Research On Disturbance Noise Analysis And Wavelet Denoising Algorithm Of Magnetic Measurement

Due to the advantages of passivity and low cost,geomagnetic positioning has become a key technology in indoor positioning research in recent years.Geomagnetic location is a method to determine the spatial position of matching points through the correlation operation between the real-time measured geomagnetic sequence and the known geomagnetic reference database.Because the real-time measurement of magnetic values is easily affected by many factors such as observation equipment and environmental interference,the observation results will contain various types of random and colored magnetic noise,which will reduce the reliability and accuracy of geomagnetic matching results to a certain extent.Aiming at the uncertainty of magnetic numerical disturbance in real-time measurement,the characteristic analysis and noise reduction of magnetic disturbance noise are carried out,which provides a theoretical basis for improving the accuracy of indoor geomagnetic matching.In order to research the disturbance characteristics of different disturbance sources on magnetic measurement values,the self-made magnetometer HBEQ-1,FVM-400 magnetometer and JHC-03 CTM magnetometer were used to carry out magnetic disturbance measurement tests on reference points under the interference of pedestrians,transporters,gate gates and ramming.The average value,variance and frequency and other indicators were selected to analyze the noise characteristics of the corresponding magnetic disturbance,and the corresponding noise models were formed.It is found that the real-time measurement of magnetic value is affected by many factors such as equipment sensor and environment interference.Magnetic perturbation fluctuates from a few to a few hundred nanotesla;environmental interference can produce instantaneous strong noise ranging from hundreds to thousands of nanotesla or periodic noise ranging from hundreds to thousands of nanotesla.HBEQ-1 lacks system correction and produces bias noise of about 10000 n T~20000n T.Aiming at the characteristics of random noise,bias noise,instantaneous noise,periodic noise and other types of noise,moving average filtering,median filtering,wavelet transform and other algorithm models are used to carry out the real-time measurement of magnetic value denoising experiment.The results show that the denoising algorithm can remove some noise to a certain extent,and the signal-to-noise ratio can be improved in different degrees after denoising.Different denoising algorithms have certain applicability to noise types.Moving average filter has better effect on periodic noise,median filter has better effect on instantaneous strong noise,and wavelet denoising has better effect on random noise and bias noise.In addition,the selection of parameters of wavelet base type,decomposition level and threshold method will directly affect the effect of the actual wavelet denoising,which needs to be flexibly determined by combining the SNR before and after denoising in the experimental process.Under the conditions of coif5 wavelet basis,5-layer decomposition and soft threshold sqtwolog function,the denoising effect of real-time measurement magnetic data is relatively optimal.Figure 50;Table 25;Reference 58.