FPGA-based dual-tone multi-frequency signal multipath parallel detection method and system
By using an FPGA-based method for parallel detection of dual-tone multi-frequency (DTMF) signals and employing a six-fold judgment criterion, the method solves the problems of misjudgment and missed judgment in noisy environments for DTMF signal detection. It achieves high recognition accuracy and robustness under low signal-to-noise ratio conditions and is suitable for parallel detection implemented in FPGA hardware.
Patent Information
- Application Number
- CN202511593620.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-03
- Publication Date
- 2026-01-23
AI Technical Summary
In noisy environments, traditional DTMF signal detection methods are easily affected by external interference, amplitude imbalance, and harmonic components, leading to misjudgments and missed judgments. In particular, in multi-channel parallel detection implemented in FPGA hardware, the robustness is insufficient, affecting communication reliability.
A parallel detection method for dual-tone multi-frequency signals based on FPGA is adopted. Through μ-law decoding and data buffering, the frequency component energy value is calculated by combining FFT or Goertzel algorithm, and a six-fold judgment standard is executed, including single signal energy intensity judgment, result reliability judgment, interference frequency judgment, effective frequency judgment, overall energy intensity judgment of sampling window and second harmonic energy judgment, and outputting a valid DTMF key signal.
It significantly reduces the false positive and false negative rates of dual-tone multi-frequency signals in noisy environments, effectively resists environmental noise and out-of-band interference, maintains high recognition accuracy and robustness, suppresses harmonic interference, and ensures the real-time performance and reliability of signal detection.
Smart Images
Figure CN121397148A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The application belongs to the technical field of communication and digital signal processing, and particularly relates to a dual-tone multi-frequency signal multi-path parallel detection method and system based on FPGA. BACKGROUND
[0002] DTMF (Dual-Tone Multi-Frequency) is a kind of telephone signaling system, which can represent numbers or symbols by combining two specific frequency audio signals, and realize telephone communication, remote control and interactive voice response function.
[0003] However, in the prior art, in a noisy environment, the traditional DTMF signal detection method is easily affected by external interference, amplitude imbalance and harmonic components, resulting in misjudgment and missed judgment, especially in the multi-path parallel detection of FPGA hardware implementation, the robustness of DTMF signal detection is insufficient, which seriously affects the communication reliability of the system. SUMMARY
[0004] The application provides a dual-tone multi-frequency signal multi-path parallel detection method and system based on FPGA, to solve the technical problem that the conventional DTMF signal detection method cannot accurately recognize DTMF signals under noise interference conditions in the prior art.
[0005] To solve the above problems, the technical scheme of the application is as follows: a dual-tone multi-frequency signal multi-path parallel detection method based on FPGA, comprising the following steps: S1: based on FPGA, parallel and real-time sampling of data stream of dual-tone multi-frequency signal, performing mu-law decoding and data buffering processing on the dual-tone multi-frequency signal; S2: for any group of data stream, based on FPGA, calculating the energy values of different frequency components in the row frequency signal and the column frequency signal by FFT or Goertzel algorithm respectively, and determining the target frequency of the row frequency signal and the column frequency signal according to the highest energy value; S3: based on FPGA, performing multiple decision processing on the row frequency signal and the column frequency signal, and outputting an effective DTMF key signal when the decision results of the row frequency signal and the column frequency signal meet the decision conditions.
[0006] Preferably, the FPGA performs decision processing on the row frequency signal and the column frequency signal, and the decision conditions include single signal energy strength decision, result credibility decision, interference frequency decision, effective frequency decision, sampling window overall energy strength decision and second harmonic energy decision. When the decision results of the row frequency signal and the column frequency signal meet all the decision conditions, an effective DTMF key signal is outputted.
[0007] Preferably, in S3, the single signal energy strength determination of the row frequency signal and the column frequency signal based on FPGA includes the following steps: S31: compare the energy values of the target frequency of the row frequency signal and the column frequency signal with the energy strength threshold value respectively, and when the energy values of the target frequency of the row frequency signal and the column frequency signal are both greater than the energy strength threshold value, it is confirmed that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0008] Preferably, a noise estimation model is preset in the FPGA, the noise estimation model is used to input the energy value distribution vector E = [P(f1),..., P(fM)] of different frequency components in the current detection window row frequency signal and column frequency signal and the statistical characteristics of the background noise, and output a dynamic energy strength threshold value.
[0009] Preferably, in S3, the result credibility determination of the row frequency signal and the column frequency signal based on FPGA includes the following steps: S32: calculate the ratio of the energy values of the target frequency of the row frequency signal and the column frequency signal, and when the ratio of the energy values of the target frequency of the row frequency signal and the column frequency signal is greater than the minimum result credibility threshold value and less than the maximum result credibility threshold value, it is confirmed that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0010] Preferably, in S3, the interference frequency determination of the row frequency signal and the column frequency signal based on FPGA includes the following steps: S33: select the larger value of the energy values of the target frequency of the row frequency signal and the column frequency signal as the reference energy value, and calculate the multiplication result of the reference energy value and the interference threshold value; Compare the energy values of the non-target frequency in the row frequency signal and the column frequency signal with the multiplication result of the reference energy value and the interference threshold value respectively, and when the energy values of all non-target frequencies in the row frequency signal and the column frequency signal are all less than the multiplication result of the reference energy value and the interference threshold value, it is confirmed that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0011] Preferably, in S3, the effective frequency determination of the row frequency signal and the column frequency signal based on FPGA includes the following steps: S34: calculate the addition result of the energy values of the target frequency of the row frequency signal and the column frequency signal, and when the addition result of the energy values of the target frequency of the row frequency signal and the column frequency signal is greater than the effective frequency threshold value, it is confirmed that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0012] Preferably, it further includes the following steps: S341: calculate the energy of the background noise and the energy of the dual-tone multi-frequency signal respectively, and the dynamic updating calculation method of the effective frequency threshold value is: T = En + β (Es - En) Wherein, T is an effective frequency threshold, En is background noise energy, β is a proportional coefficient, and Es is DTMF signal energy. Alternatively, the effective frequency threshold is adjusted based on a DTMF signal signal-to-noise ratio condition.
[0013] Preferably, in S3, the FPGA is used to perform sampling window overall energy strength determination on the row frequency signal and the column frequency signal, including the following steps: S35: Calculate the sum of squares of audio intensities of a preset number of sampling points in the sampling window, and when the sum of squares of audio intensities of the preset number of sampling points in the sampling window is greater than a minimum overall energy strength threshold of the sampling window and less than a maximum overall energy strength threshold of the sampling window, it is determined that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0014] Preferably, in S3, the FPGA is used to perform second harmonic determination on the row frequency signal and the column frequency signal, including the following steps: S36: Use the method of segmented FFT and peak tracking to obtain the target frequency point with the highest energy in each FFT result, and calculate the second harmonic energy of the target frequency point, and the calculation method of the second harmonic energy of the target frequency point is: SH[m] = |X[k 2f0 ]| 2 When the second harmonic energy of the target frequency point is less than the second harmonic threshold, it is determined that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0015] Based on the same concept, the application also provides a FPGA-based DTMF signal multi-path parallel detection system for performing the FPGA-based DTMF signal multi-path parallel detection method as described in any one of the above embodiments, comprising: A sampling module for parallel and real-time sampling of a data stream of a DTMF signal by an FPGA, performing μ-law decoding and data buffering processing on the DTMF signal; A row frequency signal and column frequency signal target frequency calculation module for calculating the energy values of different frequency components in the row frequency signal and the column frequency signal by an FPGA and using FFT or Goertzel algorithm, and determining the target frequencies of the row frequency signal and the column frequency signal according to the highest energy value; A row frequency signal and column frequency signal target frequency determination module for performing multiple determination processing on the row frequency signal and the column frequency signal by an FPGA, and outputting an effective DTMF key signal when the determination results of the row frequency signal and the column frequency signal meet the determination condition.
[0016] The application has the following advantages and positive effects compared with the prior art by adopting the above technical solutions: The application provides a dual-tone multi-frequency signal multi-path parallel detection method and system based on FPGA, which realizes multi-path parallel detection of the dual-tone multi-frequency signal by sampling the data stream of the dual-tone multi-frequency signal in parallel and real time based on FPGA, and greatly reduces the misjudgment rate and the missed judgment rate of the dual-tone multi-frequency signal in a noise environment through six judgment criteria (including single signal energy intensity judgment, result reliability judgment, interference frequency judgment, effective frequency judgment, sampling window overall energy intensity judgment and second harmonic energy judgment), effectively resists the influence of environmental noise and out-of-band interference signals, still maintains a high recognition accuracy under a low signal-to-noise ratio condition, effectively suppresses harmonic interference, and avoids false recognition caused by fundamental harmonic by combining the second harmonic energy judgment mechanism, so that the dual-tone multi-frequency signal detection maintains good real-time performance and robustness. BRIEF DESCRIPTION OF DRAWINGS
[0017] Figure 1 The application provides a first flowchart of a dual-tone multi-frequency signal multi-path parallel detection method based on FPGA. Figure 2 The application provides a second flowchart of a dual-tone multi-frequency signal multi-path parallel detection method based on FPGA. Figure 3 The application provides an audio schematic diagram of a basically noise-free scene. Figure 4 The application provides an audio schematic diagram of a weak noise scene. Figure 5 The application provides an audio schematic diagram of a strong noise scene. Figure 6 The application provides a key tone audio judgment schematic diagram. Figure 7 The application provides a block diagram of a dual-tone multi-frequency signal multi-path parallel detection system based on FPGA. DETAILED DESCRIPTION
[0018] The application provides a dual-tone multi-frequency signal multi-path parallel detection method and system based on FPGA, which will be further described in detail below in combination with the drawings and specific embodiments. The advantages and features of the application will be clearer according to the following description and claims.
[0019] First embodiment Referring to Figures 1-2 The embodiment provides a dual-tone multi-frequency signal multi-path parallel detection method based on FPGA, which is used for realizing multi-path parallel detection of the dual-tone multi-frequency signal and efficiently recognizing the dual-tone multi-frequency signal, and specifically includes the following steps: S1: Based on FPGA (Field Programmable Gate Array), a plurality of 8 kHz dual-tone multi-frequency signals are sampled in parallel and real time to obtain data streams, and μ-law decoding and data buffering are performed on the dual-tone multi-frequency signals. The μ-law is a nonlinear audio compression encoding method, which can effectively represent the dynamic range of a voice signal under a limited number of bits (usually 8 bits), and provide higher resolution for small signals and appropriate compression for large signals. In this embodiment, the transmitting end of the dual-tone multi-frequency signal can compress the 14-bit or 16-bit linear audio signal into an 8-bit nonlinear code through μ-law encoding, so the receiving end of the dual-tone multi-frequency signal needs to perform corresponding μ-law decoding, and then the decoded linear dual-tone multi-frequency signal data is temporarily stored in the storage resource inside the FPGA, thereby completing the signal sampling and preprocessing process of the dual-tone multi-frequency signal.
[0020] S2: Based on FPGA, the energy values of different frequency components in the row frequency signal and the column frequency signal are calculated for any group of data streams by FFT or Goertzel algorithm, and the target frequency of the row frequency signal and the column frequency signal is determined according to the highest energy value. The dual-tone multi-frequency signal can represent a number or a symbol by combining two audio signals of specific frequencies to form a unique tone signal, that is, in the dual-tone multi-frequency signal, a 4x4 matrix design is adopted for the DTMF keyboard, each row represents a low frequency, and each column represents a high frequency, so the row frequency signal and the column frequency signal are formed, and 16 different frequency pairs are formed by the composition of any frequency signal in the row frequency signal and the column frequency signal, each frequency pair corresponds to a specific key information. Specifically, the signal frequencies in the row frequency signal include four groups of 697 Hz, 770 Hz, 852 Hz and 941 Hz, and the signal frequencies in the column frequency signal include four groups of 1209 Hz, 1336 Hz, 1477 Hz and 1633 Hz.
[0021] Therefore, in an embodiment, the energy distribution of the entire dual-tone multi-frequency signal spectrum can be calculated by the FFT (Fast Fourier Transform) method, the energy values of the 8 DTMF frequency points in the FFT result are checked, and the target frequency of the row frequency signal and the column frequency signal is determined according to the highest energy value. In another embodiment, the energy values of the 8 DTMF frequency points can also be directly calculated by the Goertzel algorithm method, and the target frequency of the row frequency signal and the column frequency signal is determined according to the highest energy value.
[0022] S3: Based on FPGA, the row frequency signal and the column frequency signal are executed to determine the processing, and when the determination result of the row frequency signal and the column frequency signal meets the determination condition, the valid DTMF key signal is output.
[0023] In the embodiment, the row frequency signal and the column frequency signal are subjected to the determination processing based on the FPGA, the determination conditions include the single signal energy strength determination, the result reliability determination, the interference frequency determination, the effective frequency determination, the sampling window overall energy strength determination and the second harmonic energy determination, and when the determination results of the row frequency signal and the column frequency signal satisfy all the determination conditions, the effective DTMF key signal is output.
[0024] In the following, the specific implementation steps and functions of the multi-path parallel detection method for the dual-tone multi-frequency signal based on the FPGA provided in the embodiment will be described in further detail. Preferably, in an embodiment, the single signal energy strength determination of the row frequency signal and the column frequency signal based on the FPGA in S3 includes the following steps: S31: The energy values of the target frequencies of the row frequency signal and the column frequency signal are compared with the energy strength threshold value respectively, and when the energy values of the target frequencies of the row frequency signal and the column frequency signal are both greater than the energy strength threshold value, it is confirmed that the target frequencies of the row frequency signal and the column frequency signal satisfy the determination condition.
[0025] That is, the maximum energy values Pr of the four groups of frequencies of 1209 Hz, 1336 Hz, 1477 Hz and 1633 Hz in the row frequency signal are calculated, and the maximum energy values Pc of the four groups of frequencies of 697 Hz, 770 Hz, 852 Hz and 941 Hz in the column frequency signal are calculated, and if the target frequency energy value Pr of the row frequency signal and the target frequency energy value Pc of the column frequency signal are both greater than the energy strength threshold value Pt, it is considered that the target frequencies of the row frequency signal and the column frequency signal satisfy the determination condition, otherwise, it is considered that the target frequencies of the row frequency signal and the column frequency signal are invalid.
[0026] The energy strength threshold value Pt is the model training result.
[0027] Specifically, a noise estimation model (such as MLP, LSTM) is preset in the FPGA, the noise estimation model uses the detection data in the noisy environment and the quiet environment as the training set to perform the model training step, and the noise estimation model after the training is used to input the energy value distribution vector E = [P(f1),..., P(fM)] of different frequency components in the current detection window row frequency signal and column frequency signal and the statistical characteristics (mean and variance) of the background noise, and output the dynamic energy strength threshold value Pt. That is, in the embodiment, by introducing the adaptive noise threshold estimation strategy, the false positive rate and the missed detection rate of the dual-tone multi-frequency signal are reduced.
[0028] Preferably, in an embodiment, the result reliability determination of the row frequency signal and the column frequency signal based on the FPGA in S3 includes the following steps: S32: Calculate the ratio of the target frequency energy value Pr of the row frequency signal and the target frequency energy value Pc of the column frequency signal. When the ratio of the target frequency energy values of the row frequency signal and the column frequency signal is greater than a minimum result reliability threshold value (for example, 0.158) and less than a maximum result reliability threshold value (for example, 2.5), it is confirmed that the target frequencies of the row frequency signal and the column frequency signal satisfy the determination condition, otherwise, it is considered that the target frequencies of the row frequency signal and the column frequency signal are invalid.
[0029] Preferably, in an embodiment, the interference frequency determination of the row frequency signal and the column frequency signal is performed based on the FPGA in S3, including the following steps: S33: Select the larger value of the target frequency energy value Pr of the row frequency signal and the target frequency energy value Pc of the column frequency signal as a reference energy value Pm, and calculate the multiplication result of the reference energy value Pm and an interference threshold value (for example, 0.158); Compare the energy values of the non-target frequencies in the row frequency signal and the column frequency signal with the multiplication result of the reference energy value and the interference threshold value respectively. When the energy values of all the non-target frequencies in the row frequency signal and the column frequency signal are less than the multiplication result of the reference energy value and the interference threshold value, it is confirmed that the target frequencies of the row frequency signal and the column frequency signal satisfy the determination condition, otherwise, it is considered that the target frequencies of the row frequency signal and the column frequency signal are invalid.
[0030] Preferably, in an embodiment, the valid frequency determination of the row frequency signal and the column frequency signal is performed based on the FPGA in S3, including the following steps: S34: Calculate the addition result of the target frequency energy value Pr of the row frequency signal and the target frequency energy value Pc of the column frequency signal. When the addition result of the target frequency energy values of the row frequency signal and the column frequency signal is greater than a valid frequency threshold value, it is confirmed that the target frequencies of the row frequency signal and the column frequency signal satisfy the determination condition, otherwise, it is considered that the target frequencies of the row frequency signal and the column frequency signal are invalid.
[0031] In an embodiment, the following steps are further included: S341: Calculate the energy of the background noise and the energy of the dual-tone multi-frequency signal respectively. The dynamic updating calculation method of the valid frequency threshold value is: T = En + β (Es - En) Wherein, T is the valid frequency threshold value, En is the background noise energy, β is the proportional coefficient, and Es is the dual-tone multi-frequency signal energy.
[0032] Or, in another embodiment, the valid frequency threshold value can also be adjusted based on the signal-to-noise ratio condition of the dual-tone multi-frequency signal.
[0033] In this embodiment, the valid frequency threshold value can be continuously updated with the change of the valid key tone energy, so as to enhance the detection effect of the dual-tone multi-frequency signal.
[0034] Preferably, in one embodiment, the sampling window overall energy strength determination on the row frequency signal and the column frequency signal based on FPGA in S3 includes the following steps: S35: Calculate the square sum result of the audio intensity of the preset number of sampling points in the sampling window. When the square sum result of the audio intensity of the preset number of sampling points in the sampling window is greater than the minimum overall energy intensity threshold of the sampling window and less than the maximum overall energy intensity threshold of the sampling window, it is confirmed that the target frequency of the row frequency signal and the column frequency signal meets the determination condition.
[0035] Referring to Figure 6 In the present embodiment, the duration of the DTMF tone can exceed the size of the sampling window. For example, the audio actual duration of the DTMF signal is 45 ms, and a fixed length sampling window of 25 ms is used for detecting the DTMF. When the duration of the DTMF signal exceeds the size of the sampling window, a complete DTMF signal will be divided into two consecutive sampling windows, so a small length of silence will be set after the key is pressed. The position of the silence is detected to distinguish between repeated key tones and long key tones. That is, if the silence is detected, it means that the current key has ended, which is a repeated key. If the silence is not detected, it means a long key. Therefore, in the present embodiment, the square sum of all sampling points in the sampling window is calculated as the audio intensity. When the intensity exceeds the maximum overall energy intensity threshold of the sampling window, it is determined that it is an overload invalidity, which is a long key. When the intensity is lower than the minimum overall energy intensity threshold of the sampling window, it is determined that it is a silence invalidity, that is, the sampling window only samples the silence and no valid signal.
[0036] Preferably, in one embodiment, the second harmonic determination on the row frequency signal and the column frequency signal based on FPGA in S3 includes the following steps: S36: Use the method of segmented FFT and peak tracking. For example, the continuous audio stream is divided into several small segments (such as a 25 ms sampling window, corresponding to 200 sampling points), and FFT processing is performed on each segment of audio. The target frequency point (i.e. f0) with the highest energy is obtained in each FFT result, and the second harmonic energy of the target frequency point is calculated. The calculation method of the second harmonic energy of the target frequency point is as follows: SH[m] = |X[k 2f0 ]| 2 Wherein, SH[m] represents the second harmonic energy, 2f0 is the second harmonic of the target frequency point f0 (for example, f0=697 Hz, then 2f0=1394 Hz), |X[k2f0]| 2 is the amplitude square of the frequency point corresponding to 2f0 in the FFT result, that is, the energy value.
[0037] In the DTMF signal, the second harmonic energy is very low (should be close to 0), and in the voice / music signal, the second harmonic energy is strong, so in this embodiment, when the second harmonic energy of the target frequency point is less than the second harmonic threshold, it is confirmed that the target frequencies of the row frequency signal and the column frequency signal meet the determination condition, otherwise, it is considered that the target frequency detection of the row frequency signal and the column frequency signal is invalid.
[0038] Referring to Figures 3-5 , the embodiment also demonstrates several audio scenes in the actual test environment, including: a basic noise-free scene, a weak noise scene, a strong noise scene, and a key tone detection sample number of 5000. The six-determination criteria are arranged in the order of single signal energy strength determination, result reliability determination, interference frequency determination, effective frequency determination, sampling window overall energy strength determination, and second harmonic energy determination. Among them, in the basic noise-free scene, the first of the six filtering strategies can achieve a detection accuracy of 100%; in the weak noise scene, the first three of the six filtering strategies can achieve a detection accuracy of 100%; in the strong noise scene, the first, second, third, fourth, and sixth detection strategies can achieve a detection accuracy of 84.3%; in the strong noise scene, the first, second, third, fourth, and fifth detection strategies can achieve a detection accuracy of 90.4%; and in the strong noise scene, all six detection strategies can achieve a detection accuracy of 100%.
[0039] Second embodiment Referring to Figure 7 , based on the same concept, the application also provides a dual-tone multi-frequency signal multi-path parallel detection system based on FPGA, which is used to execute the dual-tone multi-frequency signal multi-path parallel detection method based on FPGA as described in any one of the first embodiment, comprising: a sampling module for sampling the data stream of the dual-tone multi-frequency signal in parallel and real time through the FPGA, and performing mu-law decoding and data caching processing on the dual-tone multi-frequency signal; a row frequency signal and column frequency signal target frequency calculation module for calculating the energy values of different frequency components in the row frequency signal and the column frequency signal through the FPGA and using the FFT or Goertzel algorithm, and determining the target frequencies of the row frequency signal and the column frequency signal according to the highest energy value; a row frequency signal and column frequency signal target frequency determination module for performing multiple determination processing on the row frequency signal and the column frequency signal through the FPGA, and outputting an effective DTMF key signal when the determination results of the row frequency signal and the column frequency signal meet the determination condition.
[0040] The function implementation of each component unit in the above-mentioned FPGA-based dual-tone multi-frequency signal multi-path parallel detection system corresponds to each step in the above-mentioned FPGA-based dual-tone multi-frequency signal multi-path parallel detection method embodiment, and the function and implementation process will not be repeated here.
[0041] The embodiment also provides an electronic device, which comprises a processor and a memory. The memory stores machine executable instructions capable of being executed by the processor. The processor executes the machine executable instructions to implement the above-mentioned FPGA-based dual-tone multi-frequency signal multi-path parallel detection method.
[0042] The embodiment also provides a computer readable storage medium, which stores computer executable instructions. When the computer executable instructions are called and executed by a processor, the computer executable instructions cause the processor to implement the above-mentioned FPGA-based dual-tone multi-frequency signal multi-path parallel detection method.
[0043] The embodiments of the present application are described in detail above with reference to the drawings, but the present application is not limited to the above-mentioned embodiments. Even if various changes are made to the present application, as long as the changes belong to the scope of the claims of the present application and the equivalent technologies, they still fall within the protection scope of the present application.
Claims
1. A method for detecting multiple DTMF signals in parallel based on FPGA, characterized in that, Comprising the following steps: S1: Based on FPGA, parallel and real-time sampling of the data stream of the dual-tone multi-frequency signal, performing μ-law decoding and data caching processing on the dual-tone multi-frequency signal; S2: For any group of data streams, based on FPGA, calculating the energy values of different frequency components in the horizontal frequency signal and the vertical frequency signal respectively through FFT or Goertzel algorithm, and determining the target frequency of the horizontal frequency signal and the vertical frequency signal according to the highest energy value; S3: Based on FPGA, performing multiple decision processing on the horizontal frequency signal and the vertical frequency signal, and outputting the valid DTMF key signal when the decision results of the horizontal frequency signal and the vertical frequency signal meet the decision conditions.
2. The method according to claim 1, wherein the method is characterized by, Based on FPGA, performing decision processing on the horizontal frequency signal and the vertical frequency signal, and the decision conditions include single signal energy strength decision, result reliability decision, interference frequency decision, valid frequency decision, sampling window overall energy strength decision and second harmonic energy decision; When the decision results of the horizontal frequency signal and the vertical frequency signal meet all the decision conditions, the valid DTMF key signal is outputted.
3. The method according to claim 2, wherein the method comprises: In S3, based on FPGA, performing single signal energy strength decision on the horizontal frequency signal and the vertical frequency signal, comprising the following steps: S31: Comparing the energy values of the target frequencies of the horizontal frequency signal and the vertical frequency signal with the energy strength threshold value respectively, and confirming that the target frequencies of the horizontal frequency signal and the vertical frequency signal meet the decision conditions when the energy values of the target frequencies of the horizontal frequency signal and the vertical frequency signal are both greater than the energy strength threshold value.
4. The method according to claim 3, wherein the method is characterized by, In the FPGA, a noise estimation model is preset, which is used to input the energy value distribution vector E = [P(f1),..., P(fM)] of different frequency components in the current detection window horizontal frequency signal and vertical frequency signal and the statistical characteristics of the background noise, and output a dynamic energy strength threshold value.
5. The method according to claim 3, wherein the method is characterized by, In S3, based on FPGA, performing result reliability decision on the horizontal frequency signal and the vertical frequency signal, comprising the following steps: S32: Calculating the ratio of the energy values of the target frequencies of the horizontal frequency signal and the vertical frequency signal, and confirming that the target frequencies of the horizontal frequency signal and the vertical frequency signal meet the decision conditions when the ratio of the energy values of the target frequencies of the horizontal frequency signal and the vertical frequency signal is greater than the minimum result reliability threshold value and less than the maximum result reliability threshold value.
6. The method according to claim 5, wherein the method is characterized by, In S3, based on FPGA, performing interference frequency decision on the horizontal frequency signal and the vertical frequency signal, comprising the following steps: S33: Selecting the larger value of the energy values of the target frequencies of the horizontal frequency signal and the vertical frequency signal as the reference energy value, and calculating the multiplication result of the reference energy value and the interference threshold value; Comparing the energy values of the non-target frequencies in the horizontal frequency signal and the vertical frequency signal with the multiplication result of the reference energy value and the interference threshold value respectively, and confirming that the target frequencies of the horizontal frequency signal and the vertical frequency signal meet the decision conditions when the energy values of all the non-target frequencies in the horizontal frequency signal and the vertical frequency signal are all less than the multiplication result of the reference energy value and the interference threshold value.
7. The method according to claim 6, wherein the method is characterized by, In S3, based on FPGA, performing valid frequency decision on the horizontal frequency signal and the vertical frequency signal, comprising the following steps: S34: Calculate the addition result of the energy values of the target frequencies of the row frequency signal and the column frequency signal, and when the addition result of the energy values of the target frequencies of the row frequency signal and the column frequency signal is greater than the effective frequency threshold value, it is determined that the target frequencies of the row frequency signal and the column frequency signal meet the determination condition.
8. The method according to claim 7, wherein the method is characterized by, Further comprising the following steps: S341: Calculate the energy of the background noise and the energy of the dual-tone multi-frequency signal respectively, and the dynamic updating calculation method of the effective frequency threshold value is: T = En + β (Es - En) Wherein, T is the effective frequency threshold value, En is the background noise energy, β is the proportional coefficient, and Es is the dual-tone multi-frequency signal energy. Or, adjust the effective frequency threshold value based on the signal-to-noise ratio condition of the dual-tone multi-frequency signal.
9. The method according to claim 7, wherein the method is characterized by, In S3, the FPGA is used to perform overall energy intensity determination on the row frequency signal and the column frequency signal, including the following steps: S35: Calculate the sum of squares of the audio intensities of a preset number of sampling points in the sampling window, and when the sum of squares of the audio intensities of the preset number of sampling points in the sampling window is greater than the minimum overall energy intensity threshold value of the sampling window and less than the maximum overall energy intensity threshold value of the sampling window, it is determined that the target frequencies of the row frequency signal and the column frequency signal meet the determination condition.
10. The method according to claim 9, wherein the method is characterized by, In S3, the FPGA is used to perform second harmonic determination on the row frequency signal and the column frequency signal, including the following steps: S36: Use the method of segmented FFT and peak tracking to obtain the target frequency point with the highest energy in each FFT result, and calculate the second harmonic energy of the target frequency point, and the calculation method of the second harmonic energy of the target frequency point is: SH[m] = |X[k 2f0 ]|2 2 When the second harmonic energy of the target frequency point is less than the second harmonic threshold value, it is determined that the target frequencies of the row frequency signal and the column frequency signal meet the determination condition.
11. A dual tone multi-frequency signal multi-path parallel detection system based on FPGA, characterized in that, The FPGA-based dual-tone multi-frequency signal multi-path parallel detection method comprises: A sampling module for sampling the data stream of the dual-tone multi-frequency signal in parallel and real time through the FPGA, performing μ-law decoding and data buffering processing on the dual-tone multi-frequency signal; A row frequency signal and column frequency signal target frequency calculation module for calculating the energy values of different frequency components in the row frequency signal and the column frequency signal through the FPGA and using FFT or Goertzel algorithm, and determining the target frequencies of the row frequency signal and the column frequency signal according to the highest energy value; A row frequency signal and column frequency signal target frequency determination module for performing multiple determination processing on the row frequency signal and the column frequency signal through the FPGA, and outputting an effective DTMF key signal when the determination results of the row frequency signal and the column frequency signal meet the determination condition.