Research On Support Vector Machine Aided Kalman Filter In Real Estate Field Investigation Technology

Based on overall consideration of basic conditions,project needs and technological possibilities,taking the real estate unit as the object,measuring and find out the real estate unit ownership,location and area etc,ensure that the real estate unit of clear ownership,clear defined location,clear relative space position,it is the main content of the real estate field survey.The main content of this paper is under the background of real estate survey technology,combining machine learning with optimal estimation theory,discusses the research of support vector machine aided kalman filter in real estate field survey technology.The specific research work is as follows:1.Research on parameter optimization of support vector machine.Selection of support vector machine parameters has significant impact on practical applications,this paper first introduced the theoretical content of genetic algorithms,then for support vector machine penalty factor C,Gauss kernel function parameter? and loss function parameter ?,this paper studies the optimization technology of parameters of support vector machine based on genetic algorithm,genetic algorithm optimization parameter process is given,and through the relevant experiments,the optimal parameters needed for the training of support vector machine are obtained,provide basic support for the use of support vector machine.2.Research on support vector machine aided adaptive kalman filter in single position alignment.First,the application of kalman filter in single position alignment is studied,then,aiming at many random factors,the accuracy of system model and noise statistical characteristics is reduced,the fuzzy adaptive kalman filter algorithm is introduced.Based on these foundations,this paper combines the support vector machine with adaptive kalman filter,determine adaptive factors through machine learning rules,then compare the support vector machine adaptive kalman filter,fuzzy adaptive kalman filter and kalman filter,analyze the performance of these three algorithms in single position alignment through simulation experiments,lastly,through semi physical simulation experiment,the effectiveness of the combined filter method in single position alignment is verified.3.Research on support vector machine aided adaptive kalman filter in loose combination in specific measurement stage.To the specific measurement stage for RTK frequent transient failure,support vector machine adaptive kalman filter is used as data fusion algorithm,speed and position is measured as the external variable,the loose combination system is designed.The simulation results show that: compared with a loose combination system with kalman filter and fuzzy adaptive kalman filter as data fusion algorithms respectively,the stability of the loose combination system based on support vector machine adaptive kalman filter is better.