基于信息融合的燃气环境监测技术研究

Research on Information Fusion-based Gas Environmental Monitoring Technology

针对传统燃气安全监测方法只针对于单一环境变量,并不能准确监测环境信息的问题,采用多传感器信息融合算法对燃气环境安全进行监测。首先使用滑动均值滤波算法消除监测数据中出现的异常数据与噪声;其次使用卡尔曼滤波算法对同质传感器数据进行数据级融合;最后,采用基于遗传算法优化BP神经网络对数据进行决策级融合。实验结果表明,相对于传统单一监测方法,基于多传感器信息融合的燃气环境监测系统准确率高,可靠性好,在燃气安全方面具有良好的应用场景。

In response to the limitations of traditional gas safety monitoring methods,which only focused on a single environmental variable and could not accurately monitor environmental information,a multi-sensor information fusion algorithm was used to monitor gas environment safety.Firstly,the sliding mean filtering algorithm was utilized to eliminate abnormal data and noise in the monitoring data.Secondly,the homogeneous sensor data underwent data-level fusion using the Kalman filtering algorithm.Finally,a decision-level fusion of data was performed using a BP neural network optimized by a genetic algorithm.The experimental results indicate that,compared to traditional single-monitoring methods,the gas environment monitoring system based on multi-sensor information fusion achieves high accuracy and reliability,demonstrating excellent applicability in the field of gas safety.