阅读说明：本技术 一种侧扫声呐海底线自动提取方法及装置 (Automatic extraction method and device for side-scan sonar sea bottom line ) 是由 于化鹏 沈同圣 李晶 李子圆 韩超 于 2020-04-07 设计创作，主要内容包括：本发明公开了一种侧扫声呐海底线自动提取方法及装置。该方法包括：获取回波数据的盲区边界；根据所述盲区的宽度及左、右舷的回波数据最大采样数,提取所有Ping扫描线的左、右舷回波数据,并消除左、右舷回波数据的异常点；利用消除异常点的左右舷回波数据检测海底点坐标；以及根据海底点坐标提取海底线。本发明在不需要人工干预设定海底线初值的条件下,能够克服较强噪声干扰,海底线提取准确率高,细节连续性好。(The invention discloses a method and a device for automatically extracting a side-scan sonar sea bottom line. The method comprises the following steps: acquiring a blind area boundary of echo data; extracting the port and starboard echo data of all Ping scanning lines according to the width of the blind zone and the maximum sampling number of the port and starboard echo data, and eliminating abnormal points of the port and starboard echo data; detecting the coordinates of the seabed points by using the port and starboard echo data for eliminating the abnormal points; and extracting a seabed line according to the seabed point coordinates. The method can overcome strong noise interference without manual intervention to set the initial value of the bottom line, and has high accuracy rate of the bottom line extraction and good detail continuity.)

1. A side scan sonar sea bottom line extraction method is characterized by comprising the following steps:

acquiring a blind area boundary of echo data;

extracting the port and starboard echo data of all Ping scanning lines according to the width of the blind zone and the maximum sampling number of the port and starboard echo data, and eliminating abnormal points of the port and starboard echo data;

detecting the coordinates of the seabed points by using the port and starboard echo data for eliminating the abnormal points; and

and extracting a seabed line according to the seabed point coordinates.

2. The side-scan sonar undersea line extraction method according to claim 1, wherein the acquiring of the dead zone boundary of the echo data includes: and respectively detecting the first echo data of each Ping scanning line on the port and the starboard, which is smaller than the blind zone boundary detection threshold value, and determining the blind zone width of the echo data according to the bilateral symmetry principle.

3. The side-scan sonar sea bottom line extraction method according to claim 1, wherein outliers of the port and starboard echo data are eliminated using median filtering.

4. The side-scan sonar sea bottom line extraction method according to claim 3, wherein the removing of the outliers of the port and starboard echo data by median filtering includes:

calculating the signal intensity median of each echo data of each Ping scanning line and a plurality of adjacent echo data of the front and rear adjacent Ping scanning lines in the port and starboard echo data;

calculating the absolute deviation of the signal intensity of each echo data of each Ping scanning line, and determining the signal intensity value of each echo data according to the absolute deviation of the signal intensity; and

and according to the relation between the absolute deviation of the signal intensity and a set value, taking the signal intensity median value or the signal intensity sampling value as the signal intensity numerical value of the echo data of each line.

5. The side-scan sonar sea bottom line extraction method according to claim 1, wherein the detecting sea bottom point coordinates using port and starboard echo data from which outliers are eliminated includes:

matching and aligning two space-time data sequences of the port and starboard echo data of each Ping scanning line;

taking logarithm of the matched and aligned port and starboard echo data;

calculating the standard deviation of the logarithmic sequence of the port and starboard echo data of each Ping scanning line, and constructing a comprehensive sequence of the port and starboard echo data according to the standard deviation;

carrying out maximum detection on the comprehensive sequence of the port and starboard echo data to obtain an extreme point coordinate; and

and acquiring the coordinates of the port and starboard seabed points according to the extreme point coordinates.

6. The side-scan sonar sea-bottom line extraction method according to claim 5, wherein matching alignment is performed based on Euclidean distances between data points of the two spatio-temporal data sequences.

7. The side-scan sonar sea bottom line extraction method according to claim 1, wherein extracting sea bottom lines from sea bottom point coordinates includes: and performing smooth filtering on the coordinate sequences of the port and starboard seabed points of all Ping scanning lines to extract accurate seabed lines.

8. The utility model provides a side scan sonar undersea line extraction element which characterized in that, the device includes:

a storage unit for storing computer instructions;

a processor, communicatively coupled to the memory unit, for executing the computer instructions; when a processor executes the computer instructions, the side-scan sonar seafloor line extraction method of any one of claims 1-7 is performed.

9. A carrier comprising side-scan sonar acoustic systems arranged on the left and right sides, respectively, and a side-scan sonar sea-bottom line extraction device according to claim 8.

10. The vector of claim 9, wherein the vector is AUV.

Technical Field

The invention relates to the technical field of underwater acoustic signal processing, in particular to a method and a device for automatically extracting a side-scan sonar sea bottom line.

Background

Side Scan Sonar (SSS) records the intensity of submarine backscattered echoes by using an echo depth sounding principle, and generates sonar images with different brightness according to the echo intensity to detect submarine landforms and underwater objects. In the side scan sonar image, an obvious boundary line between a water column area and an image area is called as a sea bottom line, which represents the distance from the side scan sonar to the sea bottom and is an important parameter for measuring the sea bottom landform or underwater objects, correcting the slant distance and equalizing the image gray level.

At present, the working mode of the side scan sonar is mainly embodied by adopting towed measurement information of towed fish. For example, the seabed line extraction may be performed by a maximum amplitude method and a gradient method. However, in order to improve the submarine line extraction effect under strong noise interference, the method needs manual intervention and sets the initial value of the submarine line.

In addition, aiming at the problem that the image in the side-scan sonar water column area is interfered by transmitted pulses, sea surface echoes, wake flows, large-area suspended matters and the like, a sea bottom line final peak value detection method and a sea bottom line repairing method are adopted, and the automatic tracking and extraction of the sea bottom line under the influence of complex ocean noise are realized. However, the algorithm is complex in calculation process, and the continuity of extracting the details of the sea bottom line is poor.

In addition, a comprehensive method integrating the technologies of seabed tracking, slant range correction, towed fish homing calculation, geocoding, gap filling and the like is provided. However, in the data processing of the method, the existence of a middle blind area of the sonar transducer array is ignored, and the precision is not high.

Disclosure of Invention

The invention aims to provide a method and a device for automatically extracting a submarine line by side scan sonar, so that stronger noise interference is overcome under the condition of not needing manual intervention to set an initial value of the submarine line, and the submarine line is high in extraction accuracy and good in detail continuity.

According to one aspect of the invention, the side-scan sonar sea bottom line extraction method comprises the following steps:

acquiring a blind area boundary of echo data;

extracting the port and starboard echo data of all Ping scanning lines according to the width of the blind zone and the maximum sampling number of the port and starboard echo data, and eliminating abnormal points of the port and starboard echo data;

detecting the coordinates of the seabed points by using the port and starboard echo data for eliminating the abnormal points; and

and extracting a seabed line according to the seabed point coordinates.

According to a further embodiment of this aspect, the acquiring a blind zone boundary of the echo data includes: and respectively detecting the first echo data of each Ping scanning line on the port and the starboard, which is smaller than the blind zone boundary detection threshold value, and determining the blind zone width of the echo data according to the bilateral symmetry principle.

According to a further embodiment of this aspect, median filtering is used to remove outliers of the port and starboard echo data.

Further, the eliminating the outlier of the port and starboard echo data by using median filtering includes:

calculating the signal intensity median of each echo data of each Ping scanning line and a plurality of adjacent echo data of the front and rear adjacent Ping scanning lines in the port and starboard echo data;

calculating the absolute deviation of the signal intensity of each echo data of each Ping scanning line, and determining the signal intensity value of each echo data according to the absolute deviation of the signal intensity; and

and according to the relation between the absolute deviation of the signal intensity and a set value, taking the signal intensity median value or the signal intensity sampling value as the signal intensity numerical value of the echo data of each line.

According to a further embodiment of this aspect, the detecting subsea point coordinates using port and starboard echo data with outlier points removed comprises:

matching and aligning two space-time data sequences of the port and starboard echo data of each Ping scanning line;

taking logarithm of the matched and aligned port and starboard echo data;

calculating the standard deviation of the logarithmic sequence of the port and starboard echo data of each Ping scanning line, and constructing a comprehensive sequence of the port and starboard echo data according to the standard deviation;

carrying out maximum detection on the comprehensive sequence of the port and starboard echo data to obtain an extreme point coordinate; and

and acquiring the coordinates of the port and starboard seabed points according to the extreme point coordinates.

Further, matching alignment is performed based on Euclidean distances between data points of the two spatiotemporal data sequences.

According to a further embodiment of this aspect, the extracting the seafloor line from the seafloor point coordinates comprises: and performing smooth filtering on the coordinate sequences of the port and starboard seabed points of all Ping scanning lines to extract accurate seabed lines.

According to another aspect of the present invention, there is provided a side-scan sonar sea floor line extracting apparatus, including:

a storage unit for storing computer instructions;

a processor, communicatively coupled to the memory unit, for executing the computer instructions; when a processor executes the computer instructions, the side-scan sonar undersea line extraction method according to any one of the preceding claims is executed.

According to still another aspect of the present invention, there is provided a carrier including side-scan sonar acoustic systems arranged on left and right sides, respectively, and a side-scan sonar undersea line extraction device according to the above-described aspect. Optionally, the vector is AUV.

According to the invention, by considering the existence of blind areas among sonar transducers in practical application, firstly, the image is calculated to extract blind area parameters and remove high-frequency outlier noise, and then a dynamic time warping algorithm and an extreme value detection method are introduced to obtain seabed points.

Other features of the present invention and advantages thereof will become apparent from the following detailed description of exemplary embodiments thereof, which proceeds with reference to the accompanying drawings.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and not to limit the invention. In the drawings:

FIG. 1 is a side scan sonar waterfall plot;

FIG. 2 is a side scan sonar image taken alone for port and starboard;

FIG. 3 is a side scan sonar waterfall image main distribution parameter;

FIG. 4 is echo data for a single Ping scanline;

FIG. 5 is a flow chart of a method for automatically extracting a side-scan sonar sea bottom line according to a first embodiment of the present invention;

FIG. 6 is a side scan sonar image with no dead zones shown on port and starboard alone;

FIG. 7 is a side scan sonar image with no dead zone shown, taken separately from port and starboard after high frequency noise cancellation;

FIG. 8 is a side scan sonar image with submarine lines marked separately on the port and starboard side and without showing dead zones;

FIG. 9 is a side scan sonar waterfall plot with sea floor lines labeled and without showing dead zones;

fig. 10 is a schematic structural view of a side-scan sonar undersea line automatic extraction device according to a second embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The relative arrangement of the components and steps, the numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present invention unless specifically stated otherwise. Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate.

The side scan sonar is an active sonar system, and the working principle is that a short sound wave pulse with a very small horizontal open angle (about 1 degree) and a very large vertical open angle is emitted to one side or two sides of the vertical direction of the course of a measuring ship, and after the pulse reaches the seabed, the pulse is continuously reflected according to the distance between the seabed and a transducer, and a sonar image with uneven gray level change is generated according to the intensity degree of a reflected signal. The information such as the change of the landform of the seabed, whether the navigation obstacle, the type of the seabed substrate and the like exist can be observed from the sonar image.

When the side scan sonar emission pulse is propagated in a water body and meets a target, the target scatters sound energy to all directions, a transducer receives backscatter echoes, and the sound energy is difficult to reach behind the target, so that a blind area can be generated. The sonar array continuously advances along with the carrier, and sonar constantly launches in the advancing process, constantly receives and forms the sonar image, and the target (the strong echo signal of target) and its shade appear in the position that corresponds on the sonar image.

The side-scan sonar system mainly comprises a side-scan sonar acoustic system, a peripheral auxiliary sensor, a data real-time acquisition and processing system and the like. The transducer is a core component of a side-scan sonar acoustic system, and for example, a piezoelectric ceramic structure can be adopted, when a voltage is applied to the transmitting transducer, the physical form of the transmitting transducer is changed, an oscillating electric field generated by the transmitter is converted into mechanical deformation, and the deformation is transmitted into water to generate oscillating pressure, namely acoustic pulse, in the water. Also, a receiving transducer is used to receive the echo signal and convert the pressure change into electrical energy by detecting the acoustic pressure change. Typically, side scan sonar systems use a transmit-receive unified linear array when designing the transducer.

An Autonomous Underwater Vehicle (AUV) is a widely used unmanned marine vehicle and can carry a large amount of detection equipment to automatically complete marine surveying and mapping tasks. Therefore, the AUV is used as a measuring ship, the side scan sonar is installed on the AUV, and data fusion can be well carried out by means of good maneuvering performance of the AUV.

For example, Sea from Marine Sonic Technology may be usedAnd carrying out offshore measurement on the ARC SCOUT Mk-II, and collecting a plurality of Ping original data. The working parameters of the equipment are set as follows: the working frequency is 900kHz, and the slant distance is 10 m.

Fig. 1 shows a waterfall diagram obtained by inverting a gray image displayed by 256 gray levels of raw data obtained by a certain measurement. Fig. 2(a), (b) are respective port and starboard side scan sonar images. Fig. 3 is a waterfall plot of several Ping echo data after the scanning in fig. 1 starts, where a1 represents a sea bottom line, a2 represents a sea surface line, A3 represents a dead zone, a4 represents a water column region, and a5 represents a scanned image region. In the figure, darker colors represent stronger echoes, whereas whiter (lighter) colors represent weaker echoes.

In the present invention, unless otherwise specified, all the figures shown are the inverse of the gray scale image in which the original data (echo data) is displayed at 256 gray scales.

Fig. 4 schematically shows a workflow of the automatic side-scan sonar undersea line extraction method according to the first embodiment of the present invention. As shown in fig. 4, the method for automatically extracting the side-scan sonar sea bottom line comprises the following steps:

step 101, acquiring a blind zone boundary of echo data;

when extracting the bottom line of a sea, the prior art often makes an ideal assumption about the geometric problem of a sonar image, and considers that no blind area exists in the middle of a sonar transducer array or considers that the blind area is small and ignores the blind area. However, as can be seen from fig. 3, in the actual work of the side scan sonar, the sonar transducer emits sound waves to both sides, the sonar image has a blind area, and the blind area occupies a large space and should not be ignored. Therefore, the present invention requires first calculating and extracting the blind area parameter.

Figure 5 shows echo data for a single Ping scanline. As shown in fig. 5, the blind area is located right below the sonar, and the echo data of the Ping scan line suddenly changes from the blind area to the outside, and the numerical cliff type drops, whereby the detection of the blind area boundary can be realized. In general, echoes directly below the side-scan sonar are strong positive emission waves, and therefore, the present invention optionally employs a threshold method for blind zone boundary detection.

For a selected side scan sonar, the Ping data bit number is fixed, e.g., 8, 16, or 32bit data. If the data bit number of the side scan sonar is Nbit, a blind area boundary detection threshold D can be set according to the measurement environment condition_mComprises the following steps:

D_m＝2^N-2

from the blind area to the outer side, the first less than the blind area boundary detection threshold D of each Ping scanning line of the port and the starboard is respectively detected_mRespectively denoted as K_{Left side of}、K_{Right side}. Because the side scan sonar port and starboard transducers are symmetrically arranged relative to the central axis of the carrier, the echo data blind area width K of the Ping scanning line is determined according to the bilateral symmetry principle_mThe determination is as follows:

102, extracting the port and starboard echo data of all Ping scanning lines according to the width of the blind zone and the maximum sampling number of the port and starboard echo data, and eliminating abnormal points of the port and starboard echo data;

for each Ping scanning line, the maximum sampling number of echo data of a carrier, such as AUV left and right sides, is respectively K_{Left max}、K_{Right max}Then the port echo data [ K ] for all Ping scanlines can be extracted_m，K_{Left max}]And starboard echo data [ K ]_m，K_{Right max}]As shown in fig. 6.

As can be seen from fig. 6, in a complex environment, for example, suspended matter exists in a water column region, more abnormal points appear in echo data, which may interfere with the submarine line extraction and may even cause false detection. The echo abnormal point is essentially high-frequency outlier noise and needs to be removed.

According to an optional embodiment of the present invention, the value filtering removes echo outliers in the present invention, and the specific process includes:

(1) calculating the signal intensity median of each echo data of each Ping scanning line and a plurality of adjacent echo data of the front and rear adjacent Ping scanning lines in the port and starboard echo data;

(2) and calculating the absolute deviation of the signal intensity of each echo data of each Ping scanning line, and determining the signal intensity value of each echo data according to the absolute deviation of the signal intensity.

And when the absolute deviation of the signal intensity is greater than a set value, taking the median of the signal intensity as the signal intensity numerical value of each echo data. Otherwise, the outlier removing operation is not performed, and the signal strength value of each echo data is the measured value (sampling value).

The set value is a calculated value based on a signal strength measurement of echo data of each Ping scanline.

For example, for port echo data [ K ]_m，K_{Left max}]Let [ i, j)]Index identifier, I, of jth echo sample data representing ith Ping scan line_ijDenotes the [ i, j ]]Signal strength value, P, of the port echo data_maxRepresenting the total number of side scan sonar Ping scans, then:

for the satisfaction of 1<i<P_max，K_m≤j<K_{Left max}All of [ i, j ] of]And combining, namely performing the following data processing:

(1) calculate [ i-1, j-1 [ ] -1]、[i-1,j]、[i-1,j+1]、[i,j-1]、[i,j]、[i,j+1]、[i+1,j-1]、[i+1,j]、[i+1,j+1]Median Z of the port echo data_ijAnd calculating the absolute deviation E_ij＝abs(I_ij-Z_ij)；

This example selects the jth echo data of the ith Ping scan line, and its 3 adjacent echo data of the preceding, i.e., ith-1 Ping scan line, i.e., jth-1, j +1 echo data, and its 3 adjacent echo data of the following, i.e., ith +1 Ping scan line, i.e., jth-1, j +1 echo data. However, it will be understood by those skilled in the art that the number of adjacent scan lines, or the number of echo data adjacent to each adjacent scan line, may be selected according to actual needs. Similarly, the same is true for the subsequently described numerical calculation of the signal intensity of the starboard echo data.

(2) Determine absolute deviation E_ij；

If E is_ij>9.78+1.44×I_ij-8.86×10^-4×I_ij ²+2.95×10^-7×I_ij ³-3.26×10^-11×I_ij ²Then adopt Z_ijAs the [ i, j ]]A new value of the signal strength of the port echo data; otherwise, no processing is done.

For port echo data [ K ]_m，K_{Right max}]Let [ i, l]Index identifier of the ith echo sample data, I, representing the ith Ping scan line_ilDenotes the [ i, l ]]Signal strength values for the individual starboard echo data.

For the satisfaction of 1<i<P_max，K_m≤l<K_{Right max}All of [ i, l ] of]And combining, namely performing the following data processing:

(1) calculation of [ i-1, l-1 [ ]]、[i-1,l]、[i-1,l+1]、[i,l-1]、[i,l]、[i,l+1]、[i+1,l-1]、[i+1,l]、[i+1,l+1]Median Z of starboard echo data_ilAnd calculating the absolute deviation E_il＝abs(I_il-Z_il)；

(2) Determine absolute deviation E_il；

If E is_il>9.78+1.44×I_il-8.86×10^-4×I_il ²+2.95×10^-7×I_il ³-3.26×10^-11×I_il ²Then adopt Z_ilAs the (i, l)]A new value of signal strength for the individual starboard echo data; otherwise, no processing is done.

Fig. 7 shows a side scan sonar image with no dead zone shown, port and starboard alone, after high frequency outlier noise cancellation. Comparing fig. 6, it can be seen that the echo abnormal point in the side-scan data water column region can be eliminated to some extent. And for the data processing, the echo abnormal point eliminating effect of the port echo data is better.

103, detecting the coordinates of the seabed points by using the port and starboard echo data for eliminating the abnormal points;

according to the sea bottom line bilateral symmetry principle, the first echoes received by the side-scan sonar port and starboard transducers are from the sea bottom right below, experience depths are the same, and the sea bottom lines extracted from the two sides are symmetrical about a navigation track line. By utilizing the principle, the sea bottom point detection can be better carried out.

According to an alternative embodiment of the invention, the subsea point detection comprises the steps of:

step 1031, matching and aligning two space-time data sequences of the left and right side echo data of each Ping scanning line;

because the maximum sampling number K of the left and right side echo data often appears in the same Ping scanning line_{Left max}、K_{Right max}Unequal problem, as shown in fig. 4, results in I in the port and starboard echo data when l is j_ijAnd I_ilThe temporal and spatial information involved is different. In order to fully utilize the sea bottom line bilateral symmetry principle and realize comprehensive extraction comparison between a port side and a starboard side so as to remove the influence of suspended matters as much as possible, the problem is firstly solved.

In the invention, the seabed line extracts the data of the water column region and the data of the abrupt change part from the water column region to the image region, therefore, in order to keep the robustness of the invention, the left and right side echo data can be considered as two space-time data sequences to be matched and aligned. When matching and aligning, the consistency of the time-space information of the abrupt change from the water column area to the image area needs to be kept. According to the optional embodiment of the present invention, the matching alignment of the two sequences in space-time by using a dynamic time warping algorithm specifically includes the following steps:

(1) for each Ping scanline, two spatio-temporal data sequences, respectively, left side echo data { I_ij，K_m≤j≤K_{Left max}And starboard echo data I_il，K_m≤l≤K_{Right max}}。

In each of the two sequences, a Euclidean distance D (I) between two points is calculated_ij,I_il) In which K is_m≤j≤K_{Left max}，K_m≤l≤K_{Right max}。

And (3) calculating Euclidean distances of all data points in the sequence to construct a Euclidean distance table, wherein the Euclidean distances are shown in the following table:

D(IiKm,IiKm)	D(IiKm,IiKm+1)	…	D(IiKm,IiK Right max)
				D(IiKm+1,IiKm)	D(IiKm+1,IiKm+1)	…	D(IiKm+1,IiK Right max)
…	…	…	…
				D(IiK left max,IiKm)	D(IiK left max,IiKm+1)	…	D(IiK left max,IiK Right max)

(2) Searching the shortest path in the Euclidean distance table;

in the Euclidean distance table, D (I) is searched for, for example, by using a dynamic programming or greedy search algorithm_iKm,I_iKm) To D (I)_{iK left max},I_{iK Right max}) Is the shortest route ofAnd (4) diameter. This shortest path needs to satisfy: if the current node is D (I)_ij,I_il) Then the next node must be at D (I)_ij+1,I_il)、D(I_ij,I_il+1)、D(I_ij+1,I_il+1) And the path must be shortest.

(3) According to the Euclidean distance table nodes passed by the shortest path, the aligned data point pairs of two discrete sequences can be obtained, and the sequences have the same length and are marked as K_max。

Step 1032, logarithm is taken on the matched and aligned port and starboard echo data;

the logarithm of the echo data after the high-frequency noise cancellation processing is performed on the port and starboard matching and aligning, for example, a common logarithm with a base of 10 is taken.

Step 1033, calculating a standard deviation of the logarithmic sequence of the port and starboard echo data of each Ping scanning line, and constructing a comprehensive sequence of the port and starboard echo data according to the standard deviation;

calculating the standard deviation of the logarithmic sequence of the port echo data and the starboard echo data of each Ping scanning line to obtain two standard deviation sequences which are respectively marked as { L }_iq，1≤q≤K_max-W +1} and { L_is，1≤s≤K_max-W +1}, and the sequence lengths are all K_max-W + 1. Wherein the value of W can also be determined by reference to the following procedure:

for example, echo outliers existing in the water column region in the port echo data in fig. 5 may cause a large fluctuation in the standard deviation sequence of the water column region when calculating the standard deviation sequence, which may affect the accuracy of the subsequent extremum detection method.

Therefore, in order to fully utilize the sea bottom line bilateral symmetry principle, the invention further takes the minimum value of the starboard and starboard logarithmic sequence standard deviation of each Ping scanning line, namely: for any s ═ q, q is more than or equal to 1 and less than or equal to K_max-W+1，1≤s≤K_max-W +1, get L_ig＝min(L_iq,L_is) And then constructing a comprehensive sequence of the port and starboard echo data { L by using the minimum value_ig，1≤g≤K_max-W+1}。

1034, carrying out maximum detection on the comprehensive sequence of the port and starboard echo data to obtain an extreme point coordinate;

according to the characteristics of the side scan sonar data, the sea bottom line is composed of the first sea bottom strong echo sequence from port to starboard in each Ping scanning line, so that the sudden change of the echo data from the water column area to the image area results in the port and starboard comprehensive sequence { L }_ig，1≤g≤K_maxThe first maximum in W +1, therefore, by synthesizing the sequence { L) for port and starboard_ig，1≤g≤K_max-W +1} to obtain an extreme point coordinate g_b。

And 1035, obtaining the left and right ship bottom point coordinates according to the extreme point coordinates.

Using extreme point coordinates g_bThe coordinate q of the standard sequence of the logarithm of the port echo data and the starboard echo data can be correspondingly obtained_bAnd s_bAnd then the coordinate q of the echo data sequence after the port and starboard matching and aligning corresponding to the extreme point can be obtained_b+ W-1 and corresponding echo data I_{b left side}、s_b+ W-1 and corresponding echo data I_{b right side}。

For each Ping scanline, search for the port echo data sequence I_ij，K_m≤j≤K_{Left max}Find the first one and I_{b left side}Equal echo data, the echo data coordinate is the port sea bottom point coordinate j_b. Likewise, starboard echo data sequence { I ] is searched_il，K_m≤l≤K_{Right max}Find the first one and I_{b right side}Equal echo data, the echo data coordinate is starboard seabed point coordinate l_b。

And 104, extracting a seabed line according to the seabed point coordinates.

The coordinate sequences { j of the port seabed points of all Ping scanning lines obtained by calculation_bStarboard sea floor point coordinate sequence{l_bAnd performing smooth filtering, namely extracting an accurate seabed line.

The side-scan sonar image which is obtained by the extraction method and separately labels the sea bottom line and does not display the blind area is shown in fig. 8, and the side-scan sonar waterfall image which is labeled the sea bottom line and does not display the blind area is shown in fig. 9.

According to a second embodiment of the present invention, there is also provided an automatic side-scan sonar undersea line extraction apparatus, as shown in fig. 10, including:

a storage unit 10 for storing computer instructions;

a processor 20, communicatively coupled to the memory unit 10, for executing the computer instructions. When the processor 20 executes the computer instructions, the automatic extraction method of the side-scan sonar undersea line described above is performed.

According to a third embodiment of the present invention, there is also provided a carrier for mounting a side scan sonar system.

The carrier comprises side-scan sonar acoustic systems which are respectively arranged on a left board and a right board, and a side-scan sonar sea bottom line extraction device.

The vector may be AUV.

The above examples are only intended to illustrate the technical solution of the present invention and not to limit it; although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art will understand that: modifications to the specific embodiments of the invention or equivalent substitutions for parts of the technical features may be made; without departing from the spirit of the present invention, it is intended to cover all aspects of the invention as defined by the appended claims.