Research On "Point Line And Surface" Distributed Sensor Information Fusion Method

In the increasingly frequent social production and economic activities,affected by regional natural environmental factors,disasters and safety accidents caused by temperature,harmful gases and other factors occur frequently.The general method is to use a single distributed sensor technology to monitor the research area,and to ensure the safety of production and life in the area by analyzing the sensor information.However,in the process of regional monitoring,in order to obtain sensor information with better precision and distribution characteristics,more than one distributed sensor is used.Affected by a variety of monitoring methods and ranges of distributed sensors,the sensor information presents a "point line surface" distribution.The "point" and "line"information are more precise,but the regional distribution characteristics are not obvious,and the "surface" information has obvious distribution characteristics.Therefore,this thesis proposes an improved kriging interpolation method to use the "point line surface" distributed sensor information,which further improves the information precision.A weighted regression Kriging method of "point line surface" information fusion is proposed.Based on the improvement of precision,improving the information distribution features.Finally obtain the results of sensor information fusion with high precision and better distribution characteristics.Main research contents are as follows:(1)This thesis proposes a scheme to improve the precision of sensing information to solve the problem of insufficient precision of sensing information.The inverse distance weighted interpolation method,spline function interpolation method and Kriging interpolation method are used to estimate the "point" and "line"information,and to fill in the missing information.By comparing the interpolation results,it can be found that for the same sensor information,the estimation error of kriging interpolation method is 5%?10%lower than the other two methods;for the same interpolation method,the interpolation estimation error of line sensor information is 8%?25%lower than that of point sensor information.The scheme to improve the precision of sensor information is that use kriging method to interpolate the "line" information,and use"point" information to correct the interpolation estimation results.(2)This thesis proposes a "point line" information fusion method based on improved kriging.The method improves the kriging interpolation algorithm of "line" sensing information.Using the "point"information to improve correction of results of kriging interpolation estimation,which solves the problem of further improving the precision of sensing information.Adjust the variable parameters of Kriging interpolation according to the fixed step size.Record the interpolation estimation results after each adjustment,and compare with the "point"sensing information at the corresponding interpolation.Select the nugget value and root mean square error as the evaluation index,when the nugget value is 0.02 and the root mean square error is 0.63,the two evaluation indexes are close to the minimum.The variable parameter is 0.5 and the root mean square error of the fusion result is 24%lower than that of the Kriging interpolation of the "line" sensing information,which further improves the precision of the sensing information.(3)This thesis proposes a weighted regression Kriging fusion method of "point line surface" information to solve the problem of lack of information distribution characteristics.The regression conversion relationship between the difference between the "point line" information and the "surface" information and the coordinate points is established,and the error of each point of the "surface" information is obtained and superimposed with it,and the sensing information estimation result of the"point line surface" weighted regression fusion is obtained.Furthermore,the information error of each original point is calculated,and the interpolation is superimposed on the "point line" fusion information to obtain the "point line surface" weighted regression Kriging fusion sensor information estimation results.By comparing and analyzing the results of each step in the process of information fusion,the average absolute error and root mean square error of the "point line area" sensor information fusion results obtained in this paper are reduced by more than 14%compared with the "point line" information fusion results,and the better sensor information distribution characteristics are retained.The "point line surface" distributed sensor information fusion method proposed in this paper has higher information precision and better distribution characteristics,and displays the information fusion results through visualization tools,which has certain research significance for the application of distributed sensor information in regional safety production and life monitoring and disaster early warning.