| Digital Elevation Model is the basic data of National Geographic Information,it is widely used in land use,urban planning,engineering planning,transportation planning and so on.High-precision DEM of cultivated land is the basic data for soil and water conservation,land renovation,black land protection,and high-standard basic farmland construction.The northeast black soil area is one of the three famous black soil areas in the world.It is famous for its deep black soil layer,a large amount of organic matter,and suitable for the growth of crops.Therefore,the northeast black soil area is called the "giant panda" in the land.The northeast black soil region is one of the three famous black soil areas in the world,which is famous for its deep black soil layer,a large amount of organic matter and suitable for the growth of crops.therefore,the northeast black soil region is called the "giant panda" in the land.However,less than a hundred years after reclamation,the thickness of the black soil layer declined rapidly,affecting the construction of high-standard basic farmland and seriously threatening the food security of our country.In accordance with the "National Sustainable Agricultural Development Plan(2015-2030)-Overall Plan for Governance of Outstanding Agricultural Environmental Issues(2014-2018)","Outline of the Northeast Black Soil Protection Plan(2017-2030)".Black land is the cornerstone of grain production capacity in Northeast China.protecting and improving the quality of black soil cultivated land and implementing comprehensive control of soil and water loss in the northeast black soil area is an important guarantee to keep the strategic bottom line of "basic self-sufficiency of grain and absolute safety of rations".It is an important ecological project clearly put forward in the outline of the 13 th five-year Plan,which is of great significance for ensuring national food security and strengthening ecological restoration.Therefore,the construction of cultivated land high-precision DEM in the black soil area to improve the geographical database of the black soil area is precisely to meet the needs of the country,urgent national urgency,brooks no delay.The redistribution of solar radiation,soil moisture and heat by topography affects the physical and chemical properties of soil and the spatial distribution characteristics of crop growth.Analyzing the relationship between topography and these variables is helpful to improve the accuracy of(DEM).The purpose of correcting Shuttle Radar Topography Mission(SRTM)data is to obtain high-precision DEM data in cultivated land.This study took the typical black soil topography as the research object,using the UAV(UAV)and a large number of measured ground elevation data to generate a high-precision reference DEM.Extract Normalized Difference Vegetation Index(NDVI),Perpendicular Drought Index(PDI)from SPOT-6 remote sensing images,and extract Potential Solar Radiation(PSR)from SRTM.Analyze the interaction between terrain and NDVI、PDI and PSR.The data sets of NDVI and SRTM,and NDVI,PDI,PSR and SRTM in June,July,August,and September 2016 were taken as independent variables,respectively,and the UAV DEM was taken as the dependent variable.Linear stepwise regression(LSR)and back propagation neural network(BPNN)were used to establish elevation prediction model.The applicability of this method was verified by Dongxing Village of Hailun and the Youyi Farm field block.The high-precision DEM obtained by this method was compared with the SRTM processed by low-pass filter and ZY3 DEM.The results indicated that:(1)Through the analysis of the relationship among topography,environment and crops from different angles of one-dimensional(point),two-dimensional(cross-section)and three-dimensional(space),we can know that topography mainly affects PDI and PSR,and then affects NDVI,it is the driving force to promote the material cycle of environment-matter-crop surface in cultivated land.(2)Whether it is using the data set of NDVI from June to September as the input or the data set of NDVI,PDI and PSR from June to September as the input.The result of nonlinear fitting was better than that of linear fitting,and BP neural network was the best method to improve the accuracy of SRTM.For example,the data sets of NDVI,PDI and PSR from June to September were used as inputs to improve SRTM.In test areas 2 and 3,the determination coefficients of the BP neural network were 0.940 and 0.920,respectively,and the root mean square errors was 0.943 and 1.234 respectively.Compared with the original SRTM,the accuracy was improved by 93% and 91%respectively.The determination coefficients of linear regression are 0.908 and 0.903 respectively,and the root means square error is 2.685 and 2.596 respectively,which is 82% and 83% higher than that of the original SRTM.At the same time,the spatial resolution is reduced from 30 m to 6m respectively.(3)Whether using linear fitting or non-linear fitting,the accuracy of improving SRTM with June-September NDVI,PSR and PDI data set is better than that of June-September NDVI data set.For example,using BP neural network method to improve SRTM.In the test areas 2 and 3,the determination coefficients of June-September NDVI as input are 0.847 and 0.913 respectively,and the root mean square error is 2.680 and 1.920 respectively.In test areas 2 and 3,the determination coefficients of NDVI,PDI,and PSR from June to September are 0.940 and 0.920,respectively,and the root mean square error is 0.943 and 1.234,respectively.In addition,the linear fitting results of NDVI,PDI and PSR from June to September indicate that the correlation coefficients of NDVI in June,July and September are-17.674,3.532 and 2.610,respectively.They are important variables to generate high-precision DEM.(4)In test area 2 and 3,the accuracy of DEM obtained by this method was higher than that of ZY-3 DEM and SRTM＿LOW,which was similar to the spatial pattern of UAV DEM,approaches the real surface,and could reflect the characteristics of erosion gully in the study area.In test area 4,the accuracy of the DEM obtained by this method is higher than that of SRTM DEM and SRTM＿LOW,which is similar to the spatial pattern of the obtained UAV DEM,and the distance between the test area 4 and the modeling area 1 is far away,which eliminates the limitation of distance to the construction of high precision DEM,and further proves the applicability of this method in space.(5)Compared the original SRTM with the improved DEM in contour planting and agricultural land regulation,the improved DEM contours were smoother and more consistent with the application of contour planting to the data.Superimposing two-dimensional cultivated land information could better supervise the unreasonable cultivation of cultivated land.According to the relationship between topography and crops and their environmental factors,discussed the feasibility of using it as an input for improving SRTM,and provided a quick method to obtain a higher precision DEM in the flooded area of cultivated land.The applicability of this method was evaluated in test areas 2,3 and 4,and discussed the application of improved DEM in cropland resource management.The results show that it is an efficient and feasible cultivated land microtopography monitoring technology,which can provide data basis for high standard basic farmland construction,black soil area protection,land renovation and so on. |
PDF Full Text Request