CN117648601A - Unmanned underwater sound target recognition system based on FPGA - Google Patents
Unmanned underwater sound target recognition system based on FPGA Download PDFInfo
- Publication number
- CN117648601A CN117648601A CN202311505539.8A CN202311505539A CN117648601A CN 117648601 A CN117648601 A CN 117648601A CN 202311505539 A CN202311505539 A CN 202311505539A CN 117648601 A CN117648601 A CN 117648601A
- Authority
- CN
- China
- Prior art keywords
- fpga
- underwater sound
- target recognition
- data
- data acquisition
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 14
- 230000005236 sound signal Effects 0.000 claims abstract description 12
- 238000007781 pre-processing Methods 0.000 claims abstract description 10
- 238000004891 communication Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 abstract description 5
- 238000004364 calculation method Methods 0.000 abstract description 4
- 230000006870 function Effects 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 abstract description 2
- 238000005516 engineering process Methods 0.000 description 11
- 238000000034 method Methods 0.000 description 8
- 238000001228 spectrum Methods 0.000 description 8
- 238000013461 design Methods 0.000 description 7
- 230000005855 radiation Effects 0.000 description 6
- 238000001514 detection method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000000605 extraction Methods 0.000 description 3
- 230000008447 perception Effects 0.000 description 3
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 3
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 238000010521 absorption reaction Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000009432 framing Methods 0.000 description 2
- 230000017525 heat dissipation Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003672 processing method Methods 0.000 description 2
- 230000035939 shock Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 230000010365 information processing Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- NGVDGCNFYWLIFO-UHFFFAOYSA-N pyridoxal 5'-phosphate Chemical compound CC1=NC=C(COP(O)(O)=O)C(C=O)=C1O NGVDGCNFYWLIFO-UHFFFAOYSA-N 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000013179 statistical model Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Landscapes
- Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
Abstract
The invention discloses an unmanned underwater sound target recognition system based on an FPGA (field programmable gate array), which is used for realizing the underwater sound target recognition reasoning function of Zynq UltraScale+MPSoCs series embedded platform based on Sitting Xilinx based on an EDA tool chain, and designing a data acquisition card, the FPGA, a CPU (Central processing Unit) board card and other accessories on a main control board; the data acquisition card sends acquired underwater sound data to an ARM core PS in the FPGA through the Ethernet, the PS transmits the underwater sound data to a logic part PL of the FPGA through an AXI bus, the PL performs underwater sound signal preprocessing by using an IP core of a Mel frequency cepstrum coefficient MFCC, and the processed data is sent to a convolutional neural network CNN to realize underwater sound target identification reasoning. The invention realizes the functions of underwater sound signal acquisition, data processing, target identification and the like based on the FPGA, improves the calculation efficiency, reduces the equipment volume, reduces the equipment power consumption, and is more beneficial to the practical application of the system.
Description
Technical Field
The invention relates to the technical field of underwater sound target recognition, in particular to an unmanned underwater sound target recognition system based on an FPGA.
Background
With the advent of sonar technology, underwater sound target recognition technology has also emerged as an important follow-up link for sonar data processing. The underwater sound target mark is an information processing technology for extracting characteristic information by using the radiation noise of the target and performing target discrimination and classification. The underwater sound destination identifier is a technology for performing multi-means and cross-domain analysis processing on time domain data collected by different sonar collecting devices, eliminating distortion caused by environments and channels, extracting characteristics capable of representing the essence of a target to be detected and classifying and identifying the characteristics.
The underwater target recognition technology is an essential key technology for the trend of the marine equipment to be intelligent, has wide application in the fields of marine resource exploration, marine environment information exploration, underwater information reconnaissance and the like, and has a large number of scholars to throw in the technology from the 50 th century. As the strategic position of the ocean is increasingly highlighted, the development and utilization of ocean resources and space is an important subject. The underwater target recognition is one of the most important functional requirements in the passive sonar system, however, as the underwater noise environment changes along with different sea areas, even if the same sea area changes along with time, the water temperature changes and the underwater depth changes, different sound field environments are formed, and the accuracy of underwater target recognition is difficult to obtain a stable effect. The acoustic wave is the only energy form capable of being transmitted in the ocean at present, so that the underwater sound target identification has great significance for ocean development and national defense safety, and is one of research hotspots in the underwater sound field. Aiming at the requirements of underwater detection in modern war, the underwater target classification and identification based on sound signals also becomes a research hotspot in the field of underwater detection. The detection and recognition of the underwater sound target plays a key role in underwater combat and underwater sound target perception, and along with informatization and intellectualization of naval equipment, the underwater sound target recognition is a prerequisite for the underwater combat on the water, so that whether the underwater sound target can be timely and accurately recognized and analyzed is an important factor for mastering the initiative of the combat in the ocean war. Because the purity of the real-time data information of the ocean audio is not high, the accuracy of the model for predicting the data is not high enough when training is carried out by applying some conventional algorithms, and the sample data can not be accurately identified well. The passive sonar has the advantages of strong concealment, long transmission distance and the like, and underwater target radiation noise is a main information source for underwater sound target identification through passive sonar acquisition. The passive sonar system has good concealment and flexibility, so how to utilize the ship radiation noise obtained by the passive sonar to detect and identify the underwater target becomes a problem to be solved urgently. However, due to the application of various ship stealth technologies and the complex and diverse marine environments, the task of underwater target identification based on ship radiation noise faces a great challenge.
Currently, aiming at different underwater information acquisition forms, the main research directions comprise two types of image recognition and signal recognition. The underwater acoustic signal recognition method taking the target radiation noise as an analysis object has the characteristics of large effective range, wide application condition and the like, and is widely applied to the technical field of underwater target recognition. The underwater sound target identification is a technology for classifying targets by utilizing target radiation noise signals, and a classification recognition method based on a traditional statistical model mainly comprises three steps of preprocessing, feature extraction and selection and a classifier, wherein common features include a power spectrum, an auditory spectrum, a noise envelope signal identification (Detection ofEnvelope Modulation OnNoise, abbreviated as DEMON) spectrum, a low-frequency analysis record (LowFrequencyAnalysis Recording, abbreviated as LOFAR) spectrum, wavelet features, loudness features, mel-Frequency Cepstral Coefficients, MFCC (Meyer's cepstrum coefficient) features, perception linear prediction (Perceptual LinearPredictive, abbreviated as PLP) features and the like, and whether reliable features can be extracted can directly influence the recognition rate of the underwater sound targets. The traditional underwater sound signal identification method takes a target feature extraction and pattern identification method as a core. The target feature extraction method with higher distinction degree is designed according to the target features, and the target recognition is realized by combining with SVM, neural network and other efficient classification methods.
In recent years, with the good development of machine learning, deep learning and other technologies, the technology of identifying underwater sound targets has also achieved some new progress and research results. The detection and recognition of the underwater sound target plays a key role in underwater combat and underwater sound target perception, and along with informatization and intellectualization of naval equipment, the underwater sound target recognition is a prerequisite for the underwater combat on the water, so that whether the underwater sound target can be timely and accurately recognized and analyzed is an important factor for mastering the initiative of the combat in the ocean war. Because the purity of the real-time data information of the ocean audio is not high, the accuracy of the model for predicting the data is not high enough when training is carried out by applying some conventional algorithms, and the sample data can not be accurately identified well. Although the neural network (CNN) based algorithm can identify data information relatively well, the structure of the algorithm can cause the algorithm to miss some time-related data information; long-short term neural networks (LSTM) have a good effect on the identification of temporal features of data information, but no CNN has a good effect on the processing of data with spatial features.
The underwater sound target recognition is based on a neural network (CNN) algorithm, and currently common computing platforms comprise a CPU, a GPU, an application specific ASIC integrated circuit and the like. The CPU has the disadvantages of low running speed and high power consumption. The GPU has the advantages of high flexibility and good universality, and has the defect of relatively low utilization rate/efficiency; application specific ASIC integrated circuits can achieve extremely high efficiency and integration of specific algorithms, but without flexibility, can support a smaller range of algorithms.
Disclosure of Invention
The invention aims to provide an unmanned underwater sound target recognition system based on an FPGA, which is used for solving the problems existing in the prior art.
In order to achieve the above purpose, the unmanned underwater sound target recognition system based on the FPGA is used for realizing the underwater sound target recognition reasoning function of Zynq UltraScale+MPSoCs series embedded platform based on the Sitting Xilinx based on an EDA tool chain, and consists of a data acquisition card, an FPGA, a CPU board card and other accessories, and the unmanned underwater sound target recognition system is used for designing the data acquisition card, the FPGA, the CPU board card and the other accessories on a main control board to improve the stability and the reliability of the system;
the main control board integrates the FPGA, the CPU board card, the atomic clock and/or the GPS module together, so that the number of system components is reduced, the FPGA, the CPU board card and the data acquisition card all adopt network switch modules to carry out network communication, and the connection between the FPGA and the CPU and the network switch modules is in-board connection, so that the connection is stable and the communication is reliable;
the data acquisition card sends acquired underwater sound data to an ARM core PS in the FPGA through the Ethernet, the PS transmits the underwater sound data to a logic part PL of the FPGA through an AXI bus, the PL performs underwater sound signal preprocessing by using an IP core of a Mel frequency cepstrum coefficient MFCC, and the processed data is sent to a convolutional neural network CNN to realize underwater sound target identification reasoning.
Furthermore, the main control board leads out two RJ45 standard network connectors, which is convenient for communication with an external data acquisition card and other devices.
Further, the atomic clock and/or the GPS module on the main control board are/is communicated with the CPU board card through a COM port.
Furthermore, the GPS antenna of the atomic clock and/or the GPS module is led in through a standard SMA, and meanwhile, the 1PPS signal of the atomic clock and/or the GPS module is led out to an external data acquisition card.
Further, the main control board is connected with a display and a keyboard and a mouse through HDMI, VGA, USB interfaces.
The method of the invention has the following advantages:
according to the invention, the FPGA is used for realizing the underwater sound target recognition algorithm based on the convolutional neural network, so that the operation time of the underwater sound target recognition algorithm is shortened, the one-time calculation time is less than 0.5s, the equipment volume is reduced, the equipment power consumption is reduced, and the method is more beneficial to the application in a system.
Drawings
FIG. 1 is a schematic diagram of a main control board;
FIG. 2 is a schematic diagram of a system connection;
FIG. 3 is a schematic diagram of the system mechanism;
FIG. 4 is a schematic diagram of the internal structure and interfaces of the system;
FIG. 5 is a schematic diagram of a system software implementation;
FIG. 6 is a flow chart of the FPGA implementation MFCC.
Detailed Description
The technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments, but it should be understood by those skilled in the art that the embodiments described below are only for illustrating the present invention and should not be construed as limiting the scope of the present invention. All other embodiments, which can be made by one of ordinary skill in the art without undue burden on the person of ordinary skill in the art based on the embodiments of the present invention, are within the scope of the present invention.
The unmanned platform underwater sound target recognition FPGA hardware system consists of a data acquisition card, an FPGA, a CPU board card, other accessories and the like, and a system main control board is designed for improving the stability and the reliability of the system.
The main control board is designed as shown in figure 1, and integrates an FPGA, a CPU, an atomic clock/GPS and the like, so that the number of system components is reduced. The FPGA, the CPU and the data acquisition card all adopt network communication, so that a network switch module is designed. The connection between the FPGA and the CPU and the network switch is in-board connection, and the connection is stable and reliable. Two RJ45 standard network connectors are led out simultaneously, so that communication with an external data acquisition card and other devices is facilitated.
The main control board is provided with an atomic clock/GPS module which is communicated with the CPU through a COM port, the GPS antenna is led in through a standard SMA, and meanwhile PPS signals of the module 1 are led out to the data acquisition card.
The main control board design HDMI, VGA, USB and other interfaces are connected with the display, the keyboard and the mouse and other devices, and the system connection is shown in figure 2.
The structure adopts the portable reinforcement machine case, as shown in fig. 3, and the integrated design is reliable in connection and can effectively protect internal equipment.
The internal structure design mainly considers the whole counterweight, heat conduction and shock absorption design of the equipment. The internal structure is shown in fig. 4, and the data acquisition card and the main control board are installed in the system. The interfaces of VGA, HDMI, USB, net mouth and the like of the main control board are connected to the outside of the shell for users to use. Two cooling fans are arranged at the lower part of the shell, and air flows through the two board cards in parallel to provide effective heat dissipation.
The internal structure design mainly considers the whole counterweight, heat conduction and shock absorption design of the equipment. The internal structure is shown in fig. 4, and the data acquisition card and the main control board are installed in the system. The interfaces of VGA, HDMI, USB, net mouth and the like of the main control board are connected to the outside of the shell for users to use. Two cooling fans are arranged at the lower part of the shell, and air flows through the two board cards in parallel to provide effective heat dissipation.
The implementation of system software is based on the underwater sound target recognition reasoning of the Sirtz (Xilinx) ZynqUltraScale+MPSoCs series embedded platform. The implementation principle is shown in fig. 5, the acquisition card sends acquired underwater sound data to PS (PS is an FPGA inner ARM core) of the FPGA through the Ethernet, the PS transmits the underwater sound data to PL (PL is an FPGA logic part) through an AXI bus, the PL uses MFCC (mei-frequency cepstral coefficients, mel frequency cepstrum coefficient) IP to verify underwater sound signal preprocessing, and the processed data is sent to CNN (Convolutional NeuralNetworks, convolutional neural network) to realize underwater sound target recognition reasoning.
MFCC pre-processing IP core design
FPGA implementation MFCC workflow as shown in figure 6,
firstly, a PS in an FPGA operates an embedded linux operating system, and underwater sound data are acquired in real time by connecting a data acquisition card through an Ethernet. The MFCC preprocessing IP core is mounted on the PS through an AXI bus, and the PS calls the MFFC preprocessing IP core to realize underwater sound data preprocessing after receiving underwater sound data.
The working principle of the MFCC preprocessing IP mainly comprises the steps of obtaining audio data, pre-emphasis, framing, windowing, fourier transformation, a Mel filter, discrete cosine transformation and the like.
Acquiring audio data: the MFCC preprocesses the IP core to acquire audio data through the AXI bus.
Pre-emphasis: the pre-emphasis passes the audio signal through a high-pass filter in order to raise the high-frequency part, flatten the spectrum of the signal, remain in the whole frequency band from low frequency to high frequency, and can use the same signal-to-noise ratio to find the spectrum. The pre-emphasis is achieved by a high pass filter after receiving the audio data.
Framing: the audio signal is framed.
Windowing: after the audio signal is framed, a window function is carried into each frame of signal so as to increase the continuity of two ends of each frame of signal and prevent frequency spectrum leakage.
Fourier transform: and carrying out Fourier transform on each frame of signals subjected to windowing to obtain each frame of frequency spectrum. The Fourier transform involves a large number of multiplication operations, in order to increase the calculation speed, a method of changing the speed by an area is adopted in the FPGA, and a parallel processing method is adopted, so that each frame signal is subjected to Fourier transform at the same time.
Mel filter: the mel filter bank also requires a large number of multiplication operations, and a parallel processing method is also adopted to increase the calculation speed.
Discrete cosine transform: the discrete cosine transform implements decorrelation and dimension reduction, and finally outputs MFCC parameters.
And the data after the MFCC pretreatment is sent to an underwater sound signal target identification IP core (CNN core in the figure) to finish the identification of the underwater sound signal and output a propulsion result to a computer.
While the invention has been described in detail in the foregoing general description and specific examples, it will be apparent to those skilled in the art that modifications and improvements can be made thereto. Accordingly, such modifications or improvements may be made without departing from the spirit of the invention and are intended to be within the scope of the invention as claimed.
Claims (5)
1. An unmanned underwater sound target recognition system based on an FPGA is used for realizing an underwater sound target recognition reasoning function of ZynqUltraScale+MPSoCs series embedded platform based on Sitting Xilinx based on an EDA tool chain, and consists of a data acquisition card, an FPGA, a CPU board card and other accessories, wherein the unmanned underwater sound target recognition system is used for designing the data acquisition card, the FPGA, the CPU board card and the other accessories on a main control board in order to improve the stability and the reliability of the system;
the main control board integrates the FPGA, the CPU board card, the atomic clock and/or the GPS module together, so that the number of system components is reduced, the FPGA, the CPU board card and the data acquisition card all adopt network switch modules to carry out network communication, and the connection between the FPGA and the CPU and the network switch modules is in-board connection, so that the connection is stable and the communication is reliable;
the data acquisition card sends acquired underwater sound data to an ARM core PS in the FPGA through the Ethernet, the PS transmits the underwater sound data to a logic part PL of the FPGA through an AXI bus, the PL performs underwater sound signal preprocessing by using an IP core of a Mel frequency cepstrum coefficient MFCC, and the processed data is sent to a convolutional neural network CNN to realize underwater sound target identification reasoning.
2. The FPGA-based unmanned underwater acoustic target recognition system of claim 1, wherein the main control board brings out two RJ45 standard network connectors for facilitating communication with external data acquisition cards and other devices.
3. The FPGA-based unmanned underwater sound target recognition system of claim 2, wherein the atomic clock and/or GPS module on the main control board communicates with the CPU board card through a COM port.
4. The FPGA-based unmanned underwater acoustic target recognition system of claim 3, wherein the GPS antenna of the atomic clock and/or GPS module is introduced through a standard SMA while the 1PPS signal of the atomic clock and/or GPS module is introduced to an external data acquisition card.
5. The unmanned underwater sound target recognition system based on the FPGA of claim 4, wherein the main control board is connected with a display and a keyboard and a mouse through HDMI, VGA, USB interfaces.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311505539.8A CN117648601A (en) | 2023-11-13 | 2023-11-13 | Unmanned underwater sound target recognition system based on FPGA |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311505539.8A CN117648601A (en) | 2023-11-13 | 2023-11-13 | Unmanned underwater sound target recognition system based on FPGA |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117648601A true CN117648601A (en) | 2024-03-05 |
Family
ID=90044173
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311505539.8A Pending CN117648601A (en) | 2023-11-13 | 2023-11-13 | Unmanned underwater sound target recognition system based on FPGA |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117648601A (en) |
-
2023
- 2023-11-13 CN CN202311505539.8A patent/CN117648601A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110503971A (en) | Time-frequency mask neural network based estimation and Wave beam forming for speech processes | |
| CN112735473B (en) | Method and system for identifying unmanned aerial vehicle based on voice | |
| Liu et al. | Deep learning and recognition of radar jamming based on CNN | |
| CN109344772A (en) | Ultrashort wave signal specific reconnaissance method based on spectrogram and depth convolutional network | |
| CN104732970A (en) | Ship radiation noise recognition method based on comprehensive features | |
| CN117388835B (en) | Multi-spelling fusion sodar signal enhancement method | |
| CN112528775A (en) | Underwater target classification method | |
| Tong et al. | Classification and recognition of underwater target based on MFCC feature extraction | |
| CN111735525B (en) | DEMON spectral feature extraction method suitable for unmanned sonar | |
| CN115510909A (en) | Unsupervised algorithm for DBSCAN to perform abnormal sound features | |
| Liu et al. | Underwater acoustic target recognition based on dual attention networks and multiresolution convolutional neural networks | |
| CN113640768B (en) | Low-resolution radar target identification method based on wavelet transformation | |
| CN114692687A (en) | Underwater sound signal identification method and training method of underwater sound signal identification model | |
| Wang et al. | Ship radiated noise recognition technology based on ML-DS decision fusion | |
| CN117648601A (en) | Unmanned underwater sound target recognition system based on FPGA | |
| CN116449328A (en) | Radar working mode identification method and system based on time-frequency analysis feature fusion | |
| CN114019452A (en) | Shallow sea target depth resolution method based on vertical beam spectrum space difference factor | |
| CN106782550A (en) | A kind of automatic speech recognition system based on dsp chip | |
| Huang et al. | A method for extracting fingerprint feature of communication satellite signal | |
| CN111624585A (en) | Underwater target passive detection method based on convolutional neural network | |
| Yi et al. | Multiresolution convolutional neural network for underwater acoustic target recognition | |
| Wang | [Retracted] Audio Signal Acquisition and Processing System Based on Model DSP Rapid Design | |
| Wang et al. | Auditory-based multi-scale amplitude-aware permutation entropy as a measure for feature extraction of ship radiated noise | |
| Wang et al. | Underwater Acoustic Target Recognition Combining Multi-scale Features and Attention Mechanism | |
| CN113777610B (en) | Low-power-consumption underwater acoustic target detection system suitable for small platform |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |