Research On Monitoring Technology Of Geologic Disaster Resistivity Method Based On Particle Filter

Landslide geological disaster is the second largest geological disaster after earthquake in China in terms of casualties,direct economic losses and percentage of total geological disasters.It is characterized by frequency,harmfulness and complexity.It is difficult for a single monitoring method to accurately reflect the whole process of landslide deformation.It is of great significance for landslide disaster prediction to combine the comprehensive data analysis and fusion obtained by different methods such as "sky-space-ground" multi-time multi-scale displacement monitoring,hydrological phenomena and geological monitoring,interaction monitoring of rock and soil mass,etc.According to the study of landslide deformation evolution mechanism,the macro surface displacement monitoring data cannot effectively reflect the internal deformation process and the deep displacement of the landslide body,and the relevant monitoring data are limited by drilling technology and other conditions,which cannot fully describe and explain the overall deformation evolution mechanism of the three-dimensional geological body in space.The physical composition of the weak surface(body)of the landslide and surrounding rock,structural porosity and water content are different,resulting in a variety of physical differences such as density,electrical property and elasticity of the landslide.In this paper,multi-dimensional imaging of electrical information of underground media is rapidly realized through apparent resistivity imaging technology and large-area four-dimensional electrode scanning.The existing landslide monitoring methods and technologies have their own advantages and disadvantages.However,most of the processing methods of collected data are based on the inversion imaging of a single collected item and a special section at each time in a known area,which to a large extent loses the correlation information between the data,resulting in poor prediction effect and failure to use the correlation information between the time series data to improve the reliability of prediction.Landslide monitoring data fusion is mainly based on surface macro displacement data such as GIS and In SAR,but the magnitude,speed and direction of surface displacement and displacement at deep sliding surface cannot be completely consistent,and the early warning index based on surface displacement time series cannot accurately reflect the deformation evolution law of deep rock and soil mass.The combined time curve using monitoring data such as borehole depth displacement,soil humidity,groundwater level and stress can reflect the deformation characteristics inside the slope body and identify the depth range and deformation trend of the potential sliding surface of the slope.However,this method is limited by technical conditions such as borehole and location,and cannot realize integral and continuous data collection in the internal space,resulting in incomplete modal data.How to integrate the existing monitoring data sets and establish an effective monitoring and prediction model for the three-dimensional complete deformation process of landslide geological bodies is the focus and difficulty of future research.In view of many problems of landslide depth displacement monitoring,high density resistivity method is adopted to monitor the deep displacement of landslide geology.Landslide slip surface is within the effective measurement range of direct current electrical exploration,and monitoring effect is better by using high density resistivity method.At the same time,the non-linear landslide resistivity data collected are passed through an improved particle filter algorithm model for landslide geological hazard monitoring to obtain the predicted value at the next moment,and the accuracy of the predicted value is analyzed through error.Through high-density resistivity inversion imaging,the electrical characteristics of landslide geology are presented through images,and the geological structure is analyzed.Landslide process is a complex dynamic evolution,and the material composition of landslide rock mass has anisotropic characteristics.The external uncertain dynamic environment affects the landslide system.Under the interaction of various factors,the landslide system presents the characteristics of nonlinearity and uncertainty that changes with time,that is,the displacement and deformation of the landslide has a nonlinear relationship with the internal and external factors of the landslide.According to the nonlinear characteristics of landslide monitoring data,the nonlinear particle filter algorithm is selected to filter the data and establish a prediction model to predict the data at the next moment.