Study On Edge Detection And Fast Automatic Inversion Methods Of Magnetic Anomaly With Inclined Magnetization

Magnetic exploration is one of the most commonly used methods in geophysical exploration,which has the advantages of fast,cheap and good application effect.However,the measured magnetic anomalies are often affected by magnetization direction,superposition anomaly and so on.It is much difficult to detect the useful information directly from the total magnetic intensity anomaly(TMI).Data processing is the key way to obtain effective signal of magnetic sources.Edge detection and fast inversion of field source parameters are two commonly used processing methods.Therefore,this paper focuses on the methods of edge recognition and fast automatic inversion applied to the interpretation of magnetic anomaly with inclined magnetization.It provides a way for the interpretation of magnetic anomaly under strong remanent magnetization.Firstly,we deduced the conventional Euler deconvolution formula for the superimposed field source,and theoretically analyzes that the anomaly superposition is the main factor that affects the effect of Euler deconvolution.Higher order derivatives can be used to overcome this problem,but which are susceptible to high frequency interference.Model tests show that the inversion result by Euler deconvolution of first vertical derivative has higher degree of calculation accuracy in the situation of noise free.however,when the original anomaly contains noise,Euler deconvolution based on vertical derivative is seriously affected by noise,and the result was not as good as that of conventional methods.Upward continuation method can be used to suppress noise interference to improve the inversion result.Secondly,we present a new method based on the reciprocal of the analytic signal amplitude and its derivatives,to directly estimate the structural index and location parameters of 2D magnetic data.The model test verifies the correctness and superiority of the new method.Especially,the inversion accuracy of the new method to the deep magnetic source is much higher than the conventional methods.Finally,we give the concept of directional tilt-gradient on the basis of tilt angle,and six improved derivatives of directional tilt-gradient are derived to obtain rich magnetic information on different directions.On this basis,the total horizontal derivative of the directional Tilt gradient method,which is less affected by the direction of magnetization,is proposed to recognize the edges of inclined magnetizedmagnetic source.Moreover,the two linear equations(named directional Tilt-Euler method),based on the combination of three directional derivatives of conventional Euler deconvolution and the six improved derivatives of the directional tilt-gradient,are constructed for fast estimation of magnetic source location,depth and structural index.Model tests show that the total horizontal derivative of the directional tiltgradient can more effectively detect the edges of cuboids and the centers of thin dykes and spheres and does not produce false anomalies,the directional Tilt-Euler method has stronger agglomeration degree,better continuity and higher accuracy than tilt-Euler method.The two methods are applied to the interpretation and processing of aeromagnetic magnetic data over Tamusu area of Inner Mongolia,and the distribution information of underground magnetic sources is obtained.The results have a good correspondence with the geological data and the interpretation results of seismic data.In addition,the magnetic anomalies over the sedimentary distribution area are related to the fault structure and the hidden volcanic rocks.From this,27 fault tectonic belts and 4 large hidden rock mass are delineated.