A miniature detection system based on the identification of ship magnetic field characteristics

Abstract

The present invention discloses a miniature detection system based on the recognition of ship magnetic field characteristics, belonging to the field of ship target recognition, which includes a power supply module, a collection and storage module, a single-chip microcomputer, an identification result transmission module, an attitude and water depth collection module, and a trigger module; by constructing a support vector machine classifier, the present invention can have stronger learning performance and generalization ability for small-sample signals, with higher recognition accuracy, lower complexity than artificial neural networks, higher real-time performance than neural networks, and less memory required for the embedded device, making it more suitable for using magnetic signals on the embedded platform, effectively improving the efficiency of ship recognition.

Description

Technical Field

[0001] The present invention relates to the field of ship target recognition, and particularly to a micro-detection system based on ship magnetic field characteristics recognition. Background Art

[0002] The accurate recognition of ships plays an irreplaceable role in protecting sovereignty. In addition, the continuous entry and exit of ships at major ports require the maritime target detection system to play a more powerful monitoring role.

[0003] Therefore, it is necessary to identify maritime targets. In the field of ship target recognition, the detection system can distinguish different types of ships through appropriate signal processing and target decision-making, and exhibit good performance. However, most detection systems based on ship magnetic field characteristics recognition require a computer to participate in the recognition process, the recognition results cannot be transmitted over long distances, are greatly affected by sea conditions and weather, and are large in size and short in battery life, that is, the autonomous recognition and wireless transmission of the detection system, the elimination of the influence of the maritime environment, the easy deployment and long-term operation of the detection system are not considered. With the improvement of the computing power of single-chip microcomputers, the deepening of the integration degree of components, and the reduction of device power consumption, it makes a micro-detection system possible. Compared with the traditional detection system, this micro-detection system is smaller in size, requires less power, and has a longer working time;

[0004] The task of the micro-detection system based on ship magnetic field characteristics recognition is to answer when a certain type of ship passes by. Different from other physical signals of ships, the magnetic signal of ships is relatively more stable, has a smaller amount of data, is less affected by the marine environment, and reflects more information. As a kind of object detection task, the recognition based on ship magnetic field characteristics depends on the selection and extraction of ship characteristics. It is certain that there is a non-linear relationship between the ship magnetic field and its structural characteristics such as length and width. Therefore, the type of ship can be inferred based on this corresponding relationship. The autonomous recognition detection system requires the target recognition algorithm to run on its own embedded platform. The ARM-Cortex microcontroller / processor software interface standard CMSIS-DSP library provides a variety of support vector machines. The trained support vector machine classification algorithm can better learn the structural information of the ship and the characteristics of draft and speed, and complete the judgment of the ship type. For this reason, we propose a micro-detection system based on ship magnetic field characteristics recognition. Summary of the Invention

[0005] The purpose of the present invention is to solve the defects existing in the prior art, and to propose a micro-detection system based on ship magnetic field characteristics recognition.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions:

[0007] A micro-detection system based on the identification of ship magnetic field characteristics, including a power supply module, a collection and storage module, a single-chip microcomputer, an identification result transmission module, an attitude and water depth collection module, and a trigger module;

[0008] Among them, the power supply module obtains the voltage required by other modules by using a 3.7V lithium battery and a voltage conversion chip, and constructs a power supply board to supply power to other modules;

[0009] The collection and storage module uses a fluxgate sensor and an ADS1220 chip to collect the magnetic signals of the ship, and uses an SD card to store the collected magnetic signals of the ship;

[0010] The single-chip microcomputer is used to control each sub-module and process data;

[0011] The identification result transmission module is used to transmit the identification result by means of Beidou short message communication;

[0012] The attitude and water depth collection module is used to collect the attitude information and water depth information of the entire detection system;

[0013] The trigger module is used to receive the single-chip microcomputer control instruction and drive the detection system to float to the water surface.

[0014] As a further solution of the present invention, the specific power supply method of the power supply module is as follows: the 3.7V battery voltage obtains a 5V power supply voltage through the LT1619 chip to supply power to the identification result transmission module and the trigger module;

[0015] The 3.7V battery voltage obtains a 3.3V power supply voltage through the LM317 chip to supply power to the single-chip microcomputer;

[0016] The 3.3V positive voltage obtains a -3.3V negative voltage through the LTC1983ES6-5 chip, and the 3.3V positive and negative voltages supply power to the fluxgate sensor;

[0017] The 3.3V positive and negative voltages respectively obtain positive and negative 2.5V power supply voltages through the LTC1983ES6-5 chip and the LT1964ES5-BYP chip to supply power to the ADS1220 chip.

[0018] As a further solution of the present invention, the fluxgate sensor is a 3-axis sensor with a range of 60μT, and it is specifically used to convert the magnetic signal into an analog signal;

[0019] The ADS1220 chip is a 24-bit analog-to-digital conversion chip, which is responsible for converting the analog signal into a digital signal and sending the digital signal to the single-chip microcomputer through SPI.

[0020] As a further solution of the present invention, the specific steps of the single-chip microcomputer data processing are as follows:

[0021] Step 1: Select appropriate resistors and capacitors to form a low-pass filter in the signal acquisition channel, and perform hardware denoising on the collected ship magnetic signals. Then, construct a digital filter and set the cut-off frequency of the digital filter at about 1 Hz according to the upper frequency limit of the ship's magnetic field.

[0022] Step 2: The digital filter performs FIR filtering on the ship magnetic signals. Then, the LMS adaptive filter receives the filtered ship magnetic signals, further filters out the colored noise, updates the current threshold through the reference value weighting function, and intercepts the useful information in the filtered ship magnetic signals according to the updated threshold.

[0023] Step 3: Collect the structural information and general speeds of various types of ships, construct the training set and test set of the support vector machine classifier, determine the type of kernel function and the parameters of the classifier in the support vector machine, train it with the training set, and evaluate the training results with the test set and accuracy rate to obtain the optimal support vector machine classifier.

[0024] Step 4: Then, according to the spatial characteristics of the ship's magnetic field, the ship is equivalent to a uniformly magnetized rotating ellipsoid, with the origin o of the ship's coordinates as the center of the ship; the positive direction of the z-axis is perpendicular to the deck and points towards the sea surface; the positive direction of the x-axis is along the bow and stern of the ship and points towards the bow; the y-axis is perpendicular to the xoz plane and points towards the starboard side.

[0025] Step 5: According to the processed ship magnetic signals, use the least squares method to calculate the parameters of the corresponding rotating ellipsoid model, judge the possible conforming ship types through the major axis and minor axis of the ellipsoid, and then import the relationship between the rotating ellipsoid and the position of the recognition module and the information such as the draft depth and speed of the ship deduced through the characteristic curve into the support vector machine classifier, and further infer the conforming ship types through the support vector machine classifier.

[0026] As a further solution of the present invention, the specific formula of the iterative formula of the weight vector of the LMS adaptive filter in Step 2 is as follows:

[0027] e(n) = d(n) - y(n) (1)

[0028] g(n) = -2e(n)x(n) ≈ Δ(n) (2)

[0029] w(n + 1) = w(n) - μg(n) = w(n) + 2μe(n)x(n) (3)

[0030] Wherein, d(n) represents the desired signal, y(n) represents the output signal, g(n) represents the gradient of the single-sample value squared error sequence, x(n) represents the input signal, and Δ(n) represents the average gradient of the multiple-sample value squared error sequence;

[0031] The specific calculation formula of the reference value weighting function described in step two is as follows:

[0032]

[0033] Wherein, B (k-1) represents the threshold before update, B (k) represents the threshold after update, α represents the filtered sample value, A (k) represents the weighting coefficient, and are the variances of B (k-1) and B (k) respectively. Among them, the value of A (k) is determined by the reference value update rate N, and is generally the average value of N - 1 data points between the updates of B (k-1) and B (k) .

[0034] As a further solution of the present invention, the specific types of ships described in step three specifically include transport ships, engineering ships, fishing ships, work ships, and military ships.

[0035] As a further solution of the present invention, the recognition result transmission module is divided into a sending end and a receiving end, and is specifically composed of a Beidou core board and sending and receiving antennas. Both the sending end and the receiving end are equipped with Beidou short message civilian level three cards and follow the RDSS communication protocol.

[0036] As a further solution of the present invention, the attitude and water depth acquisition module specifically uses a JY901 nine-axis attitude sensor and a pressure sensor to acquire the attitude information and water depth information of the entire detection system.

[0037] As a further solution of the present invention, the trigger module is composed of a motor and an airbag. Among them, the trigger module controls the movement of the motor through the I / O port of the single-chip microcomputer, fills the airbag with gas, and drives the detection system to float to the water surface through the airbag.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0039] Compared with previous detection systems, the proposed micro-detection system based on ship magnetic field characteristics constructs a training set and a test set for a support vector machine classifier by collecting the structural information and general speeds of various types of ships, determines the type of kernel function and the parameters of the classifier in the support vector machine, trains it with the training set, and evaluates the training results using the test set and accuracy rate to obtain the optimal support vector machine classifier. Then, according to the spatial characteristics of the ship magnetic field, the ship is equivalent to a uniformly magnetized rotating ellipsoid, and each group of coordinate information is determined. Based on the processed ship magnetic signals, the parameters of the corresponding rotating ellipsoid model are deduced by the least squares method. The possible ship types are judged by the major axis and minor axis of the ellipsoid. Then, the relationship between the rotating ellipsoid and the position of the recognition module and information such as the draft depth and speed of the ship deduced through characteristic curves are imported into the support vector machine classifier, and the ship types that meet the conditions are further speculated through the support vector machine classifier. By constructing a support vector machine classifier, it has stronger learning performance and generalization ability for small sample signals, higher recognition accuracy, lower complexity than artificial neural networks, higher real-time performance relative to neural networks, and less memory required for the embedded devices. It is more suitable for using magnetic signals on an embedded platform, effectively improving the efficiency of ship recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention.

[0041] Figure 1 It is a system block diagram of a micro-detection system based on ship magnetic field characteristics proposed by the present invention;

[0042] Figure 2 It is a judgment flowchart of a micro-detection system based on ship magnetic field characteristics proposed by the present invention;

[0043] Figure 3 It is a functional block diagram of a micro-detection system based on ship magnetic field characteristics proposed by the present invention;

[0044] Figure 4 It is a schematic diagram of a uniformly magnetized rotating ellipsoid model constructed by a micro-detection system based on ship magnetic field characteristics proposed by the present invention;

[0045] Figure 5 It is a micro-detection system based on ship magnetic field characteristics proposed by the present invention;

[0046] Figure 6 It is an acquisition and storage module of a micro-detection system based on ship magnetic field characteristics proposed by the present invention;

[0047] Figure 7 The recognition result transmission module of a micro detection system based on ship magnetic field characteristic recognition proposed by the present invention;

[0048] Figure 8 The attitude acquisition module of a micro detection system based on ship magnetic field characteristic recognition proposed by the present invention;

[0049] Figure 9 The water depth acquisition module of a micro detection system based on ship magnetic field characteristic recognition proposed by the present invention;

[0050] Figure 10 The trigger module of a micro detection system based on ship magnetic field characteristic recognition proposed by the present invention. Specific implementation mode

[0051] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0052] Refer to Figures 1 - 10 , a micro detection system based on ship magnetic field characteristic recognition, including a power supply module, an acquisition and storage module, a single-chip microcomputer, a recognition result transmission module, an attitude and water depth acquisition module, and a trigger module.

[0053] The power supply module obtains the voltage required by other modules by using a 3.7V lithium battery and a voltage conversion chip, and constructs a power supply board to supply power to other modules.

[0054] Specifically, the specific power supply method of the power supply module is as follows: the 3.7V battery voltage obtains a 5V power supply voltage through the LT1619 chip to supply power to the recognition result transmission module and the trigger module; the 3.7V battery voltage obtains a 3.3V power supply voltage through the LM317 chip to supply power to the single-chip microcomputer; the 3.3V positive voltage obtains a -3.3V negative voltage through the LTC1983ES6-5 chip, and the 3.3V positive and negative voltages supply power to the fluxgate sensor; the 3.3V positive and negative voltages respectively obtain positive and negative 2.5V power supply voltages through the LTC1983ES6-5 chip and the LT1964ES5-BYP chip to supply power to the ADS1220 chip.

[0055] The acquisition and storage module uses a fluxgate sensor and an ADS1220 chip to acquire the magnetic signals of the ship, and uses an SD card to store the acquired magnetic signals of the ship.

[0056] Specifically, the fluxgate sensor is a 3-axis sensor with a range of 60 μT, which is specifically used to convert magnetic signals into analog signals; the ADS1220 chip is a 24-bit analog-to-digital conversion chip, responsible for converting analog signals into digital signals and sending the digital signals to the single-chip microcomputer through SPI. To ensure that the data has sufficient information and reduce the data volume, the sampling rate of the AD conversion chip is set to 5 Hz.

[0057] It should be further noted that the SD card communicates with the single-chip microcomputer through SDIO, and the single-chip microcomputer creates a data file inside the SD card through the FATFS file system. The collected data is first cached in the RAM of the single-chip microcomputer and then written into the data file of the SD card to achieve data storage.

[0058] The single-chip microcomputer is used to control each sub-module and process data.

[0059] Specifically, researchers select appropriate resistors and capacitors to form a low-pass filter in the signal acquisition channel and perform hardware denoising processing on the collected ship magnetic signals. Then, a digital filter is constructed, and the cut-off frequency of the digital filter is set to about 1 Hz according to the upper frequency limit of the ship's magnetic field. Then, the digital filter performs FIR filtering on the ship magnetic signals. After that, the LMS adaptive filter receives the filtered ship magnetic signals and further filters out the colored noise. Then, the current threshold is updated through the reference value weighting function, and the useful information in the filtered ship magnetic signals is intercepted according to the updated threshold. The structural information and general sailing speed of various types of ships are collected to construct the training set and test set of the support vector machine classifier. The type of kernel function and the parameters of the classifier in the support vector machine are determined, and it is trained with the training set. The training results are evaluated with the test set and accuracy rate to obtain the optimal support vector machine classifier. Then, according to the spatial characteristics of the ship's magnetic field, the ship is equivalent to a uniformly magnetized rotating ellipsoid, and the origin o of the ship's coordinates is taken as the center of the ship; the positive direction of the z-axis is perpendicular to the deck and points to the sea surface; the positive direction of the x-axis is along the head and tail of the ship and points to the bow direction; the y-axis is perpendicular to the xoz plane and points to the starboard direction. Then, according to the processed ship magnetic signals, the parameters of the corresponding rotating ellipsoid model are deduced by the least squares method. The possible ship types are judged by the major axis and minor axis of the ellipsoid. Then, the relationship between the rotating ellipsoid and the position of the recognition module and the draft depth and sailing speed of the ship deduced through the characteristic curve are imported into the support vector machine classifier, and the ship types that meet the conditions are further inferred through the support vector machine classifier.

[0060] It should be further noted that the specific formula for the iterative formula of the weight vector of the LMS adaptive filter is as follows:

[0061] e(n) = d(n) - y(n) (1)

[0062] g(n) = -2e(n)x(n) ≈ Δ(n) (2)

[0063] w(n + 1) = w(n) - μg(n) = w(n) + 2μe(n)x(n) (3)

[0064] In the formula, d(n) represents the desired signal, y(n) represents the output signal, g(n) represents the gradient of the single - sample - value squared - error sequence, x(n) represents the input signal, and Δ(n) represents the gradient averaged from the multiple - sample - value squared - error sequence;

[0065] The specific calculation formula of the reference - value weighting function is as follows:

[0066]

[0067] In the formula, B (k-1) represents the threshold before update, B (k) represents the threshold after update, α represents the filtered sample value, A (k) represents the weighting coefficient, and are the variances of B (k-1) and B (k) respectively. Among them, the value of A (k) is determined by the reference - value update rate N, and is generally the average value of N - 1 data points between the update of B (k-1) and B (k)

[0068] In addition, it should be further explained that the collection methods of the structural information and general speed information of various types of ships are specifically online collection and collection through other channels. Among them, the ship types specifically include transport ships, engineering ships, fishing ships, working ships, and military ships.

[0069] In this embodiment, during the single - chip microcomputer ship recognition test, the detection system is placed on the seabed of the sea area where the ship passes, observe the types of ships passing through this sea area at certain time points, compare the recognition results received by the receiving end with the ship - passing time, check for missed detections, whether the recognition results are accurate, and calculate the missed - detection rate and the recognition accuracy rate.

[0070] The recognition - result transmission module is used to transmit the recognition result by means of Beidou short - message communication.

[0071] Specifically, this recognition - result transmission module is divided into a sending end and a receiving end, which are specifically composed of a Beidou core board and sending and receiving antennas, and both the sending end and the receiving end are equipped with Beidou short - message civilian - level three - cards and follow the RDSS communication protocol.

[0072] ​It should be further noted that the specific transmission steps of the recognition result are as follows: First, the card number, time information, and recognition result of the receiving end are combined into communication information. The single-chip microcomputer organizes the communication information into a CCTXA command and sends it to the Beidou core board of the sending end through the serial port. The Beidou core board sends the CCTXA command through the sending antenna, and then it is relayed by the Beidou satellite into the ground station. After receiving the communication information, the ground station decrypts and then encrypts it and adds it to the continuously broadcasted telegram. Finally, it is broadcasted to the receiving end by the Beidou satellite. The receiving end decrypts the communication information, and thus the transmission of the recognition result is completed once.

[0073] The attitude and water depth acquisition module is used to acquire the attitude information and water depth information of the entire detection system.

[0074] Specifically, the attitude and water depth acquisition module uses a JY901 nine-axis attitude sensor and a pressure sensor to acquire the attitude information and water depth information of the entire detection system. Among them, the JY901 nine-axis attitude sensor has been integrated into a postage stamp patch and can be directly soldered onto the single-chip microcomputer board. The acquired attitude information is sent to the single-chip microcomputer through the serial port. At the same time, the probe of the pressure sensor extends out of the housing, and the water depth information is sent to the single-chip microcomputer through IIC.

[0075] The trigger module is used to receive the single-chip microcomputer control instruction and drive the detection system to float to the water surface.

[0076] Specifically, the trigger module consists of a motor and an airbag. Among them, the trigger module controls the movement of the motor through the I / O port of the single-chip microcomputer to fill the airbag with gas, and drives the detection system to float to the water surface through the airbag.

Claims

1. A micro-detection system based on the identification of ship magnetic field characteristics, characterized in that, It includes a power supply module, a data acquisition and storage module, a single-chip microcomputer, an identification result transmission module, an attitude and water depth acquisition module, and a trigger module; Among them, the power supply module obtains the voltage required by other modules by using a 3.7V lithium battery and a voltage conversion chip, and constructs a power supply board to supply power to other modules; The data acquisition and storage module uses a fluxgate sensor and an ADS1220 chip to collect the magnetic signals of the ship, and uses an SD card to store the collected magnetic signals of the ship; The single-chip microcomputer is used to control each sub-module and process data; The identification result transmission module is used to transmit the identification result by means of Beidou short message communication; The attitude and water depth acquisition module is used to collect the attitude information and water depth information of the entire detection system; The trigger module is used to receive the control instruction of the single-chip microcomputer and drive the detection system to float to the water surface; The specific steps of the single-chip microcomputer data processing are as follows: Step 1: Select appropriate resistors and capacitors to form a low-pass filter in the signal acquisition channel, and perform hardware denoising on the collected magnetic signals of the ship. Then, construct a digital filter, and set the cut-off frequency of the digital filter to about 1Hz according to the upper limit of the frequency of the ship's magnetic field; Step 2: The digital filter performs FIR filtering on the magnetic signals of the ship. Then, the LMS adaptive filter receives the filtered magnetic signals of the ship and further filters out the colored noise. Then, the current threshold is updated through the reference value weighting function, and the useful information in the filtered magnetic signals of the ship is intercepted according to the updated threshold; Step 3: Collect the structural information and general sailing speeds of various types of ships, construct the training set and test set of the support vector machine classifier, determine the type of kernel function and the parameters of the classifier in the support vector machine, train it with the training set, and evaluate the training result with the test set and accuracy rate to obtain the optimal support vector machine classifier; Step 4: Then, according to the spatial characteristics of the ship's magnetic field, the ship is equivalent to a uniformly magnetized rotating ellipsoid, and the origin o of the ship's coordinates is taken as the center of the ship; the positive direction of the z-axis is perpendicular to the deck and points to the sea surface; the positive direction of the x-axis is along the bow and stern of the ship and points to the bow direction; the y-axis is perpendicular to the xoz plane and points to the starboard direction; Step 5: According to the processed magnetic signals of the ship, use the least squares method to calculate the parameters of the corresponding rotating ellipsoid model, judge the possible conforming ship types through the major axis and minor axis of the ellipsoid, and then import the relationship between the rotating ellipsoid and the position of the identification module and the draft depth and sailing speed information of the ship deduced through the characteristic curve into the support vector machine classifier, and further infer the conforming ship types through the support vector machine classifier; The specific formula of the iterative formula of the weight vector of the LMS adaptive filter in Step 2 is as follows: In the formula, represents the desired signal, represents the output signal, represents the gradient of the square error sequence of individual sampling values, represents the input signal, represents the gradient averaged over the square error sequences of multiple sampling values; The specific calculation formula of the reference value weighting function in Step 2 is as follows: In the formula, represents the threshold before update, represents the threshold after update, represents the sampled value after filtering, represents the weighting coefficient, and are respectively and variances of, where the value of is determined by the reference value update rate N and is to the average value of N - 1 data points between updates.

2. The micro-detection system based on the identification of the magnetic field characteristics of a ship according to claim 1, wherein The specific power supply method of the power supply module is as follows: the 3.7V battery voltage obtains a 5V power supply voltage through the LT1619 chip to supply power to the identification result transmission module and the trigger module; The 3.7V battery voltage obtains a 3.3V power supply voltage through the LM317 chip to supply power to the single-chip microcomputer; The 3.3V positive voltage obtains a -3.3V negative voltage through the LTC1983ES6-5 chip, and the 3.3V positive and negative voltages supply power to the fluxgate sensor; The 3.3V positive and negative voltages respectively obtain positive and negative 2.5V supply voltages through the LTC1983ES6-5 chip and the LT1964ES5-BYP chip to supply power to the ADS1220 chip.

3. A micro-detection system based on the identification of the magnetic field characteristics of a ship, as claimed in claim 1, wherein The fluxgate sensor is a 3-axis sensor with a range of 60 μT, and it is specifically used to convert magnetic signals into analog signals; The ADS1220 chip is a 24-bit analog-to-digital conversion chip, which is responsible for converting analog signals into digital signals and sending the digital signals to the single-chip microcomputer through SPI.

4. A miniature detection system based on the identification of ship magnetic field characteristics according to claim 1, characterized in that, The ship types described in step three specifically include transport ships, engineering ships, fishing ships, working ships, and military ships.

5. A miniature detection system based on the identification of the magnetic field characteristics of a ship, characterized in that, The recognition result transmission module is divided into a sending end and a receiving end. Specifically, it is composed of a Beidou core board and sending and receiving antennas, and both the sending end and the receiving end are equipped with Beidou short message civilian level-3 cards and follow the RDSS communication protocol.

6. The micro-detection system based on the identification of the magnetic field characteristics of a ship according to claim 1, characterized in that, The attitude and water depth acquisition module specifically uses the JY901 nine-axis attitude sensor and the pressure sensor to acquire the attitude information and water depth information of the entire detection system.

7. A miniature detection system based on the identification of the magnetic field characteristics of a ship, characterized in that, The trigger module is composed of a motor and an airbag. Among them, the trigger module controls the movement of the motor through the I / O port of the single-chip microcomputer to fill the airbag with gas, and drives the detection system to float to the water surface through the airbag.

Images