Remote Sensing Monitoring Of Spatio-temporal Distribution Of Crops In Black Soil Region

The information of the sown area and output of the crops is an important basis for the state to formulate grain policy and economic plan.The planting area of crops reflects the situation of utilizing agricultural production resources in the space.It is an effective way to understand important information about the distribution and characteristics of agricultural products,and is the basis for adjusting agricultural structure.Timely understanding crop sowing area,growth and yield is great significance for strengthening crop production management,further developing productive potential and assisting government departments in formulating scientific and reasonable food policy.The traditional way of yield estimation is mainly based on the survey and statistics method.According to the bottom up statistical report,the output data not only have poor timeliness,but also have great influence on human resources,and the accuracy of information is also not guaranteed.Monitoring crop planting information by remote sensing can timely and accurately get the planting area of crops,which is great significance for predicting crop yield,promoting farmland rotation and fallow system pilot,setting subsidy standard rationally and optimizing the utilization ratio of cultivated land.The remote sensing technology has many advantages,including quick access to speed,large covering area,abundant information,high practicality,low cost and so on.Therefore it is widely used in agricultural production activities,such as agricultural land use,agricultural resources survey,crop growth monitoring and yield estimation.In the precision agriculture,we first need to get the growth information of crops quickly.Based on the characteristics of remote sensing technology,it has obvious advantages in the rapid acquisition of crop information.At present,a large number of remote sensing satellites provide continuous and reliable data sources for agricultural applications.From high temporal resolution MODIS data to spatial resolution Landsat,HJ and GF,it provides ideal satellite data for the acquisition of crop information.This paper attempts to use remote sensing method to make a rapid and accurate estimation of the area of the typical black soil areas(Hailun,Wangkui and Lanxi).Based on ENVI and ArcGIS software platform,Hailun is selected as the study area based on the multi-source remote sensing image and the arable land area as the main data source.First,draw the sequence curves of different crop NDVI to determine the critical period of crop classification,and comparing various classification methods to determine the optimal method for crop classification,and then crop classification in the study area using the critical period of image and the optimal classification method in 2000,2009 and 2016.In order to prove the feasibility of the method,this study uses theinsured plots provided by the agricultural insurance company to verify the accuracy of the classification results.At the same time,the spatial and temporal distribution patterns and causes of the crops in the typical black soil area in the past 17 years were analyzed,and the following conclusions were drawn.(1)The main critical phase of crop extraction in June,August and September is determined through the main development period of crops and the MODIS-NDVI time series curve of different crops.At the same time,according to the study area of main crop growth periods and different crop reflectance different characteristics.We found rice need plenty of water irrigation in June,soil moisture differences led to different spectral characteristics,the farmland soil humidity in the false color display is the dark blue.The key period for the extraction of rice during this period.From NDVI time series curve,we can see that NDVI in the mid August is the highest,the NDVI of rice is the lowest.At present the soybean field is bright red,so this is the key period for the extraction of soybean.Soybean has matured in the mid September,the maturity of rice in the mid to end of September,Late September is the mature period of corn.At the same time,we can see the NDVI of soybean was lower than that of other crops,corn NDVI is the highest.The corn field is dark red and rice field is pink in this image,and this is the key period of the distinction between corn and rice.(2)Hailun as the study area,comparative supervised classification of the maximum likelihood,unsupervised classification of ISODATA algorithm and the method of combining supervised and unsupervised classification,using 1628 insurance coverage plots in Halun to evaluate the accuracy of three classification results.We found that the overall accuracy of three methods were 92.69%,86.40%,94.67%,and ultimately determine the method of combining the maximum likelihood method with ISODATA algorithm for optimal classification method.(3)The classification method using the maximum likelihood method and ISODATA algorithm combined with the classification of crop in 2000,2009,2016,Hailun,Wangkui,Lanxi,which includes the classification results of main crop types of maize,rice,soybean,wheat and other crops.The accuracy of the classification results is verified by 2972 verification samples from the agricultural insurance company.The overall accuracy classification accuracy is 95.09% and the Kappa coefficient is 0.90 in 2016.The accuracy of maize is 97.63%,the accuracy of soybean is93.25%,and the accuracy of rice is 95.65%.(4)In the three years of 2000,2009 and 2016,soybean planting area decreased first and then increased.Corn increased first and then decreased,which was mainly influenced by policy and agricultural production cost and grain income.There was a slight increase in rice until 2016,which was mainly affected by the climate.The distribution of soybean,corn and rice in space is mainly affected by topography.The distribution of wheat and other crops is mainly affected by the climate.(5)There are differences between statistical data and remote sensing monitoring.By step and step of crop statistical information,and then various administrative gather,it will waste a lot of manpower and time.The administrative statistical method ignores the spatial differences and isdifficult to apply directly to the spatial and temporal dynamic characteristics of crop spatial distribution.Remote sensing information can obtain time and space distribution of crops accurately and quickly,which has obvious advantages compared with traditional methods.