Based On Multi-Sensor Data Fusion Algorithm Of Integrated Monitoring System

Today's security technology is developing rapidly,and the market's demand for security is getting higher and higher.Traditional security technology needs technological advancement to meet the needs of society,and security needs to develop in the direction of integration,digital,and professional.The integrated security system based on multi-sensor data fusion technology can more accurately and comprehensively obtain the data information of the measured object from multiple dimensions,and improve the accuracy of system alarms.Therefore,it is of great significance to apply multi-sensor data fusion technology to integrated security systems.The main research contents of this paper are as follows:(1)In order to solve the problem of multi-value bias in the ID3 algorithm,a new method of data fusion based on improved ID3 CAC_ID3(Confidence And Correlation-ID3)algorithm in multi-sensor integrated security system is proposed.CAC ID3 algorithm introduces attribute confidence to adjust the expected entropy on the basis of the algorithm,and its value depends on the relevant domain knowledge.Then the correlation degree is introduced to adjust the information gain value and improve the classification accuracy.The experiment analyzed and compared the F1 value and correct rate of the 4 sets of UCI data sets and the alarm data of the multi-sensor integrated security system when using the CAC ID3 algorithm.The CAC_ID3 algorithm is found to be feasible and effective through experiments.(2)In order to deal with the problems in unbalanced data,optimization research is carried out,a TSMOTE+ENN hybrid sampling algorithm is proposed to balance the data set.In order to solve the multi-dimensional and unbalanced shortcomings of the monitoring data of the alarm accuracy of each sensor in the multi-sensor integrated security system,this paper adopts the under-sampling method and the over-sampling hybrid method,and applies the optimized TSMOTE+ENN hybrid sampling algorithm to a certain Multi-sensor integrated security field.Random forest is used as the classifier,and the accuracy,recall,accuracy,F1 value,and AUC are the performance evaluation indicators.Through experimental analysis and comparison of 5 sets of UCI data sets and real security alarm data when using the ROC curve graph based on TSMOTE+ENN algorithm,it shows that the effect based on TSMOTE+ENN mixed sampling is better.(3)On the basis of the above research,the C/S mode and ASP.net framework are adopted to design and implement an integrated security system based on multi-sensors.The system mainly includes six modules:information management module,warehouse access control module,personnel positioning module,material detection module,automatic inspection module and video monitoring module.At the same time,the security system based on multi-sensor data fusion technology was tested.Experimental results show that the multi-sensor fusion algorithm can improve the accuracy and reliability of system alarms,reduce the rate of false alarms and false alarms,and effectively prevent illegal intrusion by blue parties.