Study On Filtering Algorithm Of The Aerial Gravity Measurement Data Processing

The magnitude of gravity abnormity in aerial gravity measurement is less than 100 mGal,while that of the noise is a hundred or even a thousand times bigger. Thus, it is a key factor to obtain the whole information of gravity abnormity after removing the strong noise of dynamic sensor in the practice of aerial gravity measurement. In other words, filtering is an important section in the aerial gravity data processing,and a reasonable pass-band with a proper filtering method could improve the precision. This thesis mainly studies the data processing algorithms of aerial gravity and emphasizes on the filtering method of aerial disturbing gravity. Finally, survey data of CHAGS are used to validate the algorithms mentioned above. The main works and contributions are summarized as follows:1,With Newton's second law of motion, a mathematic model of aerial vector gravity measurement is established to discuss the methods of aerial gravity measurement data processing. Then, practical formulae of computing disturbing gravity based on the model are shown completely. As for the data processing flow, Etovos correction, horizontal acceleration correction, cross coupling correction, specific force of survey line, vertical acceleration, free-air correction and normal gravity computing model are studied in detail and the precision analyses are given followed.2,Two methods of obtaining the horizontal acceleration correction, that is to say, two steps and one step method are discussed in this thesis. In order to reduce the systematic errors which come from square item in the one step method, pre-filtering of the raw data used to compute horizontal acceleration correction is bound to be taken. Method of determining the scale of pre-filter based on damping system of stabilized platform is put forward, of which the pre-filter scale is correspond to that of stable period of the platform. The surveying data are pre-processed through FIR low pass filter,Haar wavelet,db4 wavelet,db6 wavelet and sym4 wavelet separately to obtain the proper wavelet used for pre-filter.3,In order to eliminate the large amount of noise,FIR low pass filter is used to process some surveying data. Due to the spectrum characteristics of aerial gravity data, technical standards are attained to achieve the window function of FIR low pass filter. Processing of surveying data shows that the FIR digital filtering has accurate linear phase and stable filtering course. Besides, the RMS of survey line crossing points is 3.25 mGal.4,To improve the precision of data processing,the method that aerial gravity data should be processed in time-frequency domain is put forward which includes data time-frequency analysis, the acquirement of time-frequency domain sustentation, the design of the time-frequency filter and the reconstruction of the time-frequency data. Then, the effect of time-frequency filter is validated through numerical experiments. The time-frequency filter designed is satisfactory to the aerial gravity data filtering processing. The processing of the surveying data validates the Wigner-Ville distribution and the feasibility of the time-frequency domain. At last, certain issues of the time-frequency domain in disturbing gravity data processing are listed.