Research On Data Mining Methode Of Agricultural Machinery Operation Information

With the emergence and application of networking technology,more and more fields begin to focus on the data itself.The precision agriculture whose core is geoinformation system becomes a subject based on information technology field.Using sensor detection and data mining techniques,people can control farming in an accurate and timely manner to maximize benefits.The data mining algorithm studied in this thesis aims at clustering and classifying the data collected by the agricultural machinery sensors in the farmland information system,so as to help the managers to analyze and make decisions.First of all,the thesis finishes farming equipment track data pretreatment in the farming system.Through the Gauss projection transform,the original ellipsoidal coordinates are transformed to Cartesian coordinates.Then agricultural track data set is established and the trajectory coordinates are obtained by using the depth information.After that,the track coordinates are split by using distance and time interval.Using interpolation method makes up the defect in actual sampling.After this processing,the accuracy of the algorithm is greatly improved.Secondly,in the light of the process of farming area and calculation indexes needed to be calculated in the platform,this thesis proposes an area measurement metho,which can quickly calculate the exact value of a farming area.This algorithm measures the farming area of heavy plow and leakage rate as well.However,time series based on trajectory grows with cropping time steadily,the data volume will become very large.So this thesis uses the optimized grid clustering algorithm to extract block edge points and geometric center point to rep resent farming records.At the same time,map stacking algorithm based on data index composites different layers in the same region,so as to detect historical problems of a heavy plow.Therefore,management platform can provide guidance and basis for cultivation.Finally,this thesis proposes a farming quality evaluation method based on support vector machine classifier.Moreover,this classifier inputs various agricultural cultivation features,which are collected by area and clustering algorithm,into sequence minimizing model to train support vectors.Then by adjusting the penalty factor and kernel parameters,a support vector machine model with high accuracy in test set can be obtained.As a result,this algorithm can provide objective evaluation criteria for farming.