A Study Of The Fusion Algorithms Of Muti-source Gravity Data In Coastal Areas

In order to effectively combine the multi-source gravity data in coastal area, a lot of methods and schemes are respectively researched in time domain and frequency domain. The main works and creations are as the following:1. The main methods about recovery of the earth gravity field are analyzed. The applications of statistical method in physical geodesy especially in data fusion are summarized. Based on the current research the core research thought is designed respectively in time domain and frequency domain. The principle and target of the gravity data fusion are also established, which are the base of this thesis.2. According to the Forsberg's covariance model of disturbed potential, the practical covariance of gravity anomaly and the covariance of the gravity disturbance are deduced, of which the expression form are identical. The methods of determining the covariance parameter are analyzed and compared, and the effect of the covariance model is verified by practical example.3. The common expression about Least Square Stepwise Collocation is derivated and the flow chart is also excogitated. The realization of the stepwise collocation not only improve the calculation efficiency but also provide the tool which can qualitative describe and quantificational analyze the covariance model of multi-source gravity data.4. Based on the lucubration, discussion about the fusion methods in time domain, the adaptive fusion algorithm about multi-source gravity data is established in Stepwise Collocation framework. The core thoughts of this algorithm are adjusting the observation covariance by adaptive factor and some fiducial gravity data. The method effectively solves the problem of fusion of multi-source gravity data which have different altitude, different spectral sensitivity and different error. The adaptive fusion algorithm is the main creations in this thesis.5. The role and application of Input-Output System theory is described in detail. the research shows that the different classical theory can be transformed into frequency domain by using Input-Output System theory. the fusion of two kind of gravity data is derivated by two-input single output system, and the realization of adaptive fusion algorithm in frequency domain is also analyzed and generalizated.In order to verify the effect of the input-output system, the gravity field estimation and establishing geoid are realized by correlation method.6. The airborne gravity data, satellite gravity data and land gravity data are combined by two fusion modes, which are the adaptive algorithm based on collocation results and the adaptive algorithm based on observation signals. The results show that the algorithm based on observation signals has the better fusion effect and the 5 ' mean gravity anomaly is acquired under the algorithm. The comparison to shipborne gravity data indicates that the accuracy of fusion data is better than 3. 7×10?5 m/s2, and the systematic error is largely eliminated.7. Based on the remove-restore technique, the local geoid is established by using stokes method and Least Square collocation, the results show that both methods have the same accuracy. In order to establish the GPS leveling geoid, the gravity geoid is fitted by suing GPS leveling point under four-parameter model. The external verifying shows that the accuracy of GPS leveling geoid is about 0.19m.