离散小波变换与重力异常多重分解

DISCRETE WAVELET TRANSFORM FOR MULTIPLE DECOMPOSITION OF GRAVITY ANOMALIES

小波变换多尺度分析是位场分解的有益工具 .离散二维小波变换产生的低阶小波细节具有尺度不变的特征 ,它们不随小波变换的总阶数改变 ,而总阶数的增加仅增加高阶小波细节的个数和改变最后的高阶逼近 .因此 ,用小波变换分解重力异常时 ,可根据地质目的来组合小波细节 ,从而实现有地质意义的异常分解 .实验表明 ,用小波变换分解的重力异常小波细节 ,同样起到功率谱分解的作用 ,用分解之后的小波细节功率谱可以精确确定场源的埋深 .

Wavelet and multiscale analysis can be a powerful tool for multiple decomposition of potential fields data. Discrete wavelet transform generates low order wavelet details that keep invariant with the total order number;while changing the total order number can only increase wavelet details of the highest orders and degree of smoothness of the final approximation. Thus,one can combine wavelet details based on regional geology to decompose gravity anomalies,producing anomalies correlated to certain geological sources. Applying the traditional spectrum analysis techniques,one can estimate the everage depth of sources for every wavelet detail,correlated statistically the decomposed wavelet detail with meaning of the source depth. Typical examples for triple decomposition of local anomalies are given by using gravity data obtained from Chinese continental scientific drilling site. [