| In the eastern part of Hulunbeier city,the A Rong Banner,the Modlidava Banner,the Elunchun Autonomous Banner and the Zhalantun city are important food production bases and commodity food export bases in Inner Mongolia.It occupies an important position in the whole regional agricultural system.The healthy and sustainable development of agriculture plays a decisive role in ensuring the income of farmers in the whole region,stabilizing agricultural production and stabilizing the overall economic situation.But because of the region’s special geographical location and climate conditions,and human activities have been increasing in recent years.It has led to the region becoming a disaster-prone area.It has obvious universality,regional,seasonal and sustainable characteristics.Timely and accurate acquisition of disaster information is of great significance to post-disaster rescue,settlement of agricultural insurance claims,optimization and adjustment of planting structure,and accurate poverty alleviation.In this paper,a fast and accurate method to extract agricultural disaster information is proposed.Use Multi-phase MOIDS reflectance products,Based on different phenological regions and cultivated land types,the NDVI standard model is constructed to perform hierarchical processing,to eliminate noise interference in disaster areas,and to conduct zoning statistics.The eastern part of Hulunbeier city,the A Rong Banner,the Modlidava Banner,the Elunchun Autonomous Banner and the Zhalantun city was taken as the research area.To extract the disaster disaster information of the study area in 2016,and calculate the disaster-hit area and disaster-hit level of non-union cities.The information extraction of agricultural disasters in large scale and long time series was realized,and the lag and inaccuracy of traditional agricultural disaster monitoring methods were avoided.The main research contents and results are as follows:(1)Use two images of bare soil and growth periods,the farmland area of the research area was extracted by the iterative self-organization data analysis algorithm of unsupervised classification.Overall accuracy Pc is 0.78,Kappa coefficient PA% is 74.32%,the results have high precision.Then the image of irrigation period of paddy field was selected to distinguish dry land from paddy field.Overall accuracy Pc is 0.90,Kappa coefficient PA% is 80.91%,high accuracy of classification results,The application of remote sensing data and classification methods in this period is applicable to the distinction between arid and paddy fields.(2)Using MODIS reflectance products for 6 consecutive time phases from late June to late September,In the case of geographical location and type of cultivated land,Through standardized processing,NDVI classification,disaster information combination and noise elimination,The disaster information of the key growth period of crops in the study area in 2016 was obtained.The overall accuracy of the monitoring results was Pc(0.63)and Kappa coefficient(PA%)(78.16%).High precision of extraction results.It is proved that this method is suitable for large-scale and long-term agricultural disaster monitoring.Moreover,the disaster area,remote sensing index of the affected area,ratio index of the affected area and disaster level index of each city are calculated.In this paper,the whole disaster situation of each city has been deeply analyzed.(3)Based on disaster monitoring results,combined with geographical location,climatic conditions,soil type and topography of the study area,The corresponding Suggestions and countermeasures for disaster prevention and control are put forward.These include building protective forests,strengthening scientific water management,developing water-saving agriculture and dry farming,and building disaster warning systems.It contributes to the healthy and sustainable development of regional agriculture.The results show that: Using long phase of MODIS reflectivity products,based on the different geographical position and the cultivated land type,through a standardized processing and grade of NDVI,disaster information merging and noise elimination method,suitable for large scale and long time series of agricultural disaster monitoring. |
