A Deep-Sea Surface and Underwater Target Identification Method Based on Information Entropy

Abstract

This invention relates to the field of target depth estimation, providing a method for identifying deep-sea surface and underwater targets at extremely close range based on information entropy. The method comprises the following steps: S1, acquiring sound source data at extremely close range in the deep sea; S2, constructing a virtual source model, calculating the measurement angle, and then obtaining the sound pressure cross-spectrum matrix; S3, calculating the sound source distance-azimuth ambiguity function; S4, calculating the sound source distance-azimuth distribution curve cluster; S5, specifying the scanning range of the sound source depth and configuring a copy field; S6, calculating the complex sound pressure received by the array under the copy field, and performing sound pressure signal matching field processing to obtain the sound source depth-distance ambiguity function; S7, finding the column corresponding to the maximum value of the sound source depth-distance ambiguity function, calculating the information entropy of that column, and determining whether the target is a surface or underwater target based on the magnitude of the information entropy. This invention can determine whether the target is a surface or underwater target, solving the ambiguity problem in deep-sea target depth estimation at extremely close range.

Description

Technical Field

[0001] This invention is applied to the field of depth estimation for deep-sea targets at extremely close range, specifically involving a method for identifying deep-sea surface and underwater targets at extremely close range based on information entropy. Background Technology

[0002] Current technologies have yielded some results in target ranging and depth determination using horizontal arrays in deep-sea environments. For example, CN115825965A proposes a three-dimensional target localization method based on a dual-horizontal array on the deep seabed for direct sound zone positioning. First, the target azimuth is estimated through beamforming of the dual-horizontal array. Then, the target distance is estimated by tracking the maximum output power of the dual-horizontal array beams and the tracking azimuth angle, combined with the positions of the two arrays. Finally, the estimated target distance, the preset sound source depth, and marine environmental parameters measured in actual sea trials are input into a ray model to obtain the time delay difference between the direct path and the first reflection path from the sea surface. A copied sound pressure field is constructed, and a depth estimation fuzzy function for the dual-horizontal array is established, the maximum value of which is the target depth. However, since CN115825965A directly uses beamforming to estimate the sound source azimuth, for extremely close-range sound sources, the sound wave does not satisfy the plane wave assumption. In this case, the selection of the reference sound velocity for beamforming may lead to a large direction-finding error, thus affecting the subsequent estimation of the sound source depth. Furthermore, since this method considers pulse sound sources, while actual signals are generally continuous, it is inconvenient to obtain the arrival delay of the direct wave in this case.

[0003] Simulations of existing technologies revealed that when a sound source approaches the water surface, the estimation of the sound source depth is interfered with by high sidelobes. Summary of the Invention

[0004] To address the issue of ambiguity in target depth estimation at extremely close range in the deep sea, this invention proposes a method for identifying surface and underwater targets at extremely close range in the deep sea based on information entropy. First, by matching the angle of arrival (AHA) and beamforming measurement angle, and combining this with translation processing, the distance-azimuth distribution curve and curve cluster of the sound source are obtained. Then, based on the array-received sound pressure signal, a matching field is performed on the depth and distance-azimuth dimensions to obtain the ambiguity function matrix. Building upon this, this invention designs a statistic called information entropy. This statistic is calculated by extracting the column corresponding to the position with the highest intensity in the ambiguity function and designing a unique calculation method to calculate the entropy value of this column. An information entropy threshold is then given based on the channel. Finally, the magnitude of the information entropy is used to identify surface and underwater targets. This invention only requires a single horizontal array and considers line spectrum signals, overcoming the limitation of previous methods that only considered transient signals. Furthermore, this method delves into how to roughly estimate target depth at extremely close range in the deep sea using a horizontal array, achieving a rough estimation of the target depth range at extremely close range.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for identifying deep-sea underwater targets at extremely close range based on information entropy includes the following steps:

[0007] S1, using N The horizontal receiving array receives sound source data from extremely close locations in the deep sea;

[0008] S2. Set up the measurement field and calculate the array-received sound pressure vector using the virtual source method model. The measurement angle is obtained through beamforming method. This leads to the sound pressure cross-spectral matrix. ;

[0009] S3. Assuming a sound source depth, calculate the angle of arrival (AHA) for different sound source distances and orientations, and perform the measured angle. By matching the angle of arrival with the field, the source distance-azimuth ambiguity function is obtained;

[0010] S4. Extract the source distance-location ambiguity function from the parts with intensity greater than […]. b The points are fitted and translated to obtain a cluster of sound source distance-azimuth distribution curves;

[0011] S5. Let the distance and orientation of the sound source be distributed on the cluster of sound source distance-orientation distribution curves, give the scanning range of the sound source depth and configure the copy field;

[0012] S6. Calculate the sound pressure received by the array under the copy field, and perform sound pressure signal matching field processing to obtain the sound source depth-distance ambiguity function;

[0013] S7. Find the column corresponding to the maximum value of the sound source depth-distance ambiguity function, calculate the information entropy of the column, and determine whether the target is a surface target or an underwater target based on the magnitude of the information entropy.

[0014] Furthermore, the sound field of the sound source data in step S1 is composed of the direct wave of the sound wave and the reflected wave from the sea surface.

[0015] Furthermore, step S2 involves calculating the array-received sound pressure vector using the virtual source method model. Specifically, it includes the following steps:

[0016] S201, Obtain the location at horizontal distance through the virtual source method model. r ,depth z The sound pressure level, expressed as follows:

[0017] ,

[0018] The time factor is ignored in the formula. The negative sign is to satisfy the boundary condition for sea surface pressure release: , For the medium wavenumber, c Take the speed of sound at the sound source. f Let r be the frequency of the sound source, and z be the horizontal distance and depth from the observed point to the sound source, respectively.

[0019] S202. Based on the geometric relationships of the virtual source method model:

[0020] , ,

[0021] when hour, and The glancing angle is approximately equal to ,but:

[0022] , ,

[0023] in, Indicates much greater than, It means approximately equal to.

[0024] S203, Assuming the denominator in the sound pressure expression is... and can be R Instead, the exponential function is expanded into a trigonometric function, and the sound pressure... Simplified to:

[0025] ;

[0026] S204. If this sound pressure expression is applied to the entire sea area, a virtual source sound field can be obtained.

[0027] S205. Obtain the sound pressure of the nth array element based on the virtual source method sound field:

[0028] ,

[0029] S206, The array-received sound pressure vector can be expressed as:

[0030] .

[0031] Furthermore, the measurement angle is obtained through beamforming method. Specifically, the steps include the following:

[0032] S211, Expression of array received signal:

[0033] ,

[0034] in, The spectrum of the sound source signal;

[0035] S212, the acoustic pressure cross-spectral density matrix of the array The expression is:

[0036] ,

[0037] S213. Express the output power of the beamformer in the frequency domain:

[0038] ,

[0039] in, It is the beamforming weight vector;

[0040] S214, Calculate The angle corresponding to the maximum value is .

[0041] Furthermore, step S3 specifically includes the following steps:

[0042] S301, The formula for calculating the angle of arrival is:

[0043] ,

[0044] in, , These are the horizontal distances to the corresponding sound sources. r and depth z The pitch and sweep angles of the lower sound waves;

[0045] S302. Based on the virtual source method model, the formula for calculating the grazing angle is obtained:

[0046] ;

[0047] S303. Assume the sound source depth used to calculate the angle of arrival is 0 to... Let any value in the distance be... r and direction From 0 to The values ​​range from -90° to 90°. Based on the formulas for angle of arrival and glancing angle, the angle of arrival of sound sources at these distances and azimuths on the array can be obtained.

[0048] S304, Sound Source Distance-Directional Ambiguity Function The calculation formula is as follows:

[0049] .

[0050] Furthermore, the method for obtaining the sound source distance-azimuth distribution curve cluster includes the following steps:

[0051] In the sound source distance-location ambiguity function, the intensity greater than b The points are extracted and fitted to obtain the sound source distance-azimuth distribution curve. ;

[0052] The sound source distance-azimuth distribution curve After translation processing, a cluster of sound source distance-azimuth distribution curves was obtained:

[0053] ,

[0054] in, This represents the translation amount.

[0055] Furthermore, the method for obtaining the sound source depth-distance ambiguity function in step S6 includes the following steps:

[0056] S601, The sound source depth range is obtained from 0 to [the specified range] using the virtual source method model. Distance / direction follows Distributed array receives sound pressure ;

[0057] S602. The expression for the depth-distance or depth-azimuth ambiguity function based on the minimum variance processor is as follows:

[0058] ;

[0059] S603, after normalization and logarithmic transformation, yields:

[0060] ;

[0061] S604, The x-coordinate corresponding to the point of maximum intensity is the estimated distance to the sound source.

[0062] Furthermore, the calculation method for information entropy in step S7 includes:

[0063] Take out the S603 described The column vector corresponding to the point of maximum intensity in the medium-depth channel, and the minimum value of the given vector elements based on the deep-sea channel. a Let those less than or equal to a The value of the point a The formula for calculating information entropy is as follows:

[0064] ,

[0065] The total number of elements in the vector is I , No. i The probability of each element is ;

[0066] The information entropy will be used to identify targets on the water surface and underwater.

[0067] Furthermore, based on the variation patterns of information entropy when the target is located at different orientations, distances, and depths, we propose to provide depth-distance two-dimensional pseudo-color maps and depth-orientation two-dimensional pseudo-color maps related to information entropy.

[0068] And based on the actual channel, the threshold of information entropy for identifying surface and underwater targets at different orientations and distances is given;

[0069] When the information entropy is greater than the threshold, it is considered a surface target; otherwise, it is considered an underwater target.

[0070] Compared with the prior art, the beneficial effects of the present invention are:

[0071] This invention proposes a method for identifying deep-sea surface and underwater targets at extremely close range based on information entropy. By assuming a sound source depth and combining it with the formula for calculating the grazing angle, the angle of arrival (AHA) of the sound source at different distances and azimuths of the array position is obtained. This AHA is then matched with the beamforming measurement angle to generate a cluster of sound source distance-azimuth distribution curves. A copy field is then designed using the depth scanning range and this cluster of curves, and a matching field processing based on the sound pressure signal is performed to obtain an ambiguity function. Finally, the column corresponding to the maximum value of the ambiguity function is extracted, and the information entropy is calculated according to the information entropy calculation process presented in this paper. The magnitude of this value is used to identify surface and underwater targets, thus solving the problem of ambiguity in depth estimation for deep-sea targets at extremely close range.

[0072] The basic principles and implementation schemes of this invention have been verified by computer numerical simulation. The results show that the deep-sea surface and underwater target identification method based on information entropy proposed in this invention can determine whether the target is a surface target or an underwater target. This method addresses the problem that horizontal arrays are difficult to estimate the depth of close-range targets in deep-sea environments and provides a method for roughly estimating the target depth range.

[0073] To more clearly illustrate the structural features and effects of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. Attached Figure Description

[0074] Figure 1 is a flowchart illustrating the deep-sea near-range underwater target identification method based on information entropy.

[0075] Figure 2 is a schematic diagram of a uniform linear array horizontally arranged near the seabed;

[0076] Figure 3. Deep-sea sound velocity profile;

[0077] Figure 4. Deep-sea propagation loss diagram (sound source depth 50m, frequency 100Hz).

[0078] Figure 5. Geometric schematic diagram of the virtual source method model;

[0079] Figure 6. Sound field propagation loss diagram of virtual source method (sound source depth 50m, frequency 100Hz).

[0080] Figure 7. Sound source distance-azimuth distribution curves and curve clusters (circles indicate the actual location of the sound source). Figure 7 Figure a shows the distance-azimuth distribution curve of the sound source before translation; Figure 7 Figure b shows the cluster of sound source distance-azimuth distribution curves after translation.

[0081] Figure 8. Matching field output based on sound pressure signal at different sound source depths ("○" represents the true position, "×" represents the estimated position, and the solid box represents the column vector required to calculate the information entropy; Figure 8 Figures a, b, c, and d represent the matched field outputs at sound source depths of 5m, 20m, 100m, and 200m, respectively.

[0082] Figure 9. Relationship between information entropy and target distance, depth, and orientation ( Figure 9 Figure a shows the information entropy depth-distance two-dimensional map, and Figure b shows the information entropy depth-direction two-dimensional map. Detailed Implementation

[0083] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0084] The specific implementation of the present invention will be described in detail below with reference to specific embodiments. Specific Implementation Example 1

[0086] Please see the appendix Figure 1 This invention provides a method for identifying underwater targets at extremely close range in deep sea based on information entropy, characterized by the following steps:

[0087] S1, using N The horizontal receiving array receives sound source data from extremely close locations in the deep sea;

[0088] S2. Set up the measurement field and calculate the array-received sound pressure vector using the virtual source method model. The measurement angle is obtained through beamforming method. This leads to the sound pressure cross-spectral matrix. ;

[0089] S3. Assuming a sound source depth, calculate the angle of arrival (AHA) for different sound source distances and orientations, and perform the measured angle. By matching the angle of arrival with the field, the source distance-azimuth ambiguity function is obtained;

[0090] S4. Extract the source distance-location ambiguity function from the parts with intensity greater than […]. b The points are fitted and translated to obtain a cluster of sound source distance-azimuth distribution curves;

[0091] S5. Let the distance and orientation of the sound source be distributed on the cluster of sound source distance-orientation distribution curves, give the scanning range of the sound source depth and configure the copy field;

[0092] S6. Calculate the sound pressure received by the array under the copy field, and perform sound pressure signal matching field processing to obtain the sound source depth-distance ambiguity function;

[0093] S7. Find the column corresponding to the maximum value of the sound source depth-distance ambiguity function, calculate the information entropy of the column, and determine whether the target is a surface target or an underwater target based on the magnitude of the information entropy.

[0094] This invention requires only a single horizontal array and considers line spectrum signals, overcoming the limitation of previous methods that only considered transient signals. Furthermore, this method delves into how to roughly estimate target depth at extremely close range in the deep sea using a horizontal array, achieving a rough estimation of target depth ranges at very close distances. Specific Implementation Example 2

[0096] See attached document Figure 1-9 As shown, a method for identifying underwater targets at extremely close range in deep sea based on information entropy includes the following steps:

[0097] Step 1: Calculate the sound field

[0098] Assuming the true depth of the sound source is The actual distance is The true location is A narrowband signal is emitted, and an N-element horizontal uniform linear array is set up at an extremely close distance in the deep sea. Since the considered environment is extremely close, the sound field mainly consists of the direct sound wave and the reflected wave from the sea surface. The deep-sea propagation loss diagram obtained using KRAKENC is attached. Figure 4 As shown, the propagation loss near the sound source exhibits a directional pattern with alternating maxima and minima, a typical deep-sea Lloyd mirror interference pattern. Each minimum propagation loss corresponds to a stripe-like pattern called a beam. This sound field can be explained using the virtual source method, which treats the sea surface as a mirror reflector. A geometrical schematic of the virtual source method model is attached. Figure 5 As shown.

[0099] The virtual source method model is used to obtain the horizontal distance. r ,depth z Sound pressure level:

[0100] ,

[0101] The time factor is ignored in the formula. The negative sign is to satisfy the boundary conditions for sea surface pressure release; , Let c be the wavenumber of the medium, c be the speed of sound at the sound source, and f be the frequency of the sound source. According to the appendix... Figure 5 Geometric relationships in:

[0102] , ;

[0103] when hour, and The glancing angle is approximately equal to ,but:

[0104] ;

[0105] ;

[0106] in, Indicates much greater than, This indicates approximately equal to. Further, assume that the denominator in the sound pressure expression... and It can be replaced by R, and the exponential function can be expanded into a trigonometric function, sound pressure. Simplified to:

[0107] ;

[0108] If this sound pressure expression is applied to the entire sea area, a new sound field can be obtained, which is the sound field of the virtual source method. Similarly, the propagation loss of the virtual source method sound field is shown in Figure 6. (From Figures 4 and 6...) Figure 6 It is easy to see that at extremely close distances, the sound field of the virtual source method and the sound field of KRAKENC are almost identical. From the perspective of simulation, this shows that it is reasonable to use the virtual source method to explain the sound field in this region. Further theoretical derivation will be carried out based on the virtual source method model.

[0109] Based on this, the sound pressure of the nth array element can be obtained:

[0110] ,

[0111] The array-received sound pressure vector can be expressed as:

[0112] .

[0113] Step 2: Calculate the measured angle

[0114] Taking a noise-free environment as an example, the array receiving signal can be written as:

[0115] ,

[0116] in, Given the sound source signal spectrum, the acoustic pressure cross-spectral density matrix of the array can be obtained as follows:

[0117] ,

[0118] The output power of the beamformer in the frequency domain is:

[0119] ,

[0120] in, This is the beamforming weight vector. For example, the weight vector for conventional beamforming is:

[0121] ,

[0122] in, Let d be the reference wavelength for beamforming and d be the element spacing; calculate The angle corresponding to the maximum value is .

[0123] Step 3: Calculate the angle of arrival

[0124] The formula for calculating the angle of arrival is given first:

[0125] ,

[0126] in, , These are the horizontal distances to the corresponding sound sources. r and depth z The pitch and sweep angles of the lower sound waves;

[0127] Taking the grazing angle as an example, based on the virtual source method model shown in Figure 5, the formula for calculating the grazing angle is obtained:

[0128] ,

[0129] Assume the sound source depth used to calculate the angle of arrival is 0 to... Let the distance r and the orientation be any value in the equation. From 0 to Taking values ​​from -90° to 90°, the angle of arrival (AHA) of the sound sources at these ranges and azimuths can be obtained on the array using the formulas for AHA and grazing angle; finally, the sound source range-azimuth ambiguity function is obtained. :

[0130] .

[0131] Step 4: Put Points with a medium intensity greater than b are extracted, and a fitting process is performed on these points. This method yields the sound source distance-azimuth distribution curve. ;

[0132] In addition, to reduce errors, After translation processing, a cluster of sound source distance-azimuth distribution curves was obtained:

[0133] ,

[0134] in, This represents the translation amount.

[0135] Step 5: Perform matched field processing based on sound pressure signals. The sound source depth range is obtained using the virtual source method model, from 0 to... Distance / direction follows Distributed array receives sound pressure The following are the depth-distance or depth-azimuth ambiguity functions based on the minimum variance processor:

[0136] ,

[0137] After normalization and logarithmic transformation, we obtain:

[0138] ,

[0139] The x-coordinate corresponding to the point of maximum intensity is the estimated distance to the sound source.

[0140] Step Six: Calculate Information Entropy

[0141] take out The column vector corresponding to the point of maximum intensity is given, and based on the minimum value 'a' of the vector elements given by the deep-sea channel, the value of the points less than or equal to 'a' is set to 'a'. Finally, the entropy calculation formula is used:

[0142] ,

[0143] Where the total number of elements in the vector is I, and the probability of the i-th element is... This statistic will be used to identify surface and underwater targets.

[0144] Finally, the variation patterns of information entropy for targets at different orientations, distances, and depths are presented, with a particular focus on its relationship with the depth of the sound source. Two-dimensional pseudo-color maps of information entropy based on depth-distance and depth-orientation are proposed, and thresholds for information entropy at different orientations and distances for identifying surface and underwater targets are given based on actual channel conditions. When the information entropy is greater than the threshold, the target is considered a surface target; otherwise, it is considered an underwater target.

[0145] Simulation Example 2

[0146] To further verify the target recognition method of the present invention, the present invention provides simulation embodiments, as shown in the appendix. Figure 1-9 As shown in Tables 1-3, the deep-sea underwater target identification method based on information entropy at extremely close range is mainly used to determine the approximate depth range of a sound source at extremely close range in the deep sea. Specifically, it includes the following steps:

[0147] The first step involves simulating the following standard environmental parameters: sound source frequency 100Hz, sound source depth range 0 to 200m, sound source distance range 0 to 2km, and sound source azimuth angle relative to the array range -70° to -40°. The uniform linear array has an element spacing of 7.5m, 35 elements, an array depth of 4950m, a signal-to-noise ratio of 30dB, a seabed P-wave velocity of 1650m / s, and a seabed density of 1.6. The attenuation of P-waves on the seabed is 0.25. Translation amount The deep-sea sound velocity profile is shown in Figure 3. First, the measurement field is set up according to standard environmental parameters, and the array-received sound pressure vector is calculated using the virtual source method model. The measurement angle is obtained through beamforming method. Finally, the sound pressure cross-spectral matrix was obtained. .

[0148] The second step involves assuming the sound source depth is any value between 0 and 300m, using 100m as the distance scanning interval and 1° as the azimuth scanning interval, calculating the angle of arrival (ADR) for different sound source distances and azimuths between 0km and 10km and between -90° and 90°. The sound source distance-directional ambiguity function is obtained by the following formula:

[0149] .

[0150] Step 3: Extraction For points with a medium intensity greater than 0, a fitting process is performed on these points, with a fitting order of 20. This yields the sound source distance-azimuth distribution curve. Then, a series of sound source distance-azimuth distribution curves are obtained by translation processing. .

[0151] Step 4: Establish a copy field to complete the matched field processing based on the sound pressure signal. The distance and orientation of the sound sources in the copy field are distributed as follows: Above, the sound source depth scanning range is 20m, with an interval of 1m, to obtain the received sound pressure of each array element. The matching field employs a minimum variance processor, and the depth-distance ambiguity function is obtained using the following formula. :

[0152] ,

[0153] ,

[0154] Finally found The column corresponding to the maximum value is determined, and using the information entropy calculation process given in this paper, the minimum value of the vector elements is set to -10 according to the deep-sea channel. All elements in the column that are less than the minimum value are made equal to the minimum value. Finally, the information entropy is calculated by combining the entropy calculation formula.

[0155] Step 5: The variation patterns of information entropy when the target is located at different orientations, distances, and depths are discussed, and depth-distance two-dimensional pseudo-color maps and depth-orientation two-dimensional pseudo-color maps of information entropy are presented. Finally, the information entropy threshold for identifying surface and underwater targets using this method in a corresponding simulation environment is given. Relevant simulation results are shown in Figure 9.

[0156] Simulation results show that information entropy is primarily affected by the depth of the sound source. Using a sound source depth of 20m as a dividing line, the information entropy decreases significantly as the sound source depth continues to increase. This property will be used to identify surface and underwater targets.

[0157] Based on the simulation results, in this simulation environment, when the information entropy is not less than 3, the target is a surface target with a depth of less than 20m, otherwise it is an underwater target.

[0158] Step Six: First, the target's orientation was fixed, and simulation analysis was conducted to examine the target identification performance of this method at different distances. Then, the target distance was fixed, and simulation analysis was conducted to examine the target identification performance of this method at different orientations. This verifies the effectiveness of this method in identifying underwater targets at extremely close range.

[0159] Table 1. Symbols related to this article

[0160]

[0161]

[0162]

[0163] Table 2. Information entropy (non-bold) and identification results (bold) of targets at different distances from a fixed orientation (-60°).

[0164]

[0165] Table 3. Information entropy (non-bold) and identification results (bold) of targets at different azimuths at a fixed distance (1.5km).

[0166]

[0167] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for identifying underwater targets at extremely close range in deep sea based on information entropy, characterized in that, Includes the following steps: S1. Receive sound source data from extremely close-range locations in the deep sea using an N-element horizontal receiving array; S2. Set up the measurement field and calculate the array-received sound pressure vector using the virtual source method model. The measurement angle is obtained through beamforming method. This leads to the sound pressure cross-spectral matrix. ; S3. Assuming a sound source depth, calculate the angle of arrival (AHA) for different sound source distances and orientations, and perform the measured angle. The source distance-azimuth ambiguity function is obtained by matching the angle of arrival with the field processing, specifically as follows: S301, the formula for calculating the angle of arrival is: ， in, , These are the horizontal distances to the corresponding sound sources. r and depth z The pitch and sweep angles of the lower sound waves; S302. Based on the virtual source method model, the formula for calculating the grazing angle is obtained: ； S303. Assume the sound source depth used to calculate the angle of arrival is 0 to... Let any value in the range be the horizontal distance. r and direction From 0 to The values ​​range from -90° to 90°. Based on the formulas for angle of arrival and glancing angle, the angle of arrival of sound sources at these distances and azimuths on the array can be obtained. S304, Sound Source Distance-Directional Ambiguity Function The calculation formula is as follows: ； S4. Extract the points with intensity greater than b in the sound source distance-azimuth ambiguity function, and perform fitting and translation processing on these points to obtain a cluster of sound source distance-azimuth distribution curves; S5. Let the distance and orientation of the sound source be distributed on the cluster of sound source distance-orientation distribution curves, give the scanning range of the sound source depth and configure the copy field; S6. Calculate the sound pressure received by the array under the copy field, and perform sound pressure signal matching field processing to obtain the sound source depth-distance ambiguity function; S7. Find the column corresponding to the maximum value of the sound source depth-distance ambiguity function, calculate the information entropy of the column, and determine whether the target is a surface target or an underwater target based on the magnitude of the information entropy.

2. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 1, characterized in that, The sound field of the sound source data in step S1 consists of the direct wave of the sound wave and the reflected wave from the sea surface.

3. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 2, characterized in that, Step S2 involves calculating the array-received sound pressure vector using the virtual source method model. Specifically, it includes the following steps: S201, Obtain the location at horizontal distance through the virtual source method model. r ,depth z The sound pressure level, expressed as follows: ， The time factor is ignored in the formula. The negative sign is to satisfy the boundary condition for sea surface pressure release: , Let f be the medium wavenumber, c be the sound velocity at the sound source, f be the sound source frequency, and r and z represent the horizontal distance and depth from the observed point to the sound source, respectively. S202. Based on the geometric relationships of the virtual source method model: ， ， when hour, and The glancing angle is approximately equal to ,but: ， ； in, Indicates much greater than, It means approximately equal to; S203, Assuming the denominator in the sound pressure expression is... and It can be replaced by R, and the exponential function can be expanded into a trigonometric function, sound pressure. Simplified to: ； S204. If this sound pressure expression is applied to the entire sea area, a virtual source sound field can be obtained. S205. Obtain the sound pressure of the nth array element based on the virtual source method sound field: ， S206, The array-received sound pressure vector can be expressed as: 。 4. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 1, characterized in that, The measurement angle is obtained through beamforming method. Specifically, the steps include the following: S211, Expression of array received signal: ， in, The spectrum of the sound source signal; S212, the acoustic pressure cross-spectral density matrix of the array The expression is: ， S213. Express the output power of the beamformer in the frequency domain: ， in, It is the beamforming weight vector; S214, Calculate The angle corresponding to the maximum value is .

5. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 4, characterized in that, The method for obtaining the sound source distance-azimuth distribution curve cluster includes the following steps: Points with intensity greater than b in the sound source distance-azimuth ambiguity function are extracted, and these points are fitted to obtain the sound source distance-azimuth distribution curve. ; The sound source distance-azimuth distribution curve After translation processing, a cluster of sound source distance-azimuth distribution curves was obtained: ， in, This represents the translation amount.

6. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 5, characterized in that, The method for obtaining the sound source depth-distance ambiguity function in step S6 includes the following steps: S601, The sound source depth range is obtained from 0 to [the specified range] using the virtual source method model. Distance / direction follows Distributed array receives sound pressure ; S602. The expression for the depth-distance or depth-azimuth ambiguity function based on the minimum variance processor is as follows: ； S603, after normalization and logarithmic transformation, yields: ； S604, The x-coordinate corresponding to the point of maximum intensity is the estimated distance to the sound source.

7. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 6, characterized in that, The method for calculating information entropy in step S7 includes: Take out the S603 described The column vector corresponding to the point of maximum intensity is given. Based on the minimum value 'a' of the vector elements given by the deep-sea channel, the value of each point less than or equal to 'a' is set to 'a'. The formula for calculating information entropy is as follows: ， Where the total number of elements in the vector is I, and the probability of the i-th element is... ; The information entropy will be used to identify targets on the water surface and underwater.

8. The method for identifying deep-sea underwater targets at extremely close range based on information entropy according to claim 7, characterized in that, The identification of surface and underwater targets includes the following steps: Based on the variation of information entropy when the target is located at different orientations, distances, and depths, we propose to provide depth-distance two-dimensional pseudo-color maps and depth-orientation two-dimensional pseudo-color maps related to information entropy. And based on the actual channel, the threshold of information entropy for identifying surface and underwater targets at different orientations and distances is given; When the information entropy is greater than the threshold, it is considered a surface target; otherwise, it is considered an underwater target.

Images