The application of multi-sensor target tracking and fusion technology to the comprehensive early warning information extraction of landslide multi-point monitoring data

Abstract Landslide disasters are the second largest geological disaster activity in addition to earthquake disasters. They cause economic losses of up to several billion yuan each year, resulting in thousands of deaths and injuries. Therefore, it is important to monitor and alert landslide disasters. Significance. In addition, after the occurrence of landslides, a large amount of deposits will be created, which will further breed mudslides. The ability to qualitatively and quantitatively extract landslide information at the first time is particularly important for disaster prevention, mitigation, and disaster management. Therefore, the extraction of landslide information is a significant measure. At the same time, due to the limitations of sensors, how to organically fuse data from multiple sensors to achieve tracking performance that cannot be achieved by a single sensor has become the focus of research and attention on multi-sensor data fusion technology in the field of target tracking. Based on the above background, the research content of this paper is the application of multi-sensor target tracking fusion technology to achieve comprehensive warning information extraction of landslide multi-point monitoring data. This paper proposes a spatiotemporal registration method and Ensemble Kalman Filter (EnKF) for target tracking Algorithm, and the prediction model is established through optimization problem and target tracking. Finally, the test simulation is carried out by selecting 30 monitoring data of a monitoring point of a new construction project as the detection value. The experimental results show that when the moving target model is unknown in actual work, the polynomial. The order of the fitting and the boundary conditions of the spline fitting are not easy to give, but you can optionally use the sensor measurement data in the middle part and then use the spline fitting algorithm to obtain a better time registration effect; In terms of spatial registration, in the Z-axis direction, the measurement error of sensor A and sensor B is large, and the maximum value exceeds 2000 m. The maximum value of sensor A error between the measured values and theoretical values of radar A and B after registration is about 1000 m, which is the Z axis direction, and the maximum error of sensor B is about 1000 m, which is also the Z axis direction. The error has been significantly reduced. EnKF is better than the UKF algorithm under non-Gaussian noise conditions. The PF algorithm itself is suitable for nonlinear and non-Gaussian systems and has the best tracking performance.