CN108508480A - Focus phase of output signal Distortion Detect method based on wavelet transformation - Google Patents

Abstract

A method for detecting phase distortion of a seismic source output signal based on wavelet transform, comprising the following steps: a) obtaining a seismic source output signal containing random noise; b) selecting a wavelet basis function and a decomposition level for the seismic source output signal to obtain the decomposed wavelet of the output signal coefficient; c) the wavelet coefficient value The absolute value and threshold of compare; d) the wavelet coefficient value after thresholding processing Perform wavelet inverse transform processing to obtain an output signal after suppressing noise; e) perform zero-crossing detection on the output signal after suppressing noise, and assign an output value, that is, phase detection is completed. By suppressing the output signal containing random noise through wavelet filtering, and then detecting the phase through zero-point phase detection, the suppression of random noise is realized. This algorithm processes data quickly and protects signal details well, which can solve zero-point drift and jump problem, so as to realize the accurate detection of the phase.

Description

Phase Distortion Detection Method of Source Output Signal Based on Wavelet Transform

technical field

The invention relates to the technical field of detection of seismic source output signals, in particular to a method for detecting phase distortion of seismic source output signals based on wavelet transform.

Background technique

When the exploration source is working, the coupling between the excitation end and the ground is often poor, or the excitation end is disturbed by human movement, etc., resulting in a large phase distortion between the actual control signal and the output signal of the source, which affects the signal-to-noise ratio of the signal and reduces the signal. Accuracy of test results. In view of this situation, most of the current phase detection adopts the zero-crossing method, but in the case of random noise, the zero point detection results have offset and jumping phenomenon, which seriously affects the accurate detection of the phase. Noise suppression methods mainly include methods based on mathematical statistical significance and adaptive filtering methods. The adaptive filtering method can adapt to the change of the environment very well, and adaptively change the filtering parameters to suppress random noise, but this algorithm is complex and affects the exploration efficiency. The algorithm of the low-pass filtering method is simple, but useful signals will be filtered out when there is frequency overlap between the signal and the noise. CN103344988A discloses a "phase detection method for vibroseis signals based on K-L decomposition" which uses a method in the sense of mathematical statistics to analyze the noise. suppression, but this method is not suitable for signals with poor stationarity.

Contents of the invention

In order to overcome the deficiencies of the above technologies, the present invention provides a phase distortion detection method of the source output signal based on wavelet transform to solve the phase distortion of the output signal of the source under random noise conditions.

The technical scheme that the present invention overcomes its technical problem adopts is:

A method for detecting phase distortion of a seismic source output signal based on wavelet transform, comprising the following steps:

a) Obtain the source output signal containing random noise;

b) select the wavelet basis function and decomposition level for the output signal of the seismic source, and obtain the wavelet coefficient after the output signal is decomposed;

c) Using the formula Compare the absolute value of the wavelet coefficient value W with the threshold δ, where W _δ is the value of the wavelet coefficient after thresholding, W ₁ is the wavelet coefficient value of the first layer, median is to obtain the median value, the points whose absolute value of the wavelet coefficient value W is greater than or equal to the threshold value δ are reserved, and the points whose absolute value of the wavelet coefficient value W is smaller than the threshold value δ are 0;

d) performing wavelet inverse transform processing on the thresholded wavelet coefficient value W to obtain an output signal after suppressing noise;

e) The zero-crossing detection is performed on the output signal after suppressing the noise, and an output value is given, that is, the phase detection is completed.

Preferably, the chirp signal is used as the seismic source output signal in step a).

Preferably, in step b), the seismic source output signal is processed by using db4 wavelet basis function and decomposing into four levels.

Preferably, the value of δ in step c) is 0.063.

Preferably, step e) assigns a value greater than 0 to 0.5 and a value less than 0 to -0.5 when assigning the output value. Preferably, the sampling rate of the chirp signal is 8000, the sweep time is 10s, the initial frequency is 40Hz, and the cutoff frequency is 500Hz.

The beneficial effect of the present invention is: by suppressing the output signal containing random noise through wavelet filtering, and then detecting the phase through zero-point phase detection, the suppression of random noise is realized. The algorithm processes data quickly, protects signal details well, and can solve zero-crossing The zero drift and jump problems that occur during detection, thus realizing the accurate detection of the phase.

Description of drawings

Fig. 1 is the curve diagram of original output signal and output noise-containing signal;

Figure 2 is a comparison diagram of the original output signal and the suppressed noise signal before and after;

Fig. 3 is the graph that contains noise output signal zero-crossing detection;

Fig. 4 is a graph of the zero-crossing detection of the output signal after the noise is suppressed.

Detailed ways

The present invention will be further described below in conjunction with accompanying drawings 1 to 4.

A method for detecting phase distortion of a seismic source output signal based on wavelet transform, comprising the following steps:

a) Obtain the source output signal containing random noise;

b) select the wavelet basis function and decomposition level for the output signal of the seismic source, and obtain the wavelet coefficient after the output signal is decomposed;

c) Using the formula Compare the absolute value of the wavelet coefficient value W with the threshold δ, where W _δ is the value of the wavelet coefficient after thresholding, W ₁ is the wavelet coefficient value of the first layer, median is to obtain the median value, the points whose absolute value of the wavelet coefficient value W is greater than or equal to the threshold value δ are reserved, and the points whose absolute value of the wavelet coefficient value W is smaller than the threshold value δ are 0;

d) performing wavelet inverse transform processing on the thresholded wavelet coefficient value W to obtain an output signal after suppressing noise;

e) The zero-crossing detection is performed on the output signal after suppressing the noise, and an output value is given, that is, the phase detection is completed.

By suppressing the output signal containing random noise through wavelet filtering, and then detecting the phase through zero-point phase detection, the suppression of random noise is realized. This algorithm processes data quickly and protects signal details well, which can solve zero-point drift and jump problem, so as to realize the accurate detection of the phase. It solves the problem that the traditional exploration seismic source is often affected by poor coupling between the excitation end and the ground when it is working, or the excitation end is disturbed by human movement, etc., resulting in large phase distortion between the actual control signal and output signal of the seismic source, which affects the accuracy of signal detection results The problem.

Example 1:

In step a), the chirp signal is used as the source output signal.

Example 2:

In step b), the output signal of the seismic source is processed by using the db4 wavelet basis function and decomposing into four levels.

Example 3:

The value of δ in step c) is 0.063.

Example 4:

Step e) When assigning the output value, assign a value greater than 0 to 0.5, and assign a value smaller than 0 to -0.5.

Example 5:

The sampling rate of the chirp signal used in step a) is 8000, the scan time is 10s, the initial frequency is 40Hz, and the cutoff frequency is 500Hz.

It can be seen from Figure 1 that the original signal is submerged in the noise due to the existence of noise, and the overall smoothness of the noisy signal is poor. In Figure 2, the noise-containing signal is suppressed to obtain the de-noising signal. It can be seen that the noise is effectively suppressed, the de-noising signal becomes smooth as a whole, and the de-noising signal is highly consistent with the original signal. In Figure 3, there are obviously multiple zero point jumps in the direct zero-crossing detection result of the noisy signal, which reduces the reliability and accuracy of phase detection. In Fig. 4, the zero-point detection is carried out after the noise-containing signal is suppressed, and it is obvious that the zero-point jump problem is effectively solved, thereby improving the reliability and accuracy of the phase detection.

Claims (6)

1. a kind of focus phase of output signal Distortion Detect method based on wavelet transformation, which is characterized in that include the following steps：

A) the focus output signal containing random noise is obtained；

B) by focus output signal selection wavelet basis function and decomposition level, the wavelet coefficient after output signal is decomposed is obtained；

C) formula is utilizedThe absolute value of wavelet coefficient values W is compared with threshold values δ, W in formula_δFor valve Wavelet coefficient values after value,W₁For first layer wavelet coefficient values, median is to seek intermediate value, wavelet systems Point of the absolute value of numerical value W more than or equal to threshold values δ retains, and point of the absolute value less than threshold values δ of wavelet coefficient values W is 0；

D) wavelet inverse transformation processing is carried out to threshold valuesization treated wavelet coefficient values W, obtains the output signal after compacting noise；

E) output signal after compacting noise is subjected to zero-crossing examination, and assigns output valve, that is, complete phase-detection.

2. the focus phase of output signal Distortion Detect method according to claim 1 based on wavelet transformation, feature exist In：Use chirp signals as focus output signal in step a).

3. the focus phase of output signal Distortion Detect method according to claim 1 based on wavelet transformation, feature exist In：Focus output signal is handled using the mode of 4 levels of db4 wavelet basis functions and decomposition in step b).

4. the focus phase of output signal Distortion Detect method according to claim 1 based on wavelet transformation, feature exist In：δ values are 0.063 in step c).

5. the focus phase of output signal Distortion Detect method according to claim 1 based on wavelet transformation, feature exist In：It is 0.5 that step e), which assigns the value more than 0 when assigning output valve, and the value less than 0 is -0.5.

6. the focus phase of output signal Distortion Detect method according to claim 2 based on wavelet transformation, feature exist In：The sample rate of chirp signals is 8000, sweep time 10s, original frequency 40Hz, cutoff frequency 500Hz.