Research On Modeling And TOF Estimation Of Snow Depth Ultrasound Echo Signal

Snowfall is one of the most difficult factors in meteorological observation.Accurate snow cover data can help relevant departments to grasp the snow distribution data,provide a basis for water resources allocation,reflect the local cold degree,and also provide protection for disaster prevention and mitigation in transportation,power communication,industrial and agricultural production and other fields.Because of the particularity of snow cover structure,uneven surface and the interference of measuring environment including drifting snow,low blowing snow,wind and other factors,there are great errors in snow depth measurement.The monitoring of snowfall by meteorological departments in China is still based on manual work,which has many problems,such as high labor cost,low accuracy and poor timeliness.Ultrasound detection technology has been widely used in biomedicine,aerospace,industrial non-destructive testing and other fields with the advantages of low cost and high sensitivity.In this thesis,the idea of mathematical modeling was introduced into snow depth detection,and TOF(Time-of-Flight)estimation of snow depth ultrasonic signal was studied and analyzed based on the echo signal model,so as to improve the measurement accuracy.According to the needs of the project,a snow echo data acquisition device based on the ultrasonic amplitude detection method was designed.The main research contents are as follows:(1)According to the definition of acoustic parameters such as sound velocity and sound attenuation,the properties of acoustic parameters related to ultrasound in air were analyzed and understood.Based on the structure characteristics of snow cover,the propagation law of ultrasonic wave in snow cover was studied.(2)After analyzing and studying the characteristics of snow structure,according to the characteristics of ultrasonic echo signal,the idea of mathematical modeling was introduced into snow depth detection for the first time.The feature extraction of echo signal envelope was carried out based on Hilbert transform.After comparing the advantages and disadvantages of several ultrasonic echo signal models,a sine-cosine model modeling method for signal envelope was proposed,and the model parameters were modified by frequency domain analysis and MVSSLMS adaptive filtering algorithm.Experiments showed that the model had a good goodness of fit for signal envelope,and the detail information was preserved completely.(3)Based on the principle of ultrasonic snow depth measurement,the factors affecting snow depth measurement were studied.The key of snow depth measurement was the accuracy of time-of-flight estimation.A time-of-flight estimation method based on EEMD(Ensemble Empirical Mode Decomposition,EEMD)algorithm was proposed.Its principle was to decompose the signal envelope into detail information and backbone information.The first-order difference method was used to solve the point where the gradient of the signal changes most in detail information.Considering that the snow structure will change in the actual environment and accompanied by snowfall,the experimental design used this method to estimate the ultrasonic flight time,and studies the impact of different distances,different reflector properties and different density drift on the time-of-flight estimation.(4)According to the need of signal waveform,the ultrasonic echo signal acquisition device was designed.The control program of ultrasonic transceiver based on STM32 was compiled to control the work of the whole system.The design idea of push-pull circuit was used to generate single frequency sinusoidal wave.The third-order Butterworth low-pass filter was designed according to the idea of fixed-K filter,and the active band-pass filter was designed by series connection of low-pass filter and high-pass filter.The temperature compensation circuit was used to correct the sound velocity and improve the measurement accuracy.