An acoustic-magnetic feature fusion device applied to intersection area identification

Abstract

The application discloses a kind of acoustic-magnetic feature fusion devices applied to intersection area identification, acoustic-magnetic feature acquisition module is used to produce electromagnetic, acoustic detection signal and collect the characteristic data of electromagnetic field and sound field in intersection process;Confidence optimization module is used to iteratively solve the optimal confidence value of each characteristic quantity;Feature combination reliability calculation module is used to solve the reliability value of each characteristic quantity combination;Support conversion module is used to convert the data collected into the area identification support data of each characteristic quantity in intersection process;Acoustic-magnetic feature data composite module is used to solve the composite characteristic value of acoustic-magnetic feature;Feature combination reliability calculation module is used to solve the reliability value of each characteristic quantity combination;Acoustic-magnetic feature fusion module is used for real-time fusion of acoustic-magnetic intersection feature;Fusion feature comprehensive decision module is used for real-time dynamic identification and decision to target area in intersection process.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of target area autonomous identification device, and particularly to a sound-magnetic feature fusion device applied to intersection area identification. BACKGROUND

[0002] Affected by the conductivity of seawater, the energy of electromagnetic wave will be severely attenuated with the increase of propagation distance when propagating in seawater, so that the detection application of electromagnetic wave in seawater is limited to low frequency and near field, and it is difficult to directly obtain the distance information of the target, but it has the advantages of being able to distinguish ferromagnetic and non-ferromagnetic targets, and not being easily affected by environmental factors and having good detection reliability. The sound field is less affected by seawater medium, and can propagate over a long distance in seawater, and it is easy to directly obtain the distance information of the target, but it is also easily affected by reverberation or non-metallic impurities in the environment, thereby reducing the identification performance of the system to the intersection area of the target. SUMMARY

[0003] To solve the problems raised in the background art, the technical scheme adopted by the present application is as follows:

[0004] A sound-magnetic feature fusion device applied to intersection area identification, comprising a sound-magnetic feature acquisition module, a confidence optimization module, a feature combination reliability calculation module, a support conversion module, a sound-magnetic feature data composite module, a feature combination reliability calculation module, a sound-magnetic feature fusion module, and a fusion feature comprehensive decision module.

[0005] The sound-magnetic feature acquisition module is used to generate electromagnetic and sound detection signals and acquire feature data of electromagnetic field and sound field during the intersection process.

[0006] The confidence optimization module is used to iteratively solve the optimal confidence value of each feature quantity.

[0007] The feature combination reliability calculation module is used to solve the reliability value of each feature quantity combination.

[0008] The support conversion module is used to convert the acquired data into area identification support data of each feature quantity during the intersection process.

[0009] The sound-magnetic feature data composite module is used to solve the composite feature value of the sound-magnetic feature.

[0010] The feature combination reliability calculation module is used to solve the reliability value of each feature quantity combination.

[0011] The sound-magnetic feature fusion module is used to real-time fuse the sound-magnetic intersection features.

[0012] The fusion feature comprehensive decision module is used to real-time dynamic identification and decision of the target area during the intersection process.

[0013] In some embodiments, the acoustic-magnetic feature acquisition module comprises an acoustic-magnetic sensor module, a preprocessing circuit, and a data acquisition circuit, the acoustic-magnetic sensor module comprising a set of electromagnetic transceiving sensors and two sets of acoustic transducer arrays;

[0014] The feature data acquired by the acoustic-magnetic feature acquisition module 1 comprises electromagnetic intersection feature quantities: z-axis electromagnetic field intensity and phase information and x-axis electromagnetic field intensity accepted by the electromagnetic sensor, and acoustic intersection feature quantities: target noise sound pressure amplitude and target azimuth interval.

[0015] In some embodiments, the confidence optimization module obtains optimal confidence values of each feature quantity by iterative optimization and solution from intersection feature learning data according to pre-set initial confidence values of each feature quantity.

[0016] The intersection feature learning data refers to electromagnetic and acoustic feature data acquired by the detection system in a prior intersection process.

[0017] In some embodiments, the feature combination reliability calculation module receives output data of the confidence optimization module, and solves reliability values of each feature quantity combination, the reliability value representing the reliability of the feature combination information source.

[0018] In some embodiments, the support conversion module receives output data of the acoustic-magnetic feature acquisition module, and converts the data into regional identification support data of each feature quantity in the intersection process, the regional identification support representing the contribution of real-time data of a certain feature quantity to the accuracy of target region identification.

[0019] In some embodiments, the acoustic-magnetic feature data complex module receives output data of the support conversion module, and performs complex on the acoustic-magnetic feature to solve a complex feature value.

[0020] In some embodiments, the feature combination reliability calculation module receives output data of the support conversion module, and solves reliability values of each feature quantity combination, the reliability value representing the reliability of the feature combination real-time data.

[0021] In some embodiments, the acoustic-magnetic feature fusion module receives output data of the feature combination reliability calculation module, the acoustic-magnetic feature data complex module, and the feature combination reliability calculation module, and completes real-time fusion of the acoustic-magnetic intersection feature.

[0022] In some embodiments, the fusion feature comprehensive decision module receives output data of the acoustic-magnetic feature fusion module, and completes real-time dynamic identification and decision of the target region in the intersection process.

[0023] The basis for target area identification and judgment in the intersection process is that when the real-time data value output by the acoustic-magnetic feature fusion module is greater than a preset value, it is judged that the system has reached the target area range.

[0024] Compared with the prior art, the acoustic-magnetic feature fusion device for intersection area identification provided by the application has the following beneficial effects:

[0025] The acoustic-magnetic feature fusion device for intersection area identification provided by the application uses electromagnetic fields and acoustic fields in combination, combines the advantages of the two physical fields in underwater detection, and fuses intersection feature information of the two physical fields, thereby effectively improving the area identification effect of the detection system on underwater targets and increasing the fault tolerance and adaptability of the detection system. BRIEF DESCRIPTION OF DRAWINGS

[0026] Figure 1 FIG. 1 is a schematic diagram of the acoustic-magnetic feature fusion device for intersection area identification provided by the application;

[0027] Figure 2a FIG. 2 is a setting diagram for collecting intersection feature quantities in a specific embodiment; Figure 2b

[0028] Figure 3a FIG. 3 is a schematic diagram of the intersection process in a specific embodiment; Figure 3b

[0029] FIG. 4 is an interval division schematic diagram of target orientation interval features in a specific embodiment. Figure 4

[0030] FIG. 5 is a workflow diagram of the acoustic-magnetic feature fusion device for intersection area identification provided by the application. Figure 5 DETAILED DESCRIPTION

[0031] In order to make the technical means, creative features, purposes and effects achieved by the application easy to understand, the following further describes how the application is implemented in combination with the drawings and specific embodiments.

[0032] Referring to FIG. 1, in one specific embodiment, the application provides an acoustic-magnetic feature fusion device for intersection area identification, which includes an acoustic-magnetic feature collection module 1, a confidence optimization module 2, a feature combination reliability calculation module 3, a support conversion module 4, an acoustic-magnetic feature data combination module 5, a feature combination confidence calculation module 6, an acoustic-magnetic feature fusion module 7, and a fusion feature comprehensive decision module 8. Figure 1

[0033] ​​​In this embodiment, the acoustic-magnetic feature acquisition module 1 is used to generate electromagnetic and acoustic detection signals and to acquire electromagnetic field and acoustic field feature data during the intersection process. The acoustic-magnetic feature acquisition module 1 includes an acoustic-magnetic sensor module, a preprocessing circuit, and a data acquisition circuit. The acoustic-magnetic sensor module includes a set of electromagnetic transceiving sensors and two sets of acoustic transducer arrays.

[0034] Referring to Figure 2a , it is shown that Figure 2a is a side view, and the detection system can be installed on an underwater vehicle. The electromagnetic sensor includes a set of electromagnetic transmitting antennas and a set of three-axis electromagnetic receiving antennas. The electromagnetic transmitting antenna is a rod-shaped spiral coil electromagnetic transmitting antenna, and its axis is parallel to the longitudinal axis of the detection system. The three-axis electromagnetic receiving antenna is composed of three spiral inductive coils with cores. The core material is manganese-zinc ferrite. The three coils correspond to the x, y, and z axes, respectively. The x-axis direction is the same as the longitudinal axis direction of the detection system, and the z-axis is the vertical direction.

[0035] Referring to Figure 2b , it is shown that Figure 2b is a top view. The two sets of acoustic transducer arrays are a low-frequency transducer and a set of medium-frequency transducer arrays. The low-frequency transducer is used to passively receive target radiated noise. The medium-frequency transducer array is a 1x20 linear transceiving array.

[0036] In addition, the electromagnetic signal in the preprocessing circuit of the acoustic-magnetic feature acquisition module 1 can be realized by an instrument amplification circuit based on the AD620 chip, and the acoustic signal can be realized by an operational amplifier circuit based on the OP07 chip. The data acquisition circuit can realize 24-channel analog signal acquisition function by using three AD7606 chips.

[0037] The feature data acquired by the acoustic-magnetic feature acquisition module 1 includes electromagnetic field intersection characteristic quantities: the electromagnetic field intensity S1 and phase information S2 of the z-axis received by the electromagnetic sensor, and the electromagnetic field intensity S3 of the x-axis; and acoustic field intersection characteristic data: the sound pressure amplitude S4 of the target noise and the target azimuth interval S5. Five-dimensional intersection feature data S={S1, S2,…S5} is formed.

[0038] The electromagnetic field intersection characteristic quantities are obtained by using a rod-shaped spiral coil electromagnetic transmitting antenna with its axis perpendicular to the longitudinal axis of the detection system, and using x, y, and z three-axis electromagnetic receiving coils. The x-axis direction is the same as the longitudinal axis direction of the detection system. The electromagnetic field intensity S1 and phase information S2 of the z-axis, and the electromagnetic field intensity S3 of the x-axis are extracted, respectively.

[0039] The acoustic field intersection characteristic data is obtained by using a low-frequency underwater acoustic receiving transducer to collect the target radiated noise during the intersection process, extracting the sound pressure amplitude S4, and using a medium-frequency active sonar to collect the target echo signal to calculate and generate the target azimuth interval S5.

[0040] In addition, the target acoustic reflection echo signal is used to calculate the target azimuth interval, which represents the specific azimuth of the target relative to the detection system during the rendezvous process. The 180° field of view of the detection system is divided into 37 intervals. For example, when the target appears in the 18th interval, it means that the target is directly above the detection system. Figure 4 This is a schematic diagram of the interval division for the target azimuth interval characteristics.

[0041] In this embodiment, the confidence optimization module 2 is used to iteratively solve for the optimal confidence value of each feature quantity. The confidence optimization module 2 obtains the optimal confidence value of each feature quantity by iteratively optimizing the intersection feature learning data according to the preset initial confidence value of each feature quantity. The intersection feature learning data refers to the electromagnetic and acoustic feature data collected by the detection system during the prior intersection process. The learning data and the initial confidence value of each feature quantity are both stored in the data storage module in advance. The data storage module is implemented using the FLASH chip S29GL01GP.

[0042] In this embodiment, the feature combination reliability calculation module 3 is used to solve the reliability value of each feature combination; the feature combination reliability calculation module 3 receives the output data of the confidence optimization module 2 and solves the reliability value of each feature combination, and the reliability value characterizes the reliability of the source of feature combination information.

[0043] The confidence optimization module 2 and the feature combination reliability calculation module 3 can be implemented using the same DSP chip, which is TMS320VC5509A.

[0044] In this embodiment, the support conversion module 4 is used to convert the collected data into regional identification support data of each feature quantity during the transaction process; the support conversion module 4 receives the output data of the acoustic-magnetic feature acquisition module 1 and converts the data into regional identification support data of each feature quantity during the transaction process. The regional identification support represents the contribution of the real-time data of a certain feature quantity to the accuracy of target region identification.

[0045] In this embodiment, the acoustic-magnetic feature data composite module 5 is used to solve the composite feature value of the acoustic-magnetic features; the acoustic-magnetic feature data composite module 5 receives the output data of the support conversion module 4, composites the acoustic-magnetic features, and solves the composite feature value.

[0046] In this embodiment, the feature combination confidence calculation module 6 is used to solve the confidence value of each feature combination; the feature combination confidence calculation module 6 receives the output data of the support conversion module 4, solves the confidence value of each feature combination, and the confidence value characterizes the degree of trustworthiness of the real-time data of the feature combination.

[0047] In this embodiment, the acoustic-magnetic feature fusion module 7 is used to fuse acoustic-magnetic intersection features in real time; the acoustic-magnetic feature fusion module 7 receives the output data from the feature combination reliability calculation module 3, the acoustic-magnetic feature data composite module 5 and the feature combination credibility calculation module 6, and completes the real-time fusion of acoustic-magnetic intersection features.

[0048] In this embodiment, the fusion feature comprehensive decision module 8 is used for real-time dynamic identification and judgment of the target area during the intersection process. The fusion feature comprehensive decision module 8 receives the output data of the acoustic-magnetic feature fusion module 7 and completes the real-time dynamic identification and judgment of the target area during the intersection process. The basis for identifying and judging the target area during the intersection process is that when the real-time data value output by the acoustic-magnetic feature fusion module 7 is greater than a preset value, it is determined that the system has reached the target area range.

[0049] Support conversion module 4, acoustic-magnetic feature data compositing module 5, feature combination credibility calculation module 6, acoustic-magnetic feature fusion module 7, and fusion feature comprehensive decision module 8 can be implemented using another DSP chip, which can be a TMS320VC5509A.

[0050] To better understand the working principle of the acoustic-magnetic feature fusion device for intersection region identification provided by this invention, its working process is described below:

[0051] Reference Figure 3a and Figure 3b As shown, the rendezvous distance between the detection system and the typical target is 5m. The rendezvous distance refers to the shortest distance between the detection system and the typical target during the rendezvous process. The rendezvous path is considered as a straight-line interval with the vertical projection point of the typical target on the rendezvous path as 0, the approach process being negative and the departure process being positive, and the rendezvous angle being 90°. The detection system travels at a speed of 5 knots. On the rendezvous path, the system collects signals and calculates feature quantities once every 2.5ms.

[0052] Reference Figure 5 As shown, the workflow is as follows:

[0053] Step 1: The device is powered on and started. The confidence optimization module loads 50 sets of electromagnetic and acoustic composite intersection feature learning data that are pre-stored in the data storage module. Each set of data has a length of 650, and the initial confidence values ​​for each feature are C = {C1, C2, ..., C5}.

[0054] Step 2: The confidence optimization module initiates the learning of 50 sets of electromagnetic and acoustic composite intersection feature learning data and iteratively optimizes the confidence of each feature. Upon completion, it outputs the optimal confidence C for each feature. best ={C b1 C b2 ,…C b5} and the reliability coefficient λ of the characteristic combination.

[0055] Step 3: The feature combination reliability calculation module calculates the reliability of all possible feature combinations based on the feature combination reliability coefficient λ, using the following formula:

[0056]

[0057] Where: {A i} is the set of all feature combinations; M({A i}) represents the reliability of all possible feature combinations; k is the reliability of feature combination A. i The number of feature quantities.

[0058] Step 4: The acoustic-magnetic feature fusion module begins to collect electromagnetic and acoustic feature data during the system intersection process, and outputs the data acquisition results S = {S1, S2, ... S5}.

[0059] Step 5: The support conversion module performs consistency preprocessing on the data S={S1,S2,…S5} output by the acoustomagnetic feature acquisition module, completes the support conversion, and outputs the support conversion result X={X1,X2,…X5}.

[0060] The formula for the consistency preprocessing operation is as follows:

[0061]

[0062] Among them, X i,j S represents the consistency processing result of the j-th data point of the i-th feature. i,j Let l represent the j-th data point of the i-th feature. j Let j represent the j-th intersection point.

[0063] Step Six: The acoustic-magnetic feature data composite module identifies the support X of the j-th (j=1,2,…m) target region. i,j (i = 1, 2, ..., n), calculate all feature combinations {A} under this composite feature data. i The composite value F j (A i ).

[0064] Calculate all feature combinations {A} i When dealing with the composite value of}, the formula is as follows:

[0065]

[0066] Step 7: The feature combination confidence calculation module calculates the support X of the target region at position j (j = 1, 2, ..., m). i,j (i = 1, 2, ..., n), calculate all feature combinations {A} under this composite feature data.i The credibility R of} j (A i ).

[0067] Calculate all feature combinations {A} i When determining the credibility of a number, the formula is as follows:

[0068]

[0069] Among them, X i,j and X l,j For feature combination A i The target region recognition support for any two features; a1 can be 0.4 and a2 can be 0.7.

[0070] Step 8: The acoustic-magnetic feature fusion module, based on feature combination A... i Credibility R j (A i ) and reliability M(A) i Solve for the fusion result W of the j-th (j=1,2,…m) n-dimensional intersection feature data of electromagnetic and acoustic composite detection. j (S).

[0071] Solving for the fusion result W of n-dimensional intersection feature data j When (S), the formula is as follows:

[0072] G j ({A i})=min(R j ({A i}),M({A i}))

[0073] W j (S)=F j (max(G j ({A i})) index )

[0074] Among them, G j ({A i}) is the feature combination {A i The fusion acceptance value of}, max(G j ({A i})) index The feature combination A corresponding to the maximum fusion acceptance value i The serial number.

[0075] Step Nine: The feature fusion and comprehensive decision-making module bases its decisions on the fusion result W. j (S) According to preset rules, the target area is dynamically identified and judged in real time during the intersection process.

[0076] The aforementioned "according to the preset rule" means that when the real-time data value output by the acoustic-magnetic feature fusion module is greater than the preset value, it is determined that the system has reached the target area; the preset value can generally be 0.85.

[0077] In summary, the acoustic-magnetic feature fusion device for intersection area identification provided by this invention combines electromagnetic and acoustic fields, leveraging the advantages of both physical fields in underwater detection, and fusing the intersection feature information of the two physical fields. This effectively improves the detection system's ability to identify underwater targets and increases the system's fault tolerance and adaptability.

[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An acoustic-magnetic feature fusion device for intersection region identification, characterized in that, It includes an acoustic-magnetic feature acquisition module (1), a confidence optimization module (2), a feature combination reliability calculation module (3), a support conversion module (4), an acoustic-magnetic feature data compositing module (5), a feature combination confidence calculation module (6), an acoustic-magnetic feature fusion module (7), and a fusion feature comprehensive decision module (8). The acoustic-magnetic feature acquisition module (1) is used to generate electromagnetic and acoustic detection signals and acquire characteristic data of electromagnetic and acoustic fields during the intersection process; The confidence optimization module (2) is used to iteratively solve for the optimal confidence value of each feature quantity; The feature combination reliability calculation module (3) is used to solve the reliability value of each feature combination; The support conversion module (4) is used to convert the collected data into regional identification support data of each feature quantity during the transaction process; The acoustic-magnetic feature data composite module (5) is used to solve the composite feature value of the acoustic-magnetic features; The feature combination confidence calculation module (6) is used to solve the confidence value of each feature combination; The acoustic-magnetic feature fusion module (7) is used to fuse acoustic-magnetic intersection features in real time; The integrated feature decision module (8) is used for real-time dynamic identification and judgment of the target area during the intersection process; The acoustic-magnetic feature acquisition module (1) acquires feature data including electromagnetic field intersection features: the electromagnetic field strength S1 and phase information S2 of the z-axis received by the electromagnetic sensor, and the electromagnetic field strength S3 of the x-axis; it also includes sound field intersection features: the sound pressure amplitude of the target noise S4 and the target azimuth interval S5, forming 5-dimensional intersection feature data S={S1, S2, ...S5}; The confidence optimization module (2) initiates the learning of electromagnetic and acoustic composite intersection feature learning data and iterative optimization of the confidence of each feature quantity. After completion, it outputs the optimal confidence C of each feature quantity. best ={C1, C2, ..., C5} and the reliability coefficient of the characteristic combination ; The feature combination reliability calculation module (3) calculates the reliability coefficient based on the feature combination. The reliability of all possible combinations of features is calculated using the following formula: ； in, The set of all feature combinations. For the reliability of all possible combinations of features, For feature combination The number of feature types; The support conversion module (4) performs consistency preprocessing on the data S={S1, S2, ... S5} output by the acoustic-magnetic feature acquisition module (1), completes the support conversion, and outputs the support conversion result X={X1, X2, ... X5}. The formula for the consistency preprocessing operation is as follows: ； in, For the first The first dimension of the feature Consistency processing results for each data point Indicates the first The first dimension of the feature Data points, Indicates the first There are 5 intersection points, where i = 1, 2, ..., 5, j = 1, 2, ..., m; The acoustic-magnetic feature data composite module (5) is composed of the first Support for target region recognition Calculate all feature combinations under this composite feature data. composite value ; Calculate all feature combinations When dealing with composite values, the formula is as follows: ； The feature combination credibility calculation module (6) is composed of the first Support for target region recognition Calculate all feature combinations under this composite feature data. Credibility ; Calculate all feature combinations When determining the credibility, the formula is as follows: ； ； in, and For feature combination Support for target region identification based on any two feature values; Take 0.4, Take 0.

7.

2. The acoustic-magnetic feature fusion device for intersection region identification according to claim 1, characterized in that, The acoustic-magnetic feature acquisition module (1) includes an acoustic-magnetic sensor module, a preprocessing circuit, and a data acquisition circuit. The acoustic-magnetic sensor module includes a set of electromagnetic transceiver sensors and two sets of acoustic transducer arrays.

3. The acoustic-magnetic feature fusion device for intersection region identification according to claim 2, characterized in that, The confidence optimization module (2) obtains the optimal confidence value of each feature by iteratively optimizing the intersection feature learning data according to the preset initial confidence values ​​of each feature. The intersection feature learning data refers to the electromagnetic and acoustic feature data collected by the detection system during the prior intersection process.

4. The acoustic-magnetic feature fusion device for intersection region identification according to claim 3, characterized in that, The feature combination reliability calculation module (3) receives the output data of the confidence optimization module (2) and solves the reliability value of each feature combination. The reliability value characterizes the reliability of the source of feature combination information.

5. The acoustic-magnetic feature fusion device for intersection region identification according to claim 4, characterized in that, The support conversion module (4) receives the output data from the acoustic-magnetic feature acquisition module (1) and converts the data into regional identification support data for each feature quantity during the transaction process. The regional identification support represents the contribution of the real-time data of a certain feature quantity to the accuracy of target region identification.

6. The acoustic-magnetic feature fusion device for intersection region identification according to claim 5, characterized in that, The acoustic-magnetic feature data composite module (5) receives the output data from the support conversion module (4), composites the acoustic-magnetic features, and solves the composite feature value.

7. The acoustic-magnetic feature fusion device for intersection region identification according to claim 6, characterized in that, The feature combination credibility calculation module (6) receives the output data from the support conversion module (4) and solves the credibility value of each feature combination. The credibility value represents the degree of credibility of the real-time data of the feature combination.

8. The acoustic-magnetic feature fusion device for intersection region identification according to claim 7, characterized in that, The acoustic-magnetic feature fusion module (7) receives the output data from the feature combination reliability calculation module (3), the acoustic-magnetic feature data composite module (5), and the feature combination credibility calculation module (6), and completes the real-time fusion of acoustic-magnetic intersection features.

9. The acoustic-magnetic feature fusion device for intersection region identification according to claim 8, characterized in that, The integrated decision-making module (8) receives the output data from the acoustic-magnetic feature fusion module (7) and completes the real-time dynamic identification and judgment of the target area during the intersection process; The basis for identifying and judging the target area during the intersection process is that when the real-time data value output by the acoustic-magnetic feature fusion module (7) is greater than the preset value, it is determined that the system has reached the target area.

Images