论文部分内容阅读
针对黄河含沙量测量易受环境因素影响而导致测量结果不准确的问题,提出基于卡尔曼和BP神经网络(Kalman-BP)的协同融合模型,将含沙量、水温和流速等传感器输出值经过卡尔曼滤波器进行滤波处理;然后经BP神经网络模型对含沙量信息和环境量信息进行多传感器数据融合;最后建立了含沙量测量的反演模型.为了比较Kalman-BP神经网络的协同处理方法的融合效果,在相同环境下还进行了一元线性回归模型和多元线性回归模型的含沙量数据处理,并进行了误差分析比较.实验结果表明,Kalman-BP神经网络协同融合模型的测量误差较小,提高了含沙量测量系统的精度.
In view of the problem that the measurement of sediment concentration in the Yellow River is sensitive to the inaccurate measurement results caused by the environmental factors, a collaborative fusion model based on Kalman-BP neural network (Kalman-BP) is proposed. The sensor output values such as sediment concentration, water temperature and flow rate Then filtered by Kalman filter, and then the multi-sensor data fusion of sediment concentration information and environmental quantity information is carried out by BP neural network model. Finally, an inversion model of sediment concentration measurement is established.In order to compare Kalman-BP neural network The synergetic effect of the co-processing method, the monotonous linear regression model and the multivariate linear regression model were also carried out to deal with the sediment concentration in the same environment, and the error analysis was carried out.Experimental results show that the Kalman-BP neural network collaborative fusion model The measurement error is small, which improves the accuracy of sediment concentration measurement system.