A rain signal denoising method based on improved wavelet threshold function

Abstract

The present invention discloses a rain sound signal denoising method based on an improved wavelet threshold function. The method performs wavelet decomposition of a noisy signal on several layers to obtain wavelet decomposition coefficients of each layer, and then applies an improved threshold function to process the high-frequency wavelet coefficients, and performs wavelet reconstruction on the obtained new high-frequency wavelet coefficients and low-frequency wavelet coefficients, thereby obtaining a de-noised signal. When performing wavelet decomposition, the optimal wavelet basis and the number of decomposition layers are determined based on the signal-to-noise ratio and the root mean square error as evaluation indicators. The present invention utilizes the attenuation characteristics of exponential and logarithmic functions to propose a signal denoising method based on an improved wavelet threshold function. The method not only achieves the effect of signal denoising, but also avoids the problem of excessive signal processing. Compared with traditional methods, the present invention has better signal processing smoothness and better noise reduction effect, and provides new ideas and guarantees for obtaining pure rain sound signals to monitor environmental rainfall conditions.

Description

Technical Field

[0001] The invention relates to a rain sound signal denoising method based on an improved wavelet threshold function, and belongs to the technical field of signal processing. Background Art

[0002] In recent years, the quality monitoring of digital signals has had a direct impact on the work in many fields. For example, telemetry signals provide a basis for the analysis of flight data, CT images also rely on digital signals for auxiliary medical treatment, and rainfall monitoring also relies on the analysis and processing of rain sound signals. However, in actual engineering, the collected rain sound signals are often interfered by noise, and denoising is required to effectively express the useful information in the original signal.

[0003] The measurement of rainfall plays a vital role in the field of meteorology and hydrology. The change of rain sound signal can invert the development and dissipation process of rainfall, which is of great significance for timely warning, disaster prevention and mitigation. In the past, traditional rain gauges were generally used to measure the amount of rainfall. Most traditional rain gauges are siphon type, weighing type and tipping bucket type. However, traditional rain gauges are mechanically installed and the detection method is relatively complicated, which has great limitations.

[0004] Today, with the advent of the era of artificial intelligence, the detection methods of rain gauges have become more diversified and intelligent. In the field of acoustics, environmental rainfall conditions can also be detected by analyzing and processing rain sound signals. Since rain sound signals have non-stationary and nonlinear characteristics, they are susceptible to various noises during the acquisition process. Therefore, how to obtain pure rain sound signals has become the primary issue in studying the measurement of rainfall using rain sound signals.

[0005] Compared with the traditional Fourier transform, the wavelet transform has a wider range of applications and can denoise most non-stationary signals. It has significant characteristics in time-frequency analysis. One of the key points of wavelet analysis denoising is the selection of threshold function. In existing methods, soft and hard thresholds are usually used as threshold functions for signal processing. Among them, the soft threshold function has good continuity when processing signals, and the processing effect is smoother, but the denoised signal will have a constant deviation from the original signal. The hard threshold function can retain the detailed features of the signal to a certain extent, but the reconstructed signal may oscillate, which is caused by its discontinuity. Therefore, it is necessary to construct an improved threshold function to obtain a rain sound signal that is as pure as possible. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a rain sound signal denoising method based on an improved wavelet threshold function, so as to effectively solve the problem that the discontinuity of the hard threshold function commonly seen in wavelet threshold denoising leads to oscillation of the denoised signal, and the deviation between the wavelet coefficient estimated by the soft threshold function and the original signal, resulting in the easy distortion of the reconstructed signal.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A rain sound signal denoising method based on an improved wavelet threshold function comprises the following steps:

[0009] Step 1, introduce the threshold adjustment factor and construct a switchable threshold function;

[0010] Step 2, using signal-to-noise ratio and root mean square error as evaluation indicators, determine the optimal wavelet basis and decomposition level;

[0011] Step 3, according to the optimal wavelet basis and decomposition layer number determined in step 2, the noisy rain sound signal is decomposed by wavelet transform to obtain the low-frequency and high-frequency wavelet coefficients corresponding to each layer;

[0012] Step 4, using the switchable threshold function constructed in step 1 to process the high-frequency wavelet coefficients corresponding to each layer to obtain new high-frequency wavelet coefficients;

[0013] Step 5: Perform inverse wavelet transform on the low-frequency wavelet coefficients corresponding to each layer and the new high-frequency wavelet coefficients, i.e., wavelet reconstruction, to obtain the denoised rain sound signal.

[0014] As a preferred solution of the present invention, in step 1, the formula of the switchable threshold function is as follows:

[0015] ,

[0016] In the formula, is the threshold value, and , is the noise standard deviation, is the length of the signal, are the high-frequency wavelet coefficients before and after processing, are threshold adjustment factors, and , .

[0017] As a preferred solution of the present invention, the specific process of step 2 is as follows:

[0018] 1) Fixed the number of decomposition levels and selected the db6, sym6, db2 and coif5 wavelet bases respectively;

[0019] 2) Using wavelet transform, the noisy rain sound signal is decomposed to obtain the low-frequency and high-frequency wavelet coefficients corresponding to each decomposition layer under each wavelet basis;

[0020] 3) Use the switchable threshold function constructed in step 1 to process the high-frequency wavelet coefficients corresponding to each decomposition layer under each wavelet basis to obtain new high-frequency wavelet coefficients;

[0021] 4) Perform inverse wavelet transform on the low-frequency wavelet coefficients and new high-frequency wavelet coefficients corresponding to each decomposition layer under each wavelet basis, i.e., wavelet reconstruction, to obtain the denoised rain signal under each wavelet basis;

[0022] 5) Calculate the signal-to-noise ratio and root mean square error under each wavelet basis, and select the wavelet basis with the best noise reduction effect as the optimal wavelet basis;

[0023] 6) When the optimal wavelet basis is determined, select different decomposition levels, perform the same operations from 2) to 4) on the noisy rain signal, calculate the signal-to-noise ratio and root mean square error under each decomposition level, and select the decomposition level with the best noise reduction effect as the optimal decomposition level.

[0024] A computer device comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor. When the processor executes the computer program, the steps of the rain sound signal denoising method based on the improved wavelet threshold function are implemented.

[0025] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the rain sound signal denoising method based on an improved wavelet threshold function are implemented.

[0026] Compared with the prior art, the present invention adopts the above technical solution and has the following technical effects:

[0027] 1. The present invention introduces a threshold adjustment factor to construct a switchable threshold function between soft and hard threshold functions. This function has excellent properties such as high-order differentiability, asymptotic and continuous, and can perform good denoising on different signals. Compared with traditional methods, it significantly improves the signal-to-noise ratio and reduces the root mean square error.

[0028] 2. The threshold function constructed by the present invention achieves a better denoising effect in the rain sound signal denoising processing. The threshold adjustment factor in the function also greatly improves the versatility of the method, so that the improved threshold function can be applied to denoising in more signal fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flow chart of a rain sound signal denoising method based on an improved wavelet threshold function of the present invention;

[0030] Figure 2 It is a comparison between the improved threshold function of the present invention and the traditional threshold function image;

[0031] Figure 3 is the constructed simulated rain sound signal and the simulated rain sound signal with added noise;

[0032] Figure 4 The coif5 wavelet basis is selected and the improved threshold function is applied to obtain the denoised signals when the decomposition levels are 1, 2 and 3 respectively. DETAILED DESCRIPTION

[0033] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be interpreted as limiting the present invention.

[0034] The present invention proposes a rain sound signal denoising method based on an improved wavelet threshold function, which performs several layers of wavelet decomposition on the noisy rain sound signal and obtains low-frequency and high-frequency wavelet coefficients of each layer. The high-frequency wavelet coefficients are processed using a new threshold function to leave more ideal coefficients. Then, the processed wavelet coefficients are reconstructed by wavelet to obtain a denoised signal, making it closer to the ideal signal, retaining effective information more completely, and significantly improving the denoising effect compared to the traditional threshold function. The overall process is as follows: Figure 1 shown.

[0035] The improved wavelet threshold denoising method includes the following steps:

[0036] 1) Apply wavelet transform and select appropriate wavelet basis and decomposition layer number to decompose the noisy rain signal. If the modulus of the corresponding wavelet coefficient is large, it means that the wavelet coefficient contains more useful signals. If the modulus of the corresponding wavelet coefficient is small, it is mainly noise signal. Therefore, the corresponding coefficients of low-frequency approximate signal coefficients and high-frequency detail signals of each layer are obtained.

[0037] In the application of wavelet transform, the correct selection of wavelet basis is crucial to the effect of signal processing. In the process of noise reduction, the present invention adjusts the threshold parameters carefully to achieve the best noise reduction effect by debugging the wavelet basis and setting a series of decomposition levels in the process of implementing the improved threshold function. Finally, the optimal wavelet basis and decomposition level are determined based on the two evaluation indicators of signal-to-noise ratio and root mean square error. The larger the signal-to-noise ratio, the smaller the root mean square error and the better the noise reduction effect. On this basis, according to the data in Table 1, the present invention adopts the coif5 wavelet basis, in which the threshold function uses the improved threshold function, and the parameter .

[0038] Table 1 Comparison of denoising effects of different wavelet bases when using the improved threshold function and the decomposition level is 1

[0039]

[0040] The number of decomposition layers can easily affect the noise reduction effect of the final output signal. When the number of decomposition layers is too large, the signal distortion is more serious, which is not conducive to signal processing. Similar to the selection of wavelet basis, the parameters of the fixed wavelet basis and the improved threshold function are debugged. It is found that when the number of decomposition layers is 1, the noise reduction effect is better. The present invention selects the decomposition layer number as 1.

[0041] 2) For high-frequency wavelet coefficients, an improved threshold function is used for processing.

[0042] The improved threshold function is shown as follows:

[0043] ,

[0044] In the formula, As the threshold adjustable factor, , They refer to the wavelet transform coefficients before and after denoising respectively; is the threshold value.

[0045] The threshold of the present invention adopts the global threshold method:

[0046] ,

[0047] In the formula, is the noise standard deviation, is the length of the signal. The noise standard deviation uses the std() function in Matlab.

[0048] By improving the threshold function formula, we can get:

[0049] 1. hour,

[0050] ;

[0051] 2. hour,

[0052] ;

[0053] 3. hour,

[0054] ;

[0055] 4. hour,

[0056] ;

[0057] So the improved threshold function is and , then the improved threshold function is continuous.

[0058] 5. hour,

[0059] ;

[0060] In summary, the threshold function is improved to is the asymptote.

[0061] The construction method of the improved threshold function is different from that of the general threshold function, mainly because a logarithmic function with attenuation characteristics is used in the construction process, which optimizes the defects of the traditional threshold function such as excessive noise reduction and discontinuity. Through cases 1-4, the function is continuous, so it is not easy to have severe oscillation. hour, and The distance is zero, so there is no constant deviation. As the threshold adjustment factor a and b Changing the parameter value of can make the method more widely applicable, so that the improved threshold function can be applied to denoising in more signal fields.

[0062] The present invention also conducts comparative analysis with four threshold functions. The new improved threshold function is compared with the existing threshold function. Figure 2 As shown, the results show that the noise reduction effect of the present invention is more superior. The threshold functions are as follows:

[0063] a) Compromise method:

[0064] ,

[0065] Where: To adjust the parameters, .

[0066] The compromise method is to adjust the parameters The threshold function is simple in form, but it does not avoid the disadvantages of soft and hard threshold functions. Oscillation and distortion will still occur after reconstruction, and the degree of distortion is affected by Influence.

[0067] b) Weighted method:

[0068] ,

[0069] Where: To adjust the parameters, .

[0070] c) Polynomial combination type:

[0071] ,

[0072] The threshold function is continuous and its first-order derivative is continuous. The reconstructed signal does not oscillate and has a certain tolerance for sudden signal changes. However, the function form is fixed and discontinuous, which is not the optimal processing for rain sound signals.

[0073] d) Improved Garrote type:

[0074] The representative ones are:

[0075] ,

[0076] in, The function is continuous, and the deviation also increases with Increases and gradually decreases, but the parameters are difficult to adjust.

[0077] Table 2 is a comparison of the processing effects of the improved threshold function and other threshold functions under the conditions of using the coif5 wavelet basis and a decomposition layer number of 2. It can be found that the improved threshold function has better processing effects than the other six more traditional threshold functions, with smooth waveforms, high signal-to-noise ratio, and small root mean square error.

[0078] Table 2 Comparison of denoising effects of various threshold functions when using the 'coif5' wavelet basis and the number of decomposition layers is 2

[0079]

[0080] 3) Perform inverse wavelet transformation on the corresponding coefficients of the high-frequency wavelets and the low-frequency wavelet coefficients of each layer obtained in step 2), and reconstruct the wavelets to output the processed signals.

[0081] The coif5 wavelet basis is selected by comparison, the decomposition layer number is 1, the improved threshold function of the present invention is applied for processing, and then wavelet reconstruction is performed to output the signal. Figure 3 is the constructed simulated rain signal and noisy rain signal. Figure 4 It is the signal after denoising using the improved threshold function and 1-3 layers of decomposition under the coif5 wavelet basis.

[0082] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the aforementioned rain sound signal denoising method based on the improved wavelet threshold function are implemented.

[0083] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the aforementioned rain sound signal denoising method based on the improved wavelet threshold function are implemented.

[0084] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0086] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0087] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0088] The above embodiments are only for illustrating the technical idea of ​​the present invention, and cannot be used to limit the protection scope of the present invention. Any changes made on the basis of the technical solution in accordance with the technical idea proposed by the present invention shall fall within the protection scope of the present invention.

Claims

1. A rain sound signal denoising method based on improved wavelet threshold function, characterized in that: The steps include: Step 1, introduce the threshold adjustment factor and construct a switchable threshold function; In step 1, the formula of the switchable threshold function is as follows: In the formula, λ is the threshold value, and σ is the standard deviation of noise, N is the length of the signal, w j,k , are the high-frequency wavelet coefficients before and after processing, respectively. a and b are threshold adjustment factors, and 0<a<1. Step 2: Using signal-to-noise ratio and root mean square error as evaluation indicators, determine the optimal wavelet basis and number of decomposition layers; the specific process is as follows: 1) Fixed the number of decomposition levels and selected the db6, sym6, db2 and coif5 wavelet bases respectively; 2) Using wavelet transform, the noisy rain sound signal is decomposed to obtain the low-frequency and high-frequency wavelet coefficients corresponding to each decomposition layer under each wavelet basis; 3) Using the switchable threshold function constructed in step 1, the high-frequency wavelet coefficients corresponding to each decomposition layer under each wavelet basis are processed to obtain new high-frequency wavelet coefficients; 4) Performing inverse wavelet transform on the low-frequency wavelet coefficients and new high-frequency wavelet coefficients corresponding to each decomposition layer under each wavelet basis, i.e., wavelet reconstruction, to obtain the denoised rain signal under each wavelet basis; 5) Calculate the signal-to-noise ratio and root mean square error under each wavelet basis, and select the wavelet basis with the best noise reduction effect, that is, the wavelet basis corresponding to the maximum signal-to-noise ratio and the minimum root mean square error as the optimal wavelet basis; 6) When the optimal wavelet basis is determined, different decomposition levels are selected, and the same operations of 2)-4) are performed on the noisy rain sound signal. The signal-to-noise ratio and root mean square error under each decomposition level are calculated, and the decomposition level with the best noise reduction effect is selected, that is, the decomposition level corresponding to the maximum signal-to-noise ratio and the minimum root mean square error is selected as the optimal decomposition level; Step 3, according to the optimal wavelet basis and decomposition layer number determined in step 2, the noisy rain sound signal is decomposed by wavelet transform to obtain the low-frequency and high-frequency wavelet coefficients corresponding to each layer; Step 4, using the switchable threshold function constructed in step 1 to process the high-frequency wavelet coefficients corresponding to each layer to obtain new high-frequency wavelet coefficients; Step 5: Perform inverse wavelet transform on the low-frequency wavelet coefficients corresponding to each layer and the new high-frequency wavelet coefficients, i.e., wavelet reconstruction, to obtain the denoised rain sound signal.

2. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the rain sound signal denoising method based on the improved wavelet threshold function as claimed in claim 1 are implemented.

3. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the rain sound signal denoising method based on an improved wavelet threshold function as claimed in claim 1 are implemented.

Images