A windowed iterative random phase modulation method and device for signal enhancement

Abstract

The invention discloses a windowed iterative random phase modulation method and device for signal enhancement, comprising: performing peak detection on an acquired image signal, and performing windowing processing based on the peak detection result to obtain each window signal; generating a random phase matrix corresponding to the signal-to-noise ratio according to the signal-to-noise ratio of each window signal, converting each window signal to a Fourier domain, and applying the phase matrix to the window signal in the Fourier domain to achieve phase modulation, iteratively processing the phase-modulated window signal multiple times using the same phase matrix, and then transferring to the spatial domain; accumulating and normalizing the window signal after each iterative processing, connecting and integrating the normalized window signal according to the window position, and then normalizing the integrated signal to obtain a final phase modulated signal, which can achieve higher signal quality and stability in a high-noise environment, and is widely applicable to various application scenarios requiring denoising and signal quality enhancement.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a windowed iterative random phase modulation method and device for signal enhancement. Background Art

[0002] Digital signal processing and image processing technology are one of the most important technologies in the field of modern science and technology. These two technologies can convert real-time or offline signals into digital form and process and analyze them on digital computers. They play a very important role in various fields.

[0003] In the field of digital signal processing and image processing, denoising and enhancing signal quality have always been research hotspots. Traditional denoising methods include filtering technology, Fourier transform, short-time Fourier transform, wavelet transform, etc. These methods have certain effects in reducing noise and improving signal quality, but they also have some limitations. For example, Fourier transform and filtering technology have limited effects on suppressing high-frequency noise, and are insufficient in retaining signal details and edge information.

[0004] In recent years, some new technologies have been introduced into the field of signal processing to improve the processing effect. Random Phase Modulation (RPM) is one of the methods, which disrupts the noise distribution by introducing random phase in the Fourier domain, so as to achieve a more uniform noise distribution after the inverse Fourier transform back to the time domain. Although this method improves the signal-to-noise ratio of the signal to a certain extent, there is currently no random phase modulation method specifically for noise reduction. Existing methods are mostly used in other fields, such as signal encryption and secure communications.

[0005] The existing random phase modulation method has limited effect when applied once, and the processing results are easily affected by the initial phase, making it difficult to ensure consistency and stability. Without segmented processing, weak signals will be suppressed by strong signals, resulting in the loss of signal details. In order to improve the processing effect, an iterative random phase modulation method is proposed, which significantly improves the signal quality and stability through multiple iterative processing. However, due to the different phase distribution of each iteration, the processing process requires a lot of computing resources and the repeatability of the results is poor.

[0006] Peak detection and segmented processing techniques play a key role in solving the above problems. Peak detection is used to identify significant feature points in the signal, and segmented processing can optimize the signal near each feature point separately, thereby preventing weak signals from being suppressed by strong signals. When these technologies are combined with random phase modulation and iterative optimization, how to effectively integrate and optimize the processing process remains a difficult problem. Summary of the invention

[0007] In view of the above, the purpose of the present invention is to provide a windowed iterative random phase modulation and device for signal enhancement, so as to achieve technical effects such as improving signal quality and stability, avoiding weak signals from being suppressed by strong signals, improving processing efficiency and effects, and reducing the impact of artifacts and noise.

[0008] To achieve the above-mentioned object of the invention, an embodiment of the present invention provides a windowed iterative random phase modulation method for signal enhancement, comprising the following steps:

[0009] Perform peak detection on the image signal, and perform windowing processing based on the peak detection result to obtain each window signal;

[0010] A phase matrix corresponding to the signal-to-noise ratio is randomly generated according to the signal-to-noise ratio of each window signal, and the window signal is subjected to multiple iterative phase modulation processes using the same phase matrix, including: converting the current signal of each window to the Fourier domain through Fourier transform, applying a fixed phase matrix to the window signal in the Fourier domain to achieve phase modulation, and then converting it to the space domain through inverse Fourier transform, wherein in the first round of iteration, the current signal is the initial window signal without phase modulation, and in other rounds of iteration, the current signal is the window signal after phase modulation in the previous round;

[0011] After accumulating and normalizing the window signals after each iteration of phase modulation, the normalized window signals are connected and integrated according to the window positions, and then the integrated signals are normalized to obtain the final phase modulated signal.

[0012] Preferably, performing peak detection on the image signal and performing windowing processing based on the peak detection result to obtain each window signal includes:

[0013] Use the peak detection algorithm to detect the peak value of the image signal and find the peak value p whose significance is greater than the threshold P. i and its corresponding position l i , and according to the position l i The peak value p is determined by i The left boundary of s i and right boundary e i , based on each peak p i The corresponding left boundary s i and right boundary e i Perform post-processing of windowing and segmentation to obtain each window signal;

[0014]

[0015]

[0016] Where i is the index of the peak, N is the total length of the signal, and n is the number of peaks detected.

[0017] Preferably, generating a random phase matrix corresponding to the signal-to-noise ratio according to the signal-to-noise ratio of each window signal includes:

[0018] A data distribution corresponding to the signal-to-noise ratio is generated according to the signal-to-noise ratio in the window signal, and a sampling matrix is ​​generated by random sampling in the data distribution, and a phase matrix is ​​constructed based on the sampling matrix, wherein the data distribution includes at least one of a normal distribution, a uniform distribution, an exponential distribution, and a Poisson distribution.

[0019] Preferably, the sampling values ​​in the sampling matrix statistically satisfy the data distribution of the source, and constructing the phase matrix based on the sampling matrix includes:

[0020] Based on each sampling value A in the sampling matrix, each phase value phi=A×π is calculated, and exp(-1i×phi) is calculated according to the phase value phi, and exp(-1i×phi) is filled into the matrix position corresponding to the sampling value A to obtain the phase matrix.

[0021] Preferably, for images with a signal-to-noise ratio of 22-33 dB, the generated normal distribution has an arbitrary mean and a standard deviation of 0.35-0.75.

[0022] Preferably, for an image with a signal-to-noise ratio of 22-33 dB, the generated exponential distribution is: the mean is between 0.5 and 5.0.

[0023] Preferably, for an image with a signal-to-noise ratio of 22-33 dB, the generated Poisson distribution is: the rate of change parameter λ is between 0.4 and 1.2.

[0024] Preferably, the window signal after each iterative processing is accumulated and normalized, including:

[0025]

[0026] Among them, y i,final represents the normalized window signal, m represents the round index, M represents the total rounds, and y i ' ,m It represents the window signal of the i-th window signal after the m-th round of phase modulation.

[0027] Preferably, normalizing the integrated signal to obtain a final phase modulation signal includes:

[0028]

[0029] Among them, y i ′ represents the signal after the ith window signal is accumulated and normalized, y′ represents the integrated signal, min and max represent the minimum window signal and the maximum window signal in the integrated signal y′, respectively. i″ represents the final phase modulated signal.

[0030] To achieve the above-mentioned purpose of the invention, an embodiment of the present invention also provides a windowed iterative random phase modulation device for signal enhancement, including a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, they are used to implement the above-mentioned windowed iterative random phase modulation method for signal enhancement.

[0031] Compared with the prior art, the technical solution provided by the present invention can solve several problems existing in the prior art, especially the denoising and signal enhancement problems in the fields of signal processing and image processing, and has the following technical effects:

[0032] 1. Improve signal quality and stability: The existing random phase modulation method has limited effect when processing high-noise signals. The processing results are easily affected by the initial phase, and it is difficult to ensure consistency and stability. The present invention significantly improves the stability and repeatability of the processing results by introducing fixed phase distribution and iterative optimization.

[0033] 2. Avoid weak signals being suppressed by strong signals: In the prior art, weak signals are often masked by strong signals, resulting in loss of signal details. The present invention uses peak detection and segmented window processing technology to effectively identify significant feature points in the signal and processes the signal segments near each feature point separately, thus avoiding the phenomenon of weak signals being suppressed by strong signals.

[0034] 3. Improve processing efficiency and effect: Traditional denoising methods are insufficient in suppressing high-frequency noise and retaining signal details. The present invention can significantly improve the signal-to-noise ratio and overall quality of the signal with lower computing resource consumption by combining smoothing, random phase modulation and iterative optimization.

[0035] 4. Reduce artifacts and noise effects: Existing denoising and enhancement methods are prone to introduce artifacts and noise amplification problems when processing complex signals. The present invention gradually reduces the impact of noise, effectively suppresses artifacts, and improves the purity and clarity of the signal through multiple iterations of optimization and signal reorganization. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0037] Figure 1 is a flow chart of a windowed iterative random phase modulation method for signal enhancement provided by an embodiment;

[0038] Figure 2 is a result diagram of the effect of the iterative process on signal enhancement provided by the embodiment;

[0039] Figure 3 This is a graph showing the effect of segmented window processing on signal enhancement provided by the embodiment. DETAILED DESCRIPTION

[0040] To make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation methods described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0041] The embodiment of the present invention provides a windowed iterative random phase modulation scheme for signal enhancement, which can achieve higher signal quality and stability in a high noise environment and is widely applicable to various application scenarios requiring denoising and signal quality enhancement. Figure 1 As shown, the following steps are included:

[0042] Step 1: Perform peak detection on the image signal, and perform windowing processing based on the peak detection result to obtain each window signal.

[0043] In the embodiment, the acquired image signal is a one-dimensional signal, wherein the first column is the coordinate x i , the second column is the signal value y i After obtaining the image signal, the peak detection algorithm is used to detect the peak value, identify the significant feature points in the signal, and specifically find the peak value p whose significance (MinPeakProminence) is greater than the threshold P. i and its corresponding position l i , which is the peak detection result.

[0044] After obtaining the peak detection result, the one-dimensional signal y i Perform segmented windowing processing, specifically including: for each peak, according to the position l i The peak value p is determined by i The left boundary of s i and right boundary e i , based on each peak p i The corresponding left boundary s i and right boundary e i Perform windowing and segmentation processing to obtain each window signal;

[0045]

[0046]

[0047] Where i is the index of the peak, N is the total length of the signal, and n is the number of peaks detected.

[0048] In the embodiment, the signal is processed in segments according to the peak position, which can improve the local accuracy of the processing, avoid the suppression of weak signals by strong signals, and enhance the retention of signal details and features.

[0049] Step 2: randomly generate a phase matrix corresponding to the signal-to-noise ratio according to the signal-to-noise ratio of each window signal, and use the same phase matrix to perform multiple iterative phase modulation processes on the window signal.

[0050] In the embodiment, the phase matrix used for noise removal of each window signal is closely related to the distributions presented by the data. Specifically, the phase matrix corresponding to the signal-to-noise ratio is randomly generated according to the signal-to-noise ratio of each window signal, including: generating a data distribution corresponding to the signal-to-noise ratio according to the signal-to-noise ratio in the window signal, and the random distribution can be a normal distribution, a uniform distribution, an exponential distribution, or a Poisson distribution, etc., or a combination of two or more thereof. Random sampling is performed from the data distribution to generate a sampling matrix, and the sampling values ​​in the sampling matrix are statistically satisfied with the data distribution of the source, assuming that the sampling values ​​of the sampling matrix come from the normal distribution, and the sampling values ​​in the statistical sampling matrix also obey the normal distribution, and the normal distribution can be a normal distribution with a mean of 0 and a standard deviation of σ.

[0051] In an embodiment, a phase matrix is ​​constructed based on a sampling matrix. The specific process includes: calculating each phase value phi=A×π based on each sampling value A in the sampling matrix, and calculating exp(-1i×phi) according to the phase value phi, and filling exp(-1i×phi) into the matrix position corresponding to the sampling value A to obtain the phase matrix.

[0052] After obtaining the phase matrix, each window signal is iteratively phase modulated using the corresponding matrix to improve the denoising effect. In specific implementation, the window signal is transformed into Fourier space, and the fixed phase matrix is ​​applied to the window signal in the Fourier domain to achieve phase modulation and then converted to the space domain through inverse Fourier transform. In the first round of iteration, the current signal is the initial window signal without phase modulation. In other rounds of iteration, the current signal is the window signal after phase modulation in the previous round. The phase modulation is expressed by the formula:

[0053]

[0054] Among them, y i ' ,m Indicates the i-th window signal The windowed signal after the phase modulation of the mth round, and represent Fourier transform and inverse Fourier transform respectively, and exp(iφ) represents phase modulation using the phase matrix φ.

[0055] In the embodiment, the Fourier transform and the inverse Fourier transform are performed using the Fast Fourier Transform (FFT) algorithm. In the iterative phase modulation process, a fixed phase matrix is ​​used for phase modulation, which reduces the change of each iteration, effectively suppresses noise, and improves the stability and repeatability of the processing results.

[0056] The data after random phase modulation is iterated T times. The phase matrix φ used in each iteration obeys the same data distribution. The data processing effect is enhanced by iterating T times. At the same time, the influence of noise is gradually reduced through multiple Fourier transforms and inverse transforms to enhance the processing effect.

[0057] Step 3, after accumulating and normalizing the window signals after each iteration of phase modulation, the normalized window signals are connected and integrated according to the window positions, and then the integrated signals are normalized to obtain the final phase modulated signal.

[0058] In the embodiment, after the signal is subjected to multiple iterative phase modulation processes, the signal segments are reorganized and the overall normalization process is performed. Specifically, the process includes:

[0059] First, the window signal after each iterative phase modulation is accumulated and normalized, which can be expressed as:

[0060]

[0061] Among them, y i,final represents the normalized window signal, m represents the round index, M represents the total rounds, and y i ' ,m represents the window signal of the i-th window signal after the m-th round of phase modulation;

[0062] Then, according to the window position, the normalized window signal y is used i,final Replace the original window signal Then connect and integrate, including:

[0063]

[0064] Finally, the integrated signal is normalized so that the range is between [0,1] to obtain the final phase modulation signal, including:

[0065]

[0066] Among them, y i ′ represents the signal after the ith window signal is accumulated and normalized, y′ represents the integrated signal, min and max represent the minimum window signal and the maximum window signal in the integrated signal y′, respectively.i ″ represents the final phase modulated signal.

[0067] The windowed iterative random phase modulation method for signal enhancement provided in the above embodiment combines peak detection, segmented windowing processing, random phase modulation with fixed phase distribution and iterative optimization, and significantly improves the performance and efficiency of signal processing in multiple key aspects:

[0068] (1) Signal enhancement: The present invention uses a fixed phase distribution to operate in the Fourier domain, thereby maintaining the consistency of the phase distribution and significantly improving the signal quality. Through iterative optimization, important details in the signal are better restored, while the noise part is effectively suppressed;

[0069] (2) Avoid signal loss: Peak detection is used to accurately segment the signal, ensuring that weak signals are not neglected during processing due to the dominance of strong signals. This strategy is particularly suitable for processing complex signals with a large dynamic range, effectively preventing the loss of weak signals during processing.

[0070] The embodiments also provide specific experimental examples. Figure 2 The influence of the iteration process on the results in the random phase modulation process is compared. The original signals are all processed by segmented random phase modulation. Under the same parameters, the effect of 100 iterations is significantly better than that of no iteration. Figure 3 The effects of segmented windowing on the results during random phase modulation were compared. The random phase modulation was iterated 100 times. Under the same parameters, segmented processing can retain the signal peak with an intensity of about 0.3 in the original signal ( Figure 3 When the signal is not processed in segments, the signal peak with an intensity of about 0.3 in the original signal disappears completely ( Figure 3 middle green line).

[0071] Based on the same inventive concept, an embodiment further provides a windowed iterative random phase modulation device for signal enhancement, comprising a memory and one or more processors, wherein an executable code is stored in the memory, and when the one or more processors execute the executable code, the windowed iterative random phase modulation method for signal enhancement is implemented, specifically comprising the following steps:

[0072] Step 1, performing peak detection on the acquired image signal, and performing windowing processing based on the peak detection result to obtain each window signal;

[0073] Step 2, convert each window signal into the Fourier domain through Fourier transform, and generate a random phase matrix corresponding to the signal-to-noise ratio according to the signal-to-noise ratio of each window signal;

[0074] Step 3, applying the phase matrix to the window signal in the Fourier domain to achieve phase modulation, iterating the phase modulated window signal multiple times using the same phase matrix, and then converting it to the spatial domain through inverse Fourier transform;

[0075] Step 4, after accumulating and normalizing the window signals after each iterative processing, the normalized window signals are connected and integrated according to the window positions, and then the integrated signals are normalized to obtain the final phase modulated signal.

[0076] The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A windowed iterative random phase modulation method for signal enhancement, characterized in that: The following steps are involved: Perform peak detection on the image signal, and perform windowing processing based on the peak detection result to obtain each window signal; A phase matrix corresponding to the signal-to-noise ratio is randomly generated according to the signal-to-noise ratio of each window signal, and the window signal is subjected to multiple iterative phase modulation processes using the phase matrix, including: converting the current signal of each window to the Fourier domain through Fourier transform, applying a fixed phase matrix to the window signal in the Fourier domain to achieve phase modulation, and then converting it to the space domain through inverse Fourier transform, wherein in the first round of iteration, the current signal is the initial window signal without phase modulation, and in other rounds of iteration, the current signal is the window signal after phase modulation in the previous round; After accumulating and normalizing the window signals after each iteration of phase modulation, the normalized window signals are connected and integrated according to the window positions, and then the integrated signals are normalized to obtain the final phase modulated signal.

2. The windowed iterative random phase modulation method for signal enhancement according to claim 1, characterized in that: The peak detection is performed on the image signal, and windowing is performed based on the peak detection result to obtain each window signal, including: Use the peak detection algorithm to detect the peak value of the image signal and find the peak value p whose significance is greater than the threshold P. i and its corresponding position l i , and according to the position l i The peak value p is determined by i The left boundary of s i and right boundary e i , based on each peak p i The corresponding left boundary s i and right boundary e i Perform windowing and segmentation processing to obtain each window signal; Where i is the index of the peak, N is the total length of the signal, and n is the number of peaks detected.

3. The windowed iterative random phase modulation method for signal enhancement according to claim 1, characterized in that: The randomly generating a phase matrix corresponding to the signal-to-noise ratio according to the signal-to-noise ratio of each window signal includes: A data distribution corresponding to the signal-to-noise ratio is generated according to the signal-to-noise ratio in the window signal, and a sampling matrix is ​​generated by random sampling in the data distribution, and a phase matrix is ​​constructed based on the sampling matrix, wherein the data distribution includes at least one of a normal distribution, a uniform distribution, an exponential distribution, and a Poisson distribution.

4. The windowed iterative random phase modulation method for signal enhancement according to claim 3, characterized in that: The sampling values ​​in the sampling matrix are statistically satisfied with the data distribution of the source, and constructing a phase matrix based on the sampling matrix includes: Based on each sampling value A in the sampling matrix, each phase value phi=A×π is calculated, and exp(-1i×phi) is calculated according to the phase value phi, and exp(-1i×phi) is filled into the matrix position corresponding to the sampling value A to obtain the phase matrix.

5. The windowed iterative random phase modulation method for signal enhancement according to claim 3, characterized in that: For images with a signal-to-noise ratio of 22-33dB, the generated normal distribution is: the mean is arbitrary and the standard deviation is between 0.35-0.

75.

6. The windowed iterative random phase modulation method for signal enhancement according to claim 3, characterized in that: For images with a signal-to-noise ratio of 22-33dB, the generated exponential distribution is: the mean is between 0.5 and 5.

0.

7. The windowed iterative random phase modulation method for signal enhancement according to claim 3, characterized in that: For images with a signal-to-noise ratio of 22-33dB, the generated Poisson distribution is: the rate of change parameter λ is between 0.4 and 1.

2.

8. The windowed iterative random phase modulation method for signal enhancement according to claim 1, characterized in that: The window signal after each iterative processing is accumulated and normalized, including: Among them, y i,final represents the normalized window signal, m represents the round index, M represents the total rounds, and y i ′ ,m It represents the window signal of the i-th window signal after the m-th round of phase modulation.

9. The windowed iterative random phase modulation method for signal enhancement according to claim 1, characterized in that: The integrated signal is normalized to obtain the final phase modulation signal, including: Among them, y i ′ It represents the signal after the ith window signal is accumulated and normalized, y ′ represents the integrated signal, min and max represent the integrated signal y ′ The minimum window signal and the maximum window signal in y i ″ represents the final phase modulated signal.

10. A windowed iterative random phase modulation device for signal enhancement, comprising a memory and one or more processors, wherein the memory stores executable code, characterized in that: When the one or more processors execute the executable code, they are used to implement the windowed iterative random phase modulation method for signal enhancement described in any one of claims 1 to 9.

Images