The Application Research Of Morphology Filter In Noise Reduction On Bridge Health Monitoring Singal

As bridge structure is usually in a complex environment,the sensor sampling signal obtained from bridge health monitoring system inevitably contains some noise components.It is the noise components that lead to reduction in the identification precision of modal parameters of the bridge structure so that the bridge structure can not be accurately evaluated.As an efficient nonlinear filter,the morphological filter has started to play a role in the vibration signal denoising field through many years’ development.There are certain applicable conditions and shortcomings in common methods of noise reduction for bridge health monitoring signal.The introduction of morphological filter into the field of noise reduction in bridge health monitoring signal means to seek a more simple and efficient way to reduce noise.The paper mainly covers the following contents:1、It introduces several common methods for reducing noise in bridge health monitoring signals and their shortcomings in application;elaborates the emergence and development of mathematical morphology,introduces the research status of one dimensional gray-scale morphology filter in the vibration signal denoising field in details,and presents the shortcomings of the existing research;2、It introduces basic theories of binary morphology,gray-scale morphology and one dimensional gray-scale morphology;3、According to the types of noise in the bridge health monitoring signal,the generalized mean morphological filter is constructed,which proves to be able to effectively remove the white noise and impulse noise in the signal through simulation test;it takes the signal-to-noise ratio(SNR)and mean square error(MSE)as the index to evaluate the noise reduction effect,and chooses triangle as the structural element through simulation test and puts forward an empirical formula for determining the size of a triangle and proposes a self-adapting morphological filtering method accordingly;4、The stochastic subspace identification method is used to identify the modal parameters of the bridge structure by using the unprocessed original sampling data,the data obtained after denoising through the generalized mean morphological filter,the data after the denoising by the traditional EMD method and the data after the AEEMD denoising,and draw the stabilization diagrams.The results prove that the denoising effect of morphological filter is superior over EMD denoising method,and the calculation efficiency is higher than AEEMD method.In addition,it can effectively reduce the noise components contained in the test signal,so as to make the modal parameter identification result of bridge structure more stable.