Remote Sensing Monitoring Research Of Irrigation Area And Planting Structure Extraction Method In Shenwu Irrigation Area In Inner Mongolia Hetao Irrigation District

China is a country with developed agriculture but lack of water resources.The rational distribution of water resources has become the basis for the healthy development of China's agriculture.The monitoring of irrigated area and planting structure is also an important basis for the rational allocation of water resources.Gaining importance.Based on the analysis of the monitoring methods of irrigated area and planting structure,the Shenwu Irrigation Area in Inner Mongolia Hetao Irrigation District was used as the research area.The feasibility of remote sensing technology in the arid and semi-arid areas,planting structure and irrigated area monitoring method was carried out.analysis.At the same time,the irrigated area of Shenwu Irrigation Area was predicted by mathematical statistics prediction model.The main achievements were as follows:(1)Based on the classification method of CART decision tree and SEE5.0 decision tree,the land use types in Shenwu Irrigation Area were classified,and the land use type map of Shenwu Irrigation Area was drawn.At the same time,based on field measured point data,two decisions were made based on the field.The comparison of the plotting accuracy of the land use type maps shows that the accuracy of the two decision tree classification methods is 83.5%and 88.3%respectively.Therefore,the classification results of the SEE5.0 decision tree are selected as the basis for the irrigation area and planting structure extraction.(2)Landsat-8 and MODIS satellite imagery were selected,and the decision tree classification method based on EVI time series was used to monitor the planting structure of Shenwu Irrigation Field.At the same time,according to the statistical data and field measured data,the accuracy was measured from both area and position.assessment.At the same time,combined with the soil distribution of each canal in the study area,the distribution relationship between planting structure and soil type was analyzed.The results showed that the classification accuracy of Landsat-8 and MODIS satellites for melon,corn and sunflower were 91.3%,92.9%and 91.4%,respectively.And 78.3%,85.9%,86.2%.It can be seen that the MODIS satellite data has poor classification results for crops with small planting area and no patch distribution.In terms of area accuracy assessment,the absolute error and relative error of Landsat-8 are lower than MODIS.Combining the soil type and the field size of each drainage area in the study area,the relationship between the two and the planting structure is analyzed and found.The distribution is less correlated with the two.(3)Based on the research of Vegetation Water Supply Index VSWI and Ts-NDVI feature space based temperature vegetation drought index TVDI,combined with field measured data,the irrigated area monitoring model based on VSWI and TVDI difference was constructed,and the SVM regression sliding prediction model and Logistic grey prediction were used.Based on the prediction results of the irrigated area of the Shenwu irrigation area,the accuracy of the remote sensing monitoring model and the mathematical statistical prediction model are evaluated.According to the accuracy evaluation results,the accuracy of the SVM regression sliding prediction model is significantly higher than that of the Logistic grey prediction model,and higher than the remote sensing.The monitoring results,but the prediction results will deviate from the actual development law with the increase of the prediction steps.In the remote sensing monitoring model,the accuracy of the irrigated area monitoring model based on VSWI and TVDI differences is 90.2%and 91.3%,respectively.The percentage of regional sampling points,with accuracy of 85.3%and 89.7%,respectively,demonstrates the feasibility of two remote sensing monitoring models,which can provide data support for irrigation management in the study area.