Intelligent Voice Noise Reduction Method and Related Devices Based on Satellite Communication

Abstract

The present invention relates to the technical field of satellite communication, and provides an intelligent voice noise reduction method and related equipment based on satellite communication, including receiving an initial satellite communication signal and classifying it into a target pure signal and a target noise signal; classifying the target noise signal into several subgroup signals, performing different noise reduction processes on different subgroup signals, and recombining them to obtain a noise reduction signal, synthesizing the target pure signal and the noise reduction signal to obtain a final communication signal, and converting it into voice data to verify the integrity of the voice data; if there is missing data in the voice data, tracing the source of the missing data and adjusting the corresponding noise reduction process until the voice data is complete. By performing different noise reduction processes on the target noise signal, and then verifying the integrity of the voice data and tracing and adjusting the missing data, it improves the problem that when dealing with complex or changing noise environments, it is impossible to effectively distinguish noise from signals, resulting in poor integrity of satellite signals.

Description

Technical Field

[0001] This application relates to the technical field of satellite communication, and in particular to an intelligent voice noise reduction method and related equipment based on satellite communication. Background Art

[0002] Satellite communication, as an important wireless communication means, is widely used in military, aerospace, ocean, emergency rescue and other fields. Satellite communication technology can cover most of the regions on the earth and provide efficient and stable communication services. With the continuous growth of global communication needs, the development of satellite communication systems has become even more rapid, especially in terms of improving communication quality and transmission efficiency, significant progress has been made.

[0003] In the prior art, voice noise reduction in satellite communication usually adopts traditional digital signal processing methods, such as filters, echo cancellers and adaptive noise suppression algorithms, etc. These methods can reduce background noise to a certain extent by processing the received satellite communication signals, and improve the clarity and quality of voice signals. For example, filters can filter out noise in specific frequency bands, and adaptive noise suppression algorithms can dynamically adjust the noise reduction strategy to adapt to different noise environments.

[0004] In view of the above technical solutions, although noise can be effectively reduced and the quality of voice signals can be improved through digital signal processing methods, in a complex noise environment, such as when multiple noise sources exist simultaneously or the noise spectrum overlaps significantly with the voice signal spectrum, there is a problem that noise and signals cannot be effectively distinguished, resulting in poor integrity of satellite signals. Summary of the Invention

[0005] In order to improve the problem that in the process of dealing with complex or changing noise environments, noise and signals cannot be effectively distinguished, resulting in poor integrity of satellite signals, this application provides an intelligent voice noise reduction method and related equipment based on satellite communication.

[0006] The present invention provides an intelligent voice noise reduction method based on satellite communication, including: receiving an initial satellite communication signal, classifying the initial satellite communication signal into a target pure signal and a target noise signal; classifying the target noise signal into several subgroup signals, and performing different noise reduction processes on different subgroup signals; recombining the subgroup signals after different noise reduction processes to obtain a noise reduction signal, synthesizing the target pure signal and the noise reduction signal to obtain a final communication signal; converting the final communication signal into voice data, and verifying the integrity of the voice data; if there is missing data in the voice data, tracing the source of the missing data, obtaining the subgroup signal corresponding to the missing data, and adjusting the noise reduction process corresponding to the subgroup signal until the voice data is complete.

[0007] As a preferred solution, the step of receiving the initial satellite communication signal and classifying the initial satellite communication signal into a target pure signal and a target noise signal includes: converting the initial satellite communication signal into the frequency domain through Fourier transform, taking the spikes at specific frequencies in the frequency domain as the initial pure signal, and taking the energy distribution on the spectrum in the frequency domain as the initial noise signal; performing time series analysis on the initial pure signal and the initial noise signal to identify the target pure signal and the target noise signal in the initial pure signal and the initial noise signal, where the target pure signal is the initial pure signal and the initial noise signal with periodicity and regularity, and the target noise signal is the initial pure signal and the initial noise signal with random or irregular fluctuations.

[0008] As a preferred solution, the step of classifying the target noise signal into several subgroup signals and performing different noise reduction processes on different subgroup signals includes: grouping according to the frequency range and intensity of the target noise signal to obtain several subgroup signals, and assigning a preset noise reduction process strategy to each subgroup signal for noise reduction processing; where the noise reduction process strategy includes any one of the following: digital filtering, adaptive noise reduction algorithm, signal processing technology.

[0009] As a preferred solution, the step of recombining the subgroup signals after different noise reduction processes to obtain a noise reduction signal includes: assigning timestamps to each subgroup signal in the subgroup signals after different noise reduction processes; sorting the subgroup signals according to the timestamps, performing synchronization adjustment on the sorted subgroup signals, and performing phase analysis on the synchronously adjusted subgroup signals to obtain the phase information of each subgroup signal at each time point; performing phase adjustment on each subgroup signal according to the phase information to align the phases of all subgroup signals at the same time point; obtaining the amplitude information of each phase-aligned subgroup signal at each time point, and performing amplitude adjustment on each phase-aligned subgroup signal according to the amplitude information to make the amplitudes of all phase-aligned subgroup signals consistent, obtaining several adjusted subgroup signals; synthesizing all the adjusted subgroup signals to obtain a noise reduction signal.

[0010] As a preferred solution, the step of synthesizing the target pure signal and the noise reduction signal to obtain the final communication signal includes: calculating the weight coefficients of the noise reduction signal and the target pure signal using the signal-to-noise ratio of the signal; performing spectral adjustment on the weight coefficients of the noise reduction signal according to the spectral characteristics of the noise reduction signal to obtain a plurality of weighted noise reduction spectral components; wherein, the spectral characteristics of the noise reduction signal are the energy levels of each frequency component and the required noise reduction degree; performing time series weighting on the weight coefficients of the target pure signal based on the priority of the target pure signal in communication to obtain a plurality of weighted pure signal components; wherein, the priority of the target pure signal in communication is the importance of the target pure signal to the information content and the contribution to the overall communication quality; using the time-frequency synchronization characteristic to perform weighting on the plurality of weighted noise reduction spectral components and the plurality of weighted pure signal components through the weighted superposition method to obtain the final communication signal.

[0011] As a preferred solution, the step of converting the final communication signal into voice data and verifying the integrity of the voice data includes: decoding the final communication signal to separate data streams carrying different information; wherein, the data streams include voice data; performing formatting processing on the voice data to convert the voice data into a standard audio format; substituting the voice data into a preset audio model, and judging the integrity of the voice data by the audio model according to the waveform, spectrum and acoustic characteristics of the audio of the voice data.

[0012] As a preferred solution, the step of, if there is missing data in the voice data, tracing the missing data, obtaining the subgroup signal corresponding to the missing data, and adjusting the noise reduction processing corresponding to the subgroup signal until the voice data is complete includes: when the audio model determines that there are abnormalities in the audio waveform or audio spectrum or audio acoustic characteristics in the voice data, it is determined that there is missing data in the voice data; obtaining the noise reduction processing record, searching for the subgroup signal associated with the missing data according to the noise reduction processing record, and adjusting the noise reduction processing process of the subgroup signal to obtain the adjusted subgroup signal; recombining the adjusted subgroup signal with the normal subgroup signal to generate an adjusted noise reduction signal, synthesizing the adjusted noise reduction signal and the target pure signal to obtain an adjusted final communication signal; converting the adjusted final communication signal into adjusted voice data; inputting the adjusted voice data into the audio model to verify the integrity of the adjusted voice data; if there is still missing data in the adjusted voice data, the above steps will be repeated until all the voice data is complete.

[0013] The present application also provides an intelligent voice noise reduction device based on satellite communication, including: a receiving module, configured to receive an initial satellite communication signal and classify the initial satellite communication signal into a target pure signal and a target noise signal; a classification module, configured to classify the target noise signal into several subgroup signals and perform different noise reduction processes on different subgroup signals; a recombination module, configured to recombine the subgroup signals after different noise reduction processes to obtain a noise reduction signal, and synthesize the target pure signal and the noise reduction signal to obtain a final communication signal; a verification module, configured to convert the final communication signal into voice data and verify the integrity of the voice data; a traceability module, configured to, if there is missing data in the voice data, trace the missing data, obtain the subgroup signal corresponding to the missing data, and adjust the noise reduction process corresponding to the subgroup signal until the voice data is complete.

[0014] The present application also provides an electronic device, including a memory and a processor. The memory stores a computer program that can run on the processor, and when the processor executes the computer program, it implements the intelligent voice noise reduction method based on satellite communication described in any one of the above.

[0015] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, the processor is caused to execute the intelligent voice noise reduction method based on satellite communication described in any one of the above.

[0016] Compared with the prior art, the present application has the following beneficial effects: good noise reduction effect and high communication integrity. By performing different noise reduction processes on the target noise signal, noise can be more accurately reduced, thereby improving the quality of the final communication signal. Recombining the subgroup signals after different noise reduction processes helps to further optimize the noise reduction effect. By verifying the integrity of the voice data and tracing and adjusting the missing data, the accuracy and integrity of the final voice data are ensured, the reliability and effectiveness of satellite communication are improved, and the problem that it is difficult to effectively distinguish noise and signals in a complex or changing noise environment, resulting in poor integrity of satellite signals, is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] The structures, proportions, sizes, etc. shown in the drawings of this specification are only used to cooperate with the content disclosed in the specification for those skilled in this technology to understand and read, and are not used to limit the limited conditions for the implementation of the present invention. Therefore, they do not have substantial technical significance. Any modification of the structure, change in the proportional relationship, or adjustment of the size, without affecting the efficacy that the present invention can produce and the purpose that can be achieved, should still fall within the scope that can be covered by the technical content disclosed in the present invention.

[0019] Figure 1 It is a schematic flowchart of the intelligent voice noise reduction method based on satellite communication provided by an embodiment of the present invention;

[0020] Figure 2 It is a schematic block diagram of the structure of the intelligent voice noise reduction device based on satellite communication provided by an embodiment of the present invention;

[0021] Figure 3 It is a schematic block diagram of the structure of the electronic device provided by an embodiment of the present invention.

[0022] Explanation of reference numerals:

[0023] 10. Intelligent voice noise reduction device based on satellite communication; 11. Receiving module; 12. Classification module; 13. Recombination module; 14. Verification module; 15. Tracing module; 20. Electronic device; 21. Memory; 22. Processor. Detailed implementation manners

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0025] The flowchart shown in the drawings is only an example illustration, and does not necessarily include all the content and operations / steps, nor does it necessarily need to be executed in the described order. For example, some operations / steps can also be decomposed, combined, or partially merged. Therefore, the actual execution order may change according to the actual situation.

[0026] It should also be understood that the terms used in this specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in this specification of the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms of "a", "an", and "the" are intended to include the plural forms.

[0027] It should be further understood that the term "and / or" used in the specification and appended claims of this application refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0028] The technical solution of the present invention will be further described below with reference to the accompanying drawings and through specific embodiments.

[0029] Embodiment 1:

[0030] As Figure 1 shown, the intelligent voice noise reduction method based on satellite communication provided by the embodiment of this application includes steps S100 to S500.

[0031] Step S100: Receive the initial satellite communication signal and classify the initial satellite communication signal into a target pure signal and a target noise signal.

[0032] In this step, the initial communication signal transmitted by the satellite is captured by a receiver; specifically, a high-speed signal processor is used to perform spectral analysis and time-domain analysis on the received communication signal to distinguish the target pure signal and the target noise signal. During the classification process, parameters such as the frequency characteristics, energy distribution, and time-domain characteristics of the signal are used as the classification basis.

[0033] For example, in satellite communication, the pure signal generally has clear frequency characteristics and a high signal-to-noise ratio, while the noise signal is distributed in a wide frequency band and has low energy. By analyzing these characteristics, the system can effectively separate the initial satellite communication signal into a target pure signal and a target noise signal.

[0034] Step S200: Classify the target noise signal into several sub-group signals and perform different noise reduction processes on different sub-group signals.

[0035] In this step, feature extraction and clustering analysis are performed on the target noise signal to divide it into several sub-group signals with similar characteristics; specifically, a clustering algorithm (such as K-means or DBSCAN) is used to classify the noise signal according to spectral characteristics and time-domain characteristics, and then the most suitable noise reduction algorithm is selected for each sub-group signal for processing.

[0036] For example, some noise signals have strong frequency characteristics and are suitable for noise reduction using frequency-domain filtering methods, while other noise signals are pulse interferences and are suitable for processing using time-domain smoothing filtering or median filtering. In this way, the accuracy and effectiveness of the noise reduction effect can be improved.

[0037] Step S300: Recombine the sub-group signals after different noise reduction processes to obtain a noise-reduced signal, and synthesize the target pure signal and the noise-reduced signal to obtain the final communication signal.

[0038] In this step, the subgroup signals processed by different noise reduction algorithms are integrated and reconstructed into a complete noise-reduced signal. Specifically, using weighted averaging or signal fusion techniques, the individual subgroup signals are weighted and synthesized according to their importance and noise reduction effect to obtain the overall noise-reduced signal, which is then synthesized with the target clean signal to preserve the integrity of the clean signal and enhance the quality of the overall signal.

[0039] For example, for a signal containing multiple noise components, after removing low-frequency noise through frequency-domain filtering and eliminating high-frequency impulse noise through time-domain filtering, the processed signal parts are weighted and synthesized, and finally merged with the target clean signal to obtain a high-quality final communication signal.

[0040] Step S400: Convert the final communication signal into voice data and verify the integrity of the voice data.

[0041] In this step, the final communication signal is demodulated and the voice signal is extracted and converted into voice data. Specifically, signal demodulation techniques are used to extract the voice data from the communication signal, and the integrity of the extracted voice data is checked through speech recognition algorithms, including the continuity of data frames, the correctness of content, and the clarity of speech.

[0042] For example, during the demodulation process, if it is found that there is missing or incorrect voice data, the system will mark these data frames and correct them in subsequent steps to ensure the integrity and coherence of the voice data.

[0043] Step S500: If there is missing data in the voice data, trace the source of the missing data, obtain the subgroup signal corresponding to the missing data, and adjust the noise reduction processing corresponding to the subgroup signal until the voice data is complete.

[0044] In this step, a traceability analysis is performed on the missing part in the voice data to determine the subgroup signal corresponding to the missing data. Specifically, by comparing the timestamps and frequency characteristics of the missing data, the subgroup of the original target noise signal is traced, and the noise reduction parameters and processing methods of these subgroup signals are readjusted until the missing data is effectively restored.

[0045] For example, if it is detected that the voice data in certain time periods in the reconstructed signal is incomplete, the system will trace back to the subgroup signals corresponding to these time periods, adjust the bandwidth of the frequency-domain filter or the window size of the time-domain filter, and process these subgroup signals again until the complete voice data is restored.

[0046] In the above embodiments, the satellite initial communication signal is received and classified into a target pure signal and a target noise signal; then, the target noise signal is further classified into several subgroup signals, and different noise reduction processes are performed on different subgroup signals; subsequently, the subgroup signals after different noise reduction processes are recombined to obtain a noise reduction signal; then, the target pure signal and the noise reduction signal are synthesized to obtain a final communication signal; finally, the final communication signal is converted into voice data, and the integrity of the voice data is verified. If there is missing data in the voice data, the missing data is traced, the subgroup signal corresponding to the missing data is obtained, and the noise reduction process corresponding to the subgroup signal is adjusted until the voice data is complete.

[0047] By classifying the target noise signal into several subgroup signals and performing different noise reduction processes, the noise can be reduced more precisely, thereby improving the quality of the final communication signal. Recombining the subgroup signals after different noise reduction processes helps to further optimize the noise reduction effect; in addition, by verifying the integrity of the voice data and tracing and adjusting the missing data, the accuracy and integrity of the final voice data are ensured, the reliability and effectiveness of satellite communication are improved, and the problem that it is difficult to effectively distinguish noise and signals in a complex or changing noise environment, resulting in poor integrity of satellite signals, is solved.

[0048] Embodiment 2:

[0049] In step S100, the satellite initial communication signal is converted into the frequency domain through Fourier transform. The spikes at specific frequencies in the frequency domain are used as the initial pure signal, and the energy distribution on the spectrum in the frequency domain is used as the initial noise signal.

[0050] The received satellite communication signal is converted from the time domain to the frequency domain through Fourier transform to more easily identify the characteristics of the signal and noise; specifically, the fast Fourier transform (FFT) algorithm is used to perform spectrum analysis on the signal, and the spikes at specific frequencies in the frequency domain are extracted. These spikes usually represent pure signals because they have high energy concentration and clear frequency characteristics, while the extensive energy distribution in the frequency domain usually represents noise signals because the energy of noise signals is usually distributed over a wider frequency band.

[0051] For example, in actual operation, it is found through FFT analysis that there are obvious spikes at a certain specific frequency, which means that the signal component corresponding to this frequency has a high signal-to-noise ratio and can be regarded as the initial pure signal. In other frequency bands in the frequency domain, the energy distribution is relatively uniform and the energy is low, and these components are regarded as the initial noise signals. In this way, the initial pure signal and the initial noise signal can be effectively distinguished.

[0052] Perform time series analysis on the initial pure signal and the initial noise signal to identify the target pure signal and the target noise signal in the initial pure signal and the initial noise signal. Among them, the target pure signal is the initial pure signal and the initial noise signal with periodicity and regularity, and the target noise signal is the initial pure signal and the initial noise signal with randomness or irregular fluctuations.

[0053] Further process the initial pure signal and the initial noise signal through time series analysis to identify the target pure signal and the target noise signal; specifically, use methods such as autocorrelation analysis and power spectral density analysis to identify the signal components with periodicity and regularity as the target pure signal, while the components with randomness or irregular fluctuations are identified as the target noise signal.

[0054] For example, during the autocorrelation analysis process, the target pure signal will show obvious periodic peaks, while the target noise signal will not have such a pattern. By comparing the graphs of the autocorrelation functions, the target pure signal and the target noise signal can be accurately separated. Such a processing method can effectively improve the accuracy of signal classification.

[0055] In step S200, group the target noise signals according to their frequency ranges and intensities to obtain several subgroup signals, and assign a preset noise reduction processing strategy to each subgroup signal for noise reduction processing; among them, the noise reduction processing strategy includes any one of the following: digital filtering, adaptive noise reduction algorithm, signal processing technology.

[0056] By analyzing the frequency ranges and intensities of the target noise signals, classify and group them so as to adopt different noise reduction strategies for different types of noise signals; specifically, use the spectral analysis method to divide the noise signals according to the frequency ranges, and at the same time further group them according to the intensities of the noise signals, and then assign appropriate noise reduction processing strategies to each subgroup signal. For example, low-frequency noise in the frequency domain can be processed using a low-pass filter, high-frequency noise can be processed using a high-pass filter, and adaptive noise can be processed using an adaptive noise reduction algorithm.

[0057] For example, if the frequency range of a certain subgroup signal mainly concentrates in the low-frequency band and has a large intensity, a low-pass filter can be selected to perform noise reduction processing on it; if the frequency range of another subgroup signal is wide and the intensity is uneven, an adaptive noise reduction algorithm can be selected to dynamically adjust the noise reduction parameters according to the noise characteristics, thereby improving the accuracy and effectiveness of the noise reduction effect.

[0058] In step S300, assign time stamps to each subgroup signal in the subgroup signals after different noise reduction processes.

[0059] By assigning timestamps to each subgroup signal, precise time synchronization of these signals can be achieved in subsequent processing steps. Specifically, signal processing techniques are used to add timestamp information to each subgroup signal after noise reduction processing to mark its time position in the original signal. This ensures that during the signal recombination process, the signals can be arranged and synthesized in the correct chronological order.

[0060] For example, when processing a complex signal containing multiple time periods, by adding timestamps to each subgroup signal, it can be ensured that these signals are arranged in the correct chronological order during subsequent recombination, thereby avoiding signal distortion or information loss caused by time misalignment.

[0061] Sort the subgroup signals according to the timestamps, perform synchronization adjustment on the arranged subgroup signals, and perform phase analysis on the synchronously adjusted subgroup signals to obtain the phase information of each subgroup signal at each time point.

[0062] Sort the subgroup signals by timestamps to ensure that the signals are processed in the correct chronological order. Specifically, use the timestamp information to arrange each subgroup signal in chronological order, then perform synchronization adjustment to ensure that all subgroup signals are aligned on the same time basis, and obtain the phase information of each subgroup signal at each time point through phase analysis.

[0063] For example, when processing multiple subgroup signals, if there is no correct time sorting and synchronization adjustment, it will lead to time misalignment of the signals, thereby affecting the synthesis quality of the signals. Through timestamp sorting and synchronization adjustment, this problem can be avoided, and the time consistency of the signals can be further ensured through phase analysis.

[0064] Perform phase adjustment on each subgroup signal according to the phase information to align the phases of all subgroup signals at the same time point.

[0065] Ensure that the phases of all subgroup signals are aligned at the same time point through phase adjustment, thereby improving the overall quality of the synthesized signal. Specifically, use the phase information obtained through phase analysis to perform corresponding phase adjustments on each subgroup signal, so that the phases of all subgroup signals are consistent at the same time point, and avoid signal interference or distortion caused by phase misalignment.

[0066] For example, during the synthesis process, if the phase of a certain subgroup signal is inconsistent with that of other signals, it will cause phase interference in the synthesized signal. Through phase adjustment, it can be ensured that the phases of all signals are consistent, thereby improving the quality of the final synthesized signal.

[0067] Obtain the amplitude information of each subgroup signal after phase alignment at each time point. According to the amplitude information, perform amplitude adjustment on each subgroup signal after phase alignment to make the amplitudes of all subgroup signals after phase alignment consistent, and obtain several adjusted subgroup signals.

[0068] By obtaining and adjusting the amplitude information of each subgroup signal, it is ensured that all subgroup signals after phase alignment are also consistent in amplitude, thereby further improving the quality of signal synthesis. Specifically, use the amplitude analysis method to obtain the amplitude information of each subgroup signal after phase alignment at each time point, and then perform amplitude adjustment on the signal according to this information to make it consistent, ensuring the stability of the overall amplitude of the signal after synthesis.

[0069] For example, if the amplitudes of some subgroup signals are too high or too low, it will cause uneven quality of the synthesized signal. Through amplitude adjustment, the amplitudes of all subgroup signals can be made consistent, thereby improving the stability and quality of the final synthesized signal.

[0070] Synthesize all the adjusted subgroup signals to obtain a noise-reduced signal.

[0071] By synthesizing all the adjusted subgroup signals, a complete noise-reduced signal is generated. Specifically, use signal processing techniques to synthesize the subgroup signals that have been adjusted in phase and amplitude, ensuring the best consistency of each subgroup signal in time, phase, and amplitude, thereby generating a high-quality noise-reduced signal.

[0072] For example, during the signal synthesis process, weighted average or signal fusion techniques can be used to synthesize all the adjusted subgroup signals, ensuring that the contribution of each subgroup signal is fully considered, thereby generating a high-quality noise-reduced signal and improving the overall quality of the communication signal.

[0073] Through the above steps, the noise signal can be effectively reduced to the lowest level and the integrity of the pure signal can be retained. In practical applications, first convert the received satellite communication signal into a frequency-domain signal through Fourier transform, extract the spikes at specific frequencies in the frequency domain as the initial pure signal, and use the energy distribution on the spectrum as the initial noise signal. Then, further identify the target pure signal and the target noise signal through time series analysis, group the target noise signal according to the frequency range and intensity, and allocate appropriate noise reduction processing strategies for each subgroup signal. The subgroup signals after different noise reduction processes are sorted by timestamp, synchronously adjusted, and phase analyzed to ensure the consistency of the signal in time, phase, and amplitude. Finally, all the adjusted subgroup signals are synthesized to obtain a high-quality noise-reduced signal, thereby ensuring the quality and integrity of the final communication signal.

[0074] In step S300, it also includes calculating the weight coefficients of the noise-reduced signal and the target pure signal using the signal-to-noise ratio of the signal.

[0075] By calculating the signal-to-noise ratio of the signal, the weight coefficients of the noise-reduced signal and the target pure signal are determined to reasonably allocate their respective contributions during signal synthesis; specifically, using the signal-to-noise ratio analysis method, the signal-to-noise ratios of the noise-reduced signal and the target pure signal are calculated, and their respective weight coefficients are determined according to the levels of the signal-to-noise ratio, so as to reasonably adjust their proportions during the synthesis process to ensure the best quality of the synthesized signal.

[0076] For example, in the actual calculation process, if a certain section of the noise-reduced signal has a relatively high signal-to-noise ratio, a higher weight coefficient can be assigned to it, while a lower weight coefficient is assigned to the target pure signal with a relatively low signal-to-noise ratio. In this way, the advantages of the high signal-to-noise ratio signal can be fully utilized during the synthesis process to improve the overall signal quality.

[0077] According to the spectral characteristics of the noise-reduced signal, spectral adjustment is performed on the weight coefficient of the noise-reduced signal to obtain a number of weighted noise-reduced spectral components; among them, the spectral characteristics of the noise-reduced signal are the energy levels of each frequency component and the required noise reduction degree.

[0078] By analyzing and adjusting the spectral characteristics of the noise-reduced signal, the weight coefficient of the noise-reduced signal is further optimized to make its distribution in the spectrum more reasonable; specifically, using spectral analysis technology, the energy levels of each frequency component in the noise-reduced signal and the required noise reduction degree are evaluated, and then the weight coefficient of the noise-reduced signal is adjusted according to these characteristics to make the effects of each frequency component after noise reduction optimal.

[0079] For example, when processing a signal containing multiple frequency components, spectral adjustment can ensure that the noise reduction effect of each frequency component reaches the best, thereby improving the quality of the entire noise-reduced signal. This method can effectively optimize the processing according to the noise characteristics of different frequency bands.

[0080] Based on the priority of the target pure signal in communication, time-series weighting is performed on the weight coefficient of the target pure signal to obtain a number of weighted pure signal components; among them, the priority of the target pure signal in communication is the importance of the target pure signal to the information content and the contribution to the overall communication quality.

[0081] By performing time-series weighting on the weight coefficient of the target pure signal according to its priority in communication, it is ensured that important pure signals receive more attention during synthesis; specifically, using time-series analysis technology, the importance of the target pure signal at different time points is evaluated, and the weight coefficient is weighted and adjusted according to its priority in communication, so that important pure signals occupy a larger proportion during the synthesis process.

[0082] For example, during the communication process, the pure signals in certain time periods contain key information. These signals will be given higher weight coefficients when weighted, so as to occupy a larger proportion in the final synthesized signal, ensuring the transmission quality and integrity of the key information.

[0083] Using the weighted superposition method, several weighted noise reduction spectrum components and several weighted pure signal components are weighted by using the time-frequency synchronization characteristic to obtain the final communication signal.

[0084] Using the weighted superposition method to perform time-frequency synchronization weighting on multiple weighted noise reduction spectrum components and weighted pure signal components to generate the final communication signal; specifically, using time-frequency analysis technology to ensure that each signal component is synchronously processed and weighted and superposed in both the time and frequency dimensions, so as to generate a high-quality final communication signal.

[0085] For example, in actual operation, through time-frequency synchronization weighting, it can be ensured that the contributions of each signal component are reasonably allocated at different frequency bands and time points, so as to generate a high-quality final communication signal. This method can effectively improve the overall quality and stability of the communication signal.

[0086] Through the above series of steps, signal noise reduction and optimization can be effectively achieved. First, use the signal-to-noise ratio analysis method to calculate the weight coefficients of the noise reduction signal and the target pure signal, and then adjust the weight coefficients spectrally according to the spectral characteristics of the noise reduction signal to obtain the weighted noise reduction spectrum components. Then, based on the priority of the target pure signal in communication, perform time-series weighting on its weight coefficients to obtain the weighted pure signal components. Finally, using the weighted superposition method, use the time-frequency synchronization characteristic to perform weighted superposition on the weighted noise reduction spectrum components and pure signal components to generate the final communication signal. This process ensures the synchronization and optimization of the signal in both the time and frequency dimensions, improves the quality and integrity of the communication signal, especially in a complex noise environment, and can effectively improve the transmission effect and reliability of the signal.

[0087] In step S400, the final communication signal is decoded to separate the data streams carrying different information; among them, the data stream includes voice data.

[0088] By decoding the final communication signal, the data streams carrying different information are separated for further processing; specifically, using the decoding algorithm to separate the different data streams in the final communication signal, especially extracting the voice data to prepare for the subsequent processing steps.

[0089] For example, during the decoding process, decoding techniques such as convolutional codes and LDPC codes can be used to separate the voice data and other types of data streams mixed in the communication signal. In this way, it can be ensured that the voice data is correctly extracted, laying a foundation for subsequent formatting processing and integrity verification.

[0090] Perform formatting processing on the voice data to convert the voice data into a standard audio format.

[0091] By performing formatting processing on the extracted voice data, it is ensured that it conforms to the standard audio format, facilitating subsequent processing and analysis; specifically, audio encoding techniques are used to convert the voice data into common audio formats such as WAV and MP3 to ensure data compatibility and ease of processing.

[0092] For example, during the formatting processing, the PCM encoding technique can be used to convert the voice data into the WAV format, which can not only guarantee the sound quality but also ensure the compatibility of the data on different devices and software, thus facilitating subsequent analysis and processing.

[0093] Substitute the voice data into a preset audio model, and the audio model judges the integrity of the voice data based on the waveform, spectrum, and acoustic features of the audio of the voice data.

[0094] By inputting the formatted voice data into a preset audio model, its integrity is verified; specifically, the audio model analyzes the waveform, spectrum, and acoustic features of the voice data to judge whether the data is complete and detects whether there are any losses or errors.

[0095] For example, the audio model can adopt a speech recognition model trained by deep learning algorithms. By analyzing the waveform and spectrum features of the voice data, the abnormal parts are identified. If it is found that there are obvious missing or distorted frames, the model will mark these frames and correct them in subsequent steps to ensure the integrity and coherence of the voice data.

[0096] Through the above steps, the voice data in the final communication signal can be effectively processed and verified. First, different data streams in the final communication signal are separated by a decoding algorithm to extract the voice data. Then, the extracted voice data is formatted and converted into a standard audio format such as WAV or MP3. Next, the formatted voice data is input into a preset audio model, and its integrity is judged by analyzing its waveform, spectrum, and acoustic features. If the audio model detects missing or incorrect parts in the voice data, these abnormal parts will be marked and corrected in subsequent steps. In this way, the integrity and coherence of the voice data can be ensured, improving the quality and reliability of the communication signal.

[0097] In step S500, when it is determined by the audio model that there are abnormalities in the audio waveform, audio spectrum, or audio acoustic features in the voice data, it is determined that there is missing data in the voice data.

[0098] The audio model analyzes the voice data. If abnormalities are found in the audio waveform, spectrum, or acoustic features, it is determined that there is missing or incorrect voice data. Specifically, the audio model performs a detailed feature analysis on the voice data. Once abnormal fluctuations or inconsistencies are detected, the model immediately marks these abnormal parts and initiates the process of repairing missing data.

[0099] For example, if the audio model finds that the spectrum features of a certain segment of voice data are significantly inconsistent with normal voice features during the analysis process, or there are obvious breaks or noise interferences in the waveform diagram, it can be determined that there is missing or incorrect data in this segment. The model will generate a corresponding error report and initiate subsequent repair steps.

[0100] Obtain the noise reduction processing record, search for the subgroup signal associated with the missing data according to the noise reduction processing record, and adjust the noise reduction processing process of the subgroup signal to obtain the adjusted subgroup signal.

[0101] By referring to the noise reduction processing record, find the subgroup signal associated with the missing data and adjust its noise reduction processing process to recover the missing data. Specifically, use the detailed information in the noise reduction processing record to trace back the noise reduction processing process, find the steps and parameters related to the missing data, and then make corresponding adjustments to generate the adjusted subgroup signal.

[0102] For example, in the noise reduction processing record, if it is found that the noise reduction parameter settings of a certain subgroup signal in a specific frequency band are inappropriate, resulting in data loss in this frequency band, the original information of this frequency band can be restored by adjusting the noise reduction parameters of this frequency band and reprocessing the signal, thereby obtaining the adjusted subgroup signal.

[0103] Recombine the adjusted subgroup signal with the normal subgroup signal to generate the adjusted noise reduction signal, and synthesize the adjusted noise reduction signal with the target pure signal to obtain the adjusted final communication signal.

[0104] By recombining the adjusted subgroup signal with the normal subgroup signal, a new noise reduction signal is generated and synthesized with the target pure signal to obtain the adjusted final communication signal. Specifically, using signal processing techniques, the adjusted subgroup signal is merged with the normal subgroup signal to ensure their consistency in time, phase, and amplitude, and then weighted synthesis is performed with the target pure signal to generate a high-quality final communication signal.

[0105] For example, in actual operation, by adjusting the subgroup signals related to the missing data, the original information can be restored. Then, these adjusted signals are merged with other normal signals to ensure that the quality of the synthesized noise-reduced signal is not affected. Subsequently, they are synthesized with the target clean signal to generate a complete and high-quality final communication signal.

[0106] Convert the adjusted final communication signal into adjusted speech data.

[0107] By processing the adjusted final communication signal, corresponding speech data is generated. Specifically, using decoding and conversion technologies, the adjusted final communication signal is decoded and converted into speech data to ensure the integrity and correctness of the data.

[0108] For example, during the decoding process, advanced decoding algorithms can be employed to convert the adjusted communication signal into speech data, and through detection and verification steps, the quality and coherence of the speech data are ensured.

[0109] Input the adjusted speech data into the audio model to verify the integrity of the adjusted speech data.

[0110] By inputting the adjusted speech data into the audio model, its integrity is verified. Specifically, the audio model is used to conduct a detailed analysis of the adjusted speech data to determine whether its waveform, spectrum, and acoustic features are complete, so as to ensure the accuracy and coherence of the data.

[0111] For example, during the verification process, if the audio model detects that there are still anomalies or missing parts in the speech data, an error report will be generated and further correction steps will be initiated to ensure that the finally generated speech data is completely complete.

[0112] If there is still missing data in the adjusted speech data, the above steps will be repeated until all the speech data is complete.

[0113] By repeating the above steps, the integrity of the speech data is ensured. Specifically, if there are still missing or incorrect parts in the adjusted speech data, steps such as noise reduction processing, signal recombination, and synthesis will be carried out again until all the speech data is complete.

[0114] For example, during multiple adjustment and verification processes, ensure that each segment of speech data undergoes strict detection and correction until finally generating high-quality speech data without any missing parts. This process ensures the reliability and accuracy of the communication signal.

[0115] For example, in practical applications, when analyzing and detecting speech data through an audio model, if missing data is found, the backtracking of the noise reduction processing record will be initiated, relevant subgroup signals will be found and their noise reduction parameters will be adjusted to restore the missing data. Then, the adjusted subgroup signals will be recombined with the normal subgroup signals to generate a new noise reduction signal, which will be synthesized with the target clean signal to obtain the final communication signal. Next, the final communication signal will be converted into speech data and input into the audio model for integrity verification. If the model detects that there is still missing data in the speech data, the above steps will be repeated until all the data is complete. Through this iterative processing method, the integrity and coherence of the speech data are ensured, and the quality and reliability of the communication signal are improved.

[0116] In this embodiment, the initial communication signal is converted into a frequency-domain signal through Fourier transform, the spikes at specific frequencies in the frequency domain are extracted as the initial clean signal, and the energy distribution on the spectrum is used as the initial noise signal. Then, the target clean signal and the target noise signal are identified through time series analysis, grouped according to the frequency range and intensity of the noise signal, and different noise reduction processing strategies are adopted. Next, time stamps are assigned to each subgroup signal for synchronization adjustment and phase analysis to ensure the consistency of the signals in terms of time, phase, and amplitude. Subsequently, the adjusted subgroup signals are synthesized to generate a high-quality noise reduction signal, and the weight coefficients of the signal are optimized through signal-to-noise ratio calculation and spectrum adjustment. Finally, through decoding, formatting processing, and audio model verification, the integrity and coherence of the speech data are ensured. The entire process ensures the high quality and reliability of the communication signal through repeated iterative processing, especially in a complex noise environment, and can effectively improve the transmission effect and accuracy of the signal.

[0117] Embodiment 3:

[0118] As Figure 2 shown, the present application also provides an intelligent speech noise reduction device 10 based on satellite communication. The intelligent speech noise reduction device 10 based on satellite communication includes a receiving module 11, a classification module 12, a recombination module 13, a verification module 14, and a traceability module 15.

[0119] The receiving module 11 is mainly used to receive the satellite initial communication signal and classify the satellite initial communication signal into a target clean signal and a target noise signal.

[0120] The classification module 12 is mainly used to classify the target noise signal into several subgroup signals and perform different noise reduction processing on different subgroup signals.

[0121] The recombination module 13 is mainly used to recombine the subgroup signals after different noise reduction processing to obtain a noise reduction signal, and synthesize the target clean signal and the noise reduction signal to obtain the final communication signal.

[0122] The verification module 14 is mainly used to convert the final communication signal into voice data and verify the integrity of the voice data.

[0123] The traceability module 15 is mainly used to trace the missing data if there is any in the voice data, obtain the subgroup signals corresponding to the missing data, and adjust the noise reduction processing corresponding to the subgroup signals until the voice data is complete.

[0124] In this embodiment, the intelligent voice noise reduction function is realized through modular design. Specifically, the receiving module 11 uses a highly sensitive receiver and a spectrum analyzer to accurately receive and classify the initial communication signal; the classification module 12 adopts an advanced signal processing algorithm to carefully group the target noise signals and uses a variety of noise reduction techniques to optimize the processing of different subgroup signals; the recombination module 13 precisely recombines the processed subgroup signals through phase synchronization and amplitude matching algorithms to ensure the quality of the synthesized communication signal; the verification module 14 efficiently verifies the final voice data by combining voice recognition technology and integrity detection algorithms; the traceability module 15 traces back and adjusts the missing data by recording and analyzing each step in the processing process to ensure the integrity and reliability of the voice data. Through the collaborative work of these modules, the noise reduction performance and data processing ability of the device are significantly improved, meeting the requirements of high-quality satellite communication.

[0125] It should be noted that those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described device and each module can refer to the corresponding processes in the embodiment of the intelligent voice noise reduction method based on satellite communication described above, and will not be elaborated here.

[0126] Embodiment 4:

[0127] As Figure 3 shown, the present application also provides an electronic device 20, including a memory 21 and a processor 22. The memory 21 stores a computer program that can run on the processor 22, and when the processor 22 executes the computer program, it implements the intelligent voice noise reduction method based on satellite communication in Embodiment 1.

[0128] In this embodiment, by integrating the intelligent voice noise reduction method into the electronic device 20, efficient noise reduction processing is achieved. Specifically, optimized algorithms and parameter configurations are pre-stored in the memory 21, and the processor 22 calls these programs and configurations during operation to quickly execute steps such as Fourier transform, signal classification, noise reduction processing, signal recombination, and integrity verification; the device can also adaptively adjust the algorithm parameters according to the actual communication environment to cope with different noises and signal changes; in addition, the electronic device 20 significantly improves the processing speed and efficiency through efficient hardware acceleration and parallel processing technologies. In practical applications, this integrated solution can be widely applied to devices such as satellite phones and satellite data terminals to provide stable and reliable voice communication services.

[0129] Embodiment 5:

[0130] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is run by a processor, the processor is caused to execute the intelligent voice noise reduction method based on satellite communication as in Embodiment 1.

[0131] In this embodiment, by storing the intelligent voice noise reduction program on a computer-readable storage medium, a flexible noise reduction solution deployment is achieved. Specifically, the computer program includes detailed algorithm steps and optimized parameters and can run on different processor platforms to adapt to various hardware environments; the storage medium can be a hard disk, a solid-state drive, a USB flash drive, etc., which is convenient for program transmission and installation; during program operation, by calling the preset noise reduction algorithm, the received communication signal is accurately processed to ensure the quality and integrity of voice data. In this way, users can conveniently deploy the intelligent voice noise reduction method on different devices and widely apply it to fields such as satellite communication and wireless communication to improve communication quality and user experience.

[0132] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent voice noise reduction method based on satellite communication, characterized in that, Including: Receiving an initial satellite communication signal and classifying the initial satellite communication signal into a target pure signal and a target noise signal; wherein, the target pure signal is an initial pure signal and an initial noise signal with periodicity and regularity, and the target noise signal is an initial pure signal and an initial noise signal with randomness or irregular fluctuations. Classifying the target noise signal into several subgroup signals and performing different noise reduction processes on different subgroup signals. Recombining the subgroup signals after different noise reduction processes to obtain a noise reduction signal, and synthesizing the target pure signal and the noise reduction signal to obtain a final communication signal. Converting the final communication signal into voice data and verifying the integrity of the voice data. If there is missing data in the voice data, tracing the missing data, obtaining the subgroup signal corresponding to the missing data, and adjusting the noise reduction process corresponding to the subgroup signal until the voice data is complete.

2. The intelligent voice noise reduction method based on satellite communication according to claim 1, wherein The step of receiving an initial satellite communication signal and classifying the initial satellite communication signal into a target pure signal and a target noise signal includes: Converting the initial satellite communication signal into the frequency domain through Fourier transform, taking the spikes at specific frequencies in the frequency domain as the initial pure signal, and taking the energy distribution on the spectrum in the frequency domain as the initial noise signal; wherein, the specific frequency is a frequency feature with high energy concentration. Performing time series analysis on the initial pure signal and the initial noise signal to identify the target pure signal and the target noise signal in the initial pure signal and the initial noise signal, wherein, the target pure signal is an initial pure signal and an initial noise signal with periodicity and regularity, and the target noise signal is an initial pure signal and an initial noise signal with randomness or irregular fluctuations.

3. The intelligent voice noise reduction method based on satellite communication according to claim 1, wherein The step of classifying the target noise signal into several subgroup signals and performing different noise reduction processes on different subgroup signals includes: Grouping according to the frequency range and intensity of the target noise signal to obtain several subgroup signals, and assigning a preset noise reduction process strategy to each subgroup signal for noise reduction; wherein, the noise reduction process strategy includes any one of the following: digital filtering, adaptive noise reduction algorithm.

4. The intelligent voice noise reduction method based on satellite communication according to claim 1, characterized in that, The step of recombining the subgroup signals after different noise reduction processes to obtain a noise reduction signal includes: Assigning timestamps to each subgroup signal in the subgroup signals after different noise reduction processes. Sorting the subgroup signals according to the timestamps, performing synchronization adjustment on the sorted subgroup signals, and performing phase analysis on the synchronously adjusted subgroup signals to obtain the phase information of each subgroup signal at each time point. ​ ​ Combine all the adjusted subgroup signals to obtain a noise-reduced signal.

5. The intelligent voice noise reduction method based on satellite communication according to claim 1, wherein The step of combining the target clean signal and the noise-reduced signal to obtain a final communication signal includes: Calculate the weight coefficients of the noise-reduced signal and the weight coefficients of the target clean signal using the signal-to-noise ratio of the signal; Perform spectral adjustment on the weight coefficients of the noise-reduced signal according to the spectral characteristics of the noise-reduced signal to obtain a number of weighted noise-reduced spectral components; wherein, the spectral characteristics of the noise-reduced signal are the energy levels of each frequency component and the required noise reduction degree; Perform time-series weighting on the weight coefficients of the target clean signal based on the priority of the target clean signal in communication to obtain a number of weighted clean signal components; wherein, the priority of the target clean signal in communication is the importance of the target clean signal to the information content and the contribution to the overall communication quality; Use the time-frequency synchronization characteristic to perform weighting on the number of weighted noise-reduced spectral components and the number of weighted clean signal components through the weighted superposition method to obtain a final communication signal.

6. The intelligent voice noise reduction method based on satellite communication according to claim 1, characterized in that The step of converting the final communication signal into voice data and verifying the integrity of the voice data includes: Decode the final communication signal to separate data streams carrying different information; wherein, the data streams include voice data; Perform formatting processing on the voice data to convert the voice data into a standard audio format; Substitute the voice data into a preset audio model, and the audio model determines the integrity of the voice data according to the waveform, spectrum, and acoustic characteristics of the audio of the voice data.

7. The intelligent voice noise reduction method based on satellite communication according to claim 6, characterized in that The step of, if there is missing data in the voice data, tracing the source of the missing data, obtaining the subgroup signal corresponding to the missing data, and adjusting the noise reduction process corresponding to the subgroup signal until the voice data is complete includes: When the audio model determines that there are abnormalities in the audio waveform or audio spectrum or audio acoustic characteristics in the voice data, it is determined that there is missing data in the voice data; Obtain the noise reduction processing record, search for the subgroup signal associated with the missing data according to the noise reduction processing record, and adjust the noise reduction process of the subgroup signal to obtain the adjusted subgroup signal; Recombine the adjusted subgroup signal with the normal subgroup signal to generate an adjusted noise-reduced signal, combine the adjusted noise-reduced signal with the target clean signal to obtain an adjusted final communication signal; Convert the adjusted final communication signal into adjusted voice data; Input the adjusted voice data into the audio model to verify the integrity of the adjusted voice data; If there is still missing data in the adjusted voice data, the above steps will be repeated until all the voice data is complete.

8. An intelligent voice noise reduction device based on satellite communication, characterized in that, Including: A receiving module, configured to receive an initial satellite communication signal and classify the initial satellite communication signal into a target pure signal and a target noise signal; wherein, the target pure signal is an initial pure signal and an initial noise signal with periodicity and regularity, and the target noise signal is an initial pure signal and an initial noise signal with randomness or irregular fluctuations; A classification module, configured to classify the target noise signal into several subgroup signals and perform different noise reduction processes on different subgroup signals; A recombination module, configured to recombine the subgroup signals after different noise reduction processes to obtain a noise reduction signal, and synthesize the target pure signal and the noise reduction signal to obtain a final communication signal; A verification module, configured to convert the final communication signal into voice data and verify the integrity of the voice data; A traceability module, configured to, if there is missing data in the voice data, trace the missing data, obtain the subgroup signal corresponding to the missing data, and adjust the noise reduction process corresponding to the subgroup signal until the voice data is complete.

9. An electronic device, characterized in that, It includes a memory and a processor, the memory stores a computer program running on the processor, and when the processor executes the computer program, it implements the intelligent voice noise reduction method based on satellite communication according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon, and when the computer program is run by the processor, the processor is caused to execute the intelligent voice noise reduction method based on satellite communication according to any one of claims 1 to 7.

Images