Research On Delineation Management Zones Of Precision Agriculture Based On Multi-Source Data

The rapid growth of the world's population has brought about food shortages.Countries have increased their food production by applying large amounts of chemical fertilizers,however,excessive application of water and fertilizers has caused problems of land pollution and waste of resources.Precision agriculture aims to improve such problems through variable fertilization,that is,to assess the precise needs of different plots and apply the correct amount of water and fertilizer according to the soil characteristics or yield of the farmland.Management zone delineation is the main method of precision agriculture.Most previous studies divided the management zones through a single variable or the synergy of multi-source variables.The indicators involved in the management zone delineation have evolved from soil grids sampled data to satellite imagery data,aerial flight data,and so on.Soil data can accurately reflect the nutritional status of cultivated land,but its application is limited because the data collection process requires a lot of manpower,material resources and time.In contrast,management zoning delineated by satellite image data and aerial data can be used for larger areas without destroying the cultivated land,and is more time-efficient.This paper takes the typical wheatmaize rotation field in North China as the research object,and uses the multi-year winter wheat production data,soil sampling data from different growth stages of wheat,and high-resolution satellite image data as data sources,the following studies are planned to be carried out:(1)Research on management zone delineation methods and scale effects based on years of yield data.Compared different clustering methods in delineate management zones based on years of yield data in the research area,and filter the delineation results.A set of evaluation indicators for accurate management scale effects were selected to evaluate the filtering results of different window sizes and different times,and try to find the best filtering scale,including the significance index of regional differences(variance reduction rate VR,mean change rate Mcv,standard deviation change rate STDv,coefficient of variation change rate CRv),and spatial consistency index(Kappa coefficient Kc).And spatial fragmentation index(plaque density PD,core area TCA,average core area MCA,aggregation index AI).The research results show that the delineation result of the Fuzzy C-means algorithm is better than the Iterative Selforganizing Data Analysis Techniques Algorithm.Filtering the delineation results can effectively remove isolated pixels or fragments caused by random mutation.According to the scale effect evaluation index,the window size in 9×9 and the filtering time in third are the best filtering scales,and the delineation results can provide a decision basis for the variable fertilization management of crop yield in the next year.(2)Research on delineation of management zones based on a variety of soil nutrients.Use the data interpolation results of the winter wheat jointing stage and the difference between the jointing and harvesting seasons respectively as the data source to delineate management zones.The optimal number of zones was determined by FPI and NCE.The changes in the coefficient of variation as well as other indicators of soil nutrient within the delineated zones were counted,and the significance test of the differences between the zones was performed using the F test.The research results show that the coefficient of variation of each zone reduced by 63.81%-88.57% when based on the soil nutrient interpolation results at the jointing stage;reduced by 30.45%-79.70% when based on the difference between the jointing and harvesting seasons.It is shown that the statistical data of most soil nutrients corresponding to each zone are extremely significant.In general,the homogeneity of soil nutrients in each zone are increased after delineation,and the differences between zones were obvious.The delineation results based on soil nutrients at the jointing stage can provide a basis for variable fertilizer topdressing at the later stage.The delineation results based on the soil nutrient difference between the jointing and harvesting seasons can guide the variable fertilization of early wheat in the next year.(3)Research on delineation of management zones based on high-resolution satellite imagery and multi-source data.The vegetation indices image data and multi-source data are delineated separately based on the Fuzzy C-means clustering algorithm,and the satellite image is segmented based on a multi-scale segmentation algorithm.Comparing and analyzing the spatial consistency of the delineation results,the results show that: the spatial consistency of the management zones delineation results based on the vegetation indices and based on the multi-scale segmentation technology reaches 73.536%;The spatial consistency reached 77.053% and 83.060% respectively when the delineation results based on multi-source data and soil nutrients and vegetation at the jointing stage.It proves that wheat growth in the same period has a greater correlation with soil nutrient content,and also proves the feasibility and superiority of management partitioning based on high-resolution satellite imagery.