PROBLEM TO BE SOLVED: To calculate the fractal dimension of a time series data with high precision by removing the influence of noise components which the time series data comprises.;SOLUTION: The frequency spectra of a time series data X[j] such as an altitude data outputted from a radio altimeter 1 or the like are calculated by an FET 3, and a cut-off frequency detector 4 calculates a cut-off frequency fc which divides a white region (noise component) and a color region (signal component) from the frequency spectra by a specified algorism. In a differential circuit 6, an LPF 7, an inverse fractal filter 8, and a white inspection portion 9, whether or not a frequency region exceeding the cut-off frequency fc out of the frequency spectra is white noise is determined. If the frequency region is not white noise, the detector 4 computes the cut-off frequency fc again. A fractal dimension calculator 5 calculates the fractal dimension of the time series data X [j], on the basis of the frequency spectra of a frequency region which is not higher than the cut-off frequency fc.;COPYRIGHT: (C)2003,JPO
展开▼
