Abstract: Derivative conversion, as a commonly used data processing method of potential field, has important physical significance. Due to the measurement error and noise, the derivative conversion in wavenumber domain is unstable,and the accuracy of derivative results is poor. In order to reduce noise interference and to improve the accuracy of derivative calculation, an arbitrary-order derivative conversion combining Chebyshev low-pass filter with routine derivative operator in wavenumber domain is proposed. Through analyzing the filtering properties of the filter and the characteristics of the radially averaged power spectrum of potential field anomalies, the parameters of the Chebyshev low-pass filter method are determined, which can reduce the interference of human choice. The 2D and 3D model tests show that compared with the other three methods,routine FFT (fast Fourier transformation), upward continuation and ISVD (integrated second vertical derivative), the root mean square error (RMSE) of the derivative calculated by this method and the theoretical value is the smallest, and the accuracy of its derivation result is higher. In the data processing of Bouguer gravity anomaly in Hulin basin, the second vertical derivative calculated by the proposed method is less affected by noise, and has higher reliability. Eight large faults and eleven small faults recognized based on this result are confirmed by previous studies.

Key words: Chebyshev low-pass filter, wavenumber domain, derivative conversion, radially averaged power spectrum