Wavelet Threshold Denoising Parameter Selection Method Based on Composite Evaluation Index and Wavelet Entropy

Abstract

The object of the present invention is to provide a method for selecting a wavelet threshold denoising parameter based on a composite evaluation index and wavelet entropy, which includes the following steps: obtaining a backup wavelet base, performing wavelet denoising processing, calculating the composite evaluation index, and decomposing each wavelet base with different decomposition layers. The data sets composed of two traditional evaluation indicators are normalized respectively, each wavelet base obtains a set of composite evaluation indicators, and the optimal number of decomposition layers is determined. The optimal number of decomposition layers is determined by calculating the wavelet entropy of the low-frequency coefficients to determine the optimal wavelet base decomposed at each layer, and the optimal denoising scheme among multiple types of wavelet bases is determined by comparing the composite evaluation indicators. The invention determines the optimal number of decomposition layers by constructing a composite evaluation index, and calculates the low-frequency coefficient wavelet entropy to determine the optimal wavelet base decomposed at each layer. The optimization problem of two denoising parameters, and the number of decomposition layers.

Description

technical field [0001] The present invention relates to a denoising method, specifically a wavelet threshold denoising method. Background technique [0002] Observation data is always mixed with a large amount of noise information due to various factors, which brings difficulties to the extraction and identification of signal features. Therefore, signal denoising, which is one of the classic problems in the field of signal processing, is particularly important. The traditional denoising methods mainly include linear filtering and non-linear filtering. The disadvantages are that the entropy of the transformed signal increases, the non-stationarity of the signal cannot be described, and the correlation of the signal cannot be obtained. At present, in practical engineering applications, for non-stationary signals whose true value is unknown, the denoising methods mainly include wavelet denoising, Kalman filter denoising, particle filter denoising and curve threshold denoising. ...

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a wavelet threshold denoising parameter selection method based on composite evaluation index and wavelet entropy, which solves the optimization problem of wavelet base and decomposition layer number in wavelet denoising under the premise of reducing the amount of calculation as much as possible

Images