The Study On The Application Of Multi-scale Wavelet Transforms Modulus Maxima In Boundary Detection Of Potential Fields

The boundaries detection of causative sources is an essential job in geophysicalexploration area, which can not only locate the boundaries of causative sources butprovide constraints for some inversion methods. The previous methods for boundariesdetection almost make use of different order derivatives of potential field data or theircombinations to locate the boundaries projection of causative sources on surface,however, the higher-derivative of potential field data is usually unstable, which willeasily distort the accuracy of boundaries detection result. In this paper, we present anew method for boundaries detection based on theory of wavelet—WaveletTransform Modulus Maxima (WTMM), which abandons using derivatives ofpotential field data to avoid the unstability of higher derivatives of potential field dataand gives a more accurate result.Wavelet technique is widely used in denoising of potential field signals andseparation of local and regional potential field anomGalies, however, it is very rare tosee its application in boundaries detection of potential field data. To test the stabilityof WTMM method in the application of boundaries detection, we apply WTMMmethod to the synthetic models data and real gravity data of Meishan iron ore and geta good performance for both the boundaries detection of models and iron ore.To get the forward models data conveniently, we make a Matlab GUI with Matlabprogramming language, which can model the synthetic forward gravity anomGalythat is caused by any number of right polygonal prisms. The user of the GUI caneasily get forward gravity anomGaly by just inputing corresponding parameters intothe GUI and can also choose plot function you want to display the anomGaly.In this paper, we deal with three different kinds of model gravity anomGalieswith WTMM method. First of all, we decompose the model gravity anomGalies forseveral scales with SWT2wavelet function which belongs to the wavelet toolbox ofMatlab software, then we detect the boundaries of models with WTMM method, however, the boundaries detection results of WTMM based on the SWT2function arealways with shifts which are bigger for larger scales. We abandon the use of SWT2function and program another wavelet function based on the consideration of thefeature of DWT2and SWT2, whose name is Advanced_Wavelet to do the multi-scaledecomposition for the model gravity anomalies. Applying the WTMM method basedon the Advanced_Wavelet function to the model data and real data of iron ore both getthe good boundaries detection results without shifts. By comparison of detectedboundaries of Meishan iron ore with WTMM method and previous boundariesdetection result of this ore by Xiao demonstrates the advantage and accuracy ofWTMM method in practical use.The successful application of WTMM method in boundaries detection ofpotential fields is mainly due to many good properties of wavelet technique. Themulti-resolution of wavelet allows analyzing the signal for different scales, bycomparing the boundaries detection result for each scale, we can give more reliableboundaries detection results. The classical principle of how to choose the thresholdprovides an effective tool for eliminating the fake boundaries. Due to the filteringproperty of wavelet, WTMM method also has the ability of denoising which makesboundaries detection results more accurate.