 image,application of support vector machine(SVM)method and the CART decision tree method,successfully extract the huanghuai-hai plain nearly three years within the scope of the main food crop planting area,and analyzes the spatial distribution pattern,for the relevant departments supervise agriculture development,sum up the experiences of farming,agricultural policy to provide certain reference value. |
PDF Full Text Request