Underwater terrain matching positioning method, device and storage medium

Abstract

The present invention provides an underwater terrain matching positioning method, device and storage medium. The method comprises the following steps: receiving an actual depth data map; dividing the actual depth data map into grid units, and calculating depth parameters of the grid unit centroids; dividing a preset positioning area into multiple candidate area maps; constructing a two-dimensional coordinate map of the actual depth data map and the candidate area map, inputting the map into a first model, and converting the map into a first vector; screening the candidate area map based on the distance between the first vectors, and expanding the screened candidate area map to obtain a candidate area map; obtaining a covariance matrix of each grid unit, and calculating a linear value, a flatness value and a scattering value of the grid unit based on an eigenvalue; extracting a matching area map from the candidate area map and the actual depth data map, and inputting the map into a second model to obtain a second vector; calculating the sum of the distances between the actual depth data map and each candidate area map based on the second vector, and matching the candidate area map based on the distance sum.

Description

Technical Field

[0001] The present invention relates to the field of underwater positioning technology, and in particular to an underwater terrain matching positioning method, device and storage medium. Background Art

[0002] In traditional methods, due to the limitations of underwater communications, autonomous underwater vehicles (AUVs) need to surface to receive satellite signals for position correction after a long period of underwater navigation. This undoubtedly increases the time and energy consumption of AUVs for deep diving and long-range missions.

[0003] In order to reduce the time and energy cost of AUV position correction, some acoustic positioning correction methods have also been used for AUV position correction, but the existing acoustic positioning schemes have large errors. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide an underwater terrain matching positioning method, apparatus, and storage medium to eliminate or improve one or more defects in the prior art.

[0005] A first aspect of the present invention provides an underwater terrain matching positioning method, the method comprising the steps of:

[0006] Receiving an actual depth data map transmitted by a sonar device, wherein the actual depth data map is marked with a sounding point having latitude and longitude parameters and depth parameters;

[0007] Dividing the actual depth data map into a plurality of grid units based on a preset first resolution to obtain a grid depth map, calculating the position of the centroid of each grid unit, and calculating the depth parameter of the centroid based on the depth parameter of the first depth measurement point close to the centroid;

[0008] Divide the preset positioning area into multiple candidate area maps according to the size of the actual depth data map, each candidate area map is divided into grid cells of the first resolution size, and the center of gravity of each grid cell is marked with a depth parameter and a latitude and longitude parameter;

[0009] Constructing a two-dimensional coordinate graph of the actual depth data graph and the candidate area graph based on the growth direction of the longitude and latitude data, inputting the two-dimensional coordinate graph into a preset first model, and converting each two-dimensional coordinate graph into a first vector;

[0010] Filtering the plurality of candidate area maps based on a distance between a first vector corresponding to the actual depth data map and a first vector corresponding to the candidate area map, and expanding the range of the filtered candidate area maps within the positioning area to obtain a candidate area map;

[0011] Based on the preset neighborhood size, the covariance matrix of each grid cell in the selected area map and the actual depth data map is obtained, the eigenvalues ​​of each covariance matrix are calculated, and the linearity value, flatness value and scattering value of each grid cell are calculated based on the eigenvalues, and the linearity value, flatness value and scattering value are assigned to the corresponding grid cell;

[0012] Based on a preset extraction window, a plurality of matching area maps of the size of the extraction window are extracted from the candidate area map and the actual depth data map respectively, wherein grid cells having a positive vertical coordinate value of a center of gravity, a depth parameter, a linear value, a flatness value, and a scattering type value are set in the matching area;

[0013] The matching area map is input into a preset second model to obtain a second vector corresponding to each matching area map, and the distance sum between the matching area map of the actual depth data map and the matching area map of each candidate area map is calculated based on the second vector, and the candidate area map is matched based on the distance sum.

[0014] Using the above scheme, this scheme first obtains multiple candidate area maps through a first vector including horizontal and vertical coordinates and depth parameters, and further adds three parameters, linear value, flatness value and scattering type value, to each grid cell through the covariance matrix. Among them, the linear value describes the degree of elongation of the neighborhood, while the flatness value evaluates the degree of a plane fitting, and the scattering value corresponds to an isotropic spherical neighborhood. The final candidate area map is further obtained through grid cells with positive vertical coordinate values, depth parameters, linear values, flatness values ​​and scattering type values ​​to complete precise underwater positioning.

[0015] In some embodiments of the present invention, the method further comprises:

[0016] Obtaining a second vector corresponding to the matching area map of the center of gravity position of the actual depth data map;

[0017] Calculating the distance between the second vector and the second vector of the matching area graph in the matched area graph to be selected;

[0018] The position of the matching area map in the candidate area map with the smallest distance value is selected as the position of the actual depth data map.

[0019] In some embodiments of the present invention, the step of receiving the actual depth data map transmitted by the sonar device further includes the following steps:

[0020] Obtain multiple sounding points closest to each sounding point, and calculate the average value of the depth parameters of the sounding point and the multiple sounding points closest to it;

[0021] Determine whether the absolute value of the difference between the depth parameter of the sounding point and the average value is greater than a preset first threshold;

[0022] If it is greater, the sounding point is determined to be an outlier, and the depth parameter of the sounding point is modified to the average of the depth parameters of the multiple sounding points closest to the sounding point.

[0023] In some embodiments of the present invention, the step of calculating the depth parameter of the center of gravity based on the depth parameter of the first sounding point close to the center of gravity includes:

[0024] Calculate the distance between the center of gravity and each sounding point in the first sounding point respectively;

[0025] A weight is assigned to each of the first number of sounding points based on the distance value, and the weighted average value of the depth parameters of the first number of sounding points is calculated as the depth parameter of the centroid point.

[0026] In some embodiments of the present invention, the weighted average value of the depth parameter of the first sounding point is calculated using the following formula:

[0027]

[0028] s represents the weighted average value, n represents the value of the first number, i represents any of the sounding points of the first number, s i Indicates the depth parameter of the i sounding point, d i Indicates the distance between the sounding point i and the center of gravity.

[0029] In some embodiments of the present invention, the step of calculating the linearity value, the flatness value, and the scattering value of each grid cell based on the eigenvalue includes:

[0030] Sort the obtained multiple eigenvalues ​​from large to small according to their numerical values;

[0031] The three largest eigenvalues ​​are obtained, where eigenvalue one ≥ eigenvalue two ≥ eigenvalue three, and the linearity value, the flatness value, and the scattering type value are calculated based on the eigenvalue one, the eigenvalue two, and the eigenvalue three.

[0032] In some embodiments of the present invention, the linearity value, the flatness value, and the scattering value are calculated based on the eigenvalue 1, the eigenvalue 2, and the eigenvalue 3 according to the following formula:

[0033]

[0034]

[0035]

[0036] In some embodiments of the present invention, the distance sum between the matching area map of the actual depth data map and the matching area map of each candidate area map is calculated based on the second vector, and the step of matching the candidate area map based on the distance sum is as follows:

[0037] Calculate the sum of the distances between the matching area map of the actual depth data map and the matching area map of each candidate area map respectively;

[0038] The distance sum of the actual depth data map for each matching area map is added to obtain the distance sum, and the matching area map with the smallest distance sum is the final matching candidate area map.

[0039] The second aspect of the present invention provides an underwater terrain matching and positioning device, which includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, the processor is used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the device implements the steps of the above method.

[0040] A third aspect of the present invention provides a computer-readable storage medium, characterized in that a computer program is stored thereon, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0041] Additional advantages, objects, and features of the present invention will be described in part in the following description and will become apparent to those skilled in the art after studying the following or may be learned by practice of the present invention. The objects and other advantages of the present invention may be particularly pointed out and attained in the description and drawings.

[0042] Those skilled in the art will understand that the purposes and advantages that can be achieved by the present invention are not limited to the above specific descriptions, and the above and other purposes that can be achieved by the present invention will be more clearly understood based on the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The drawings described herein are used to provide a further understanding of the present invention, constitute a part of this application, and do not constitute a limitation of the present invention.

[0044] Figure 1 Schematic diagram of a first embodiment of the underwater terrain matching and positioning method of the present invention;

[0045] Figure 2 This is a schematic diagram of a second embodiment of the underwater terrain matching and positioning method of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0047] It should also be noted that, in order to avoid obscuring the present invention due to unnecessary details, the accompanying drawings only show structures and / or processing steps closely related to the solutions according to the present invention, while other details that are not closely related to the present invention are omitted.

[0048] It should be emphasized that the term "include / comprises" when used herein refers to the existence of features, elements, steps or components, but does not exclude the existence or addition of one or more other features, elements, steps or components.

[0049] It should also be noted that, unless otherwise specified, the term "connection" herein may refer not only to a direct connection but also to an indirect connection involving an intermediate.

[0050] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the accompanying drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

[0051] Traditional underwater terrain matching positioning algorithms have low positioning accuracy and slow speed, often misjudge similar terrains, and are prone to falling into local optimality and difficult to accurately locate; this application provides an underwater terrain matching positioning method with high accuracy, fast speed and the ability to achieve precise positioning.

[0052] To solve the above problems, Figure 1 As shown, the present invention proposes an underwater terrain matching positioning method, the steps of the method include:

[0053] Step S100, receiving an actual depth data map transmitted by a sonar device, wherein the actual depth data map is marked with a depth measurement point having latitude and longitude parameters and depth parameters;

[0054] In some embodiments of the present invention, the actual depth data map transmitted by the sonar device is an actual depth data map obtained by the sonar device emitting sonar waves toward the bottom of the water and receiving feedback sonar waves.

[0055] In some embodiments of the present invention, the actual depth data map is a two-dimensional image including a plurality of sounding points as a point cloud, projected from the water surface toward the bottom of the water.

[0056] Step S200, dividing the actual depth data map into a plurality of grid units based on a preset first resolution to obtain a grid depth map, calculating the position of the centroid of each grid unit, and calculating the depth parameter of the centroid based on the depth parameter of the first depth measurement point close to the centroid;

[0057] In some embodiments of the present invention, the shape of the grid unit can be rectangular or square, preferably square. The preset first resolution is the size of each grid, which can be the side length of the square grid unit. Each grid unit can include a depth measurement point within its range.

[0058] Step S300: Divide the preset positioning area into multiple candidate area maps according to the size of the actual depth data map, each candidate area map is divided into grid cells of a first resolution size, and the center of gravity of each grid cell is marked with a depth parameter and a latitude and longitude parameter;

[0059] In some embodiments of the present invention, the preset positioning area is an area map that is much larger than the actual depth data map, the position of the actual depth data map is within the range of the positioning area, and the positioning area is pre-divided into multiple grid units of the same size, and the longitude and latitude parameters and depth parameters are marked at the center of gravity of each grid unit.

[0060] In some embodiments of the present invention, the size of the candidate region is equal to the size of the actual depth data map.

[0061] In some embodiments of the present invention, the step of dividing the preset positioning area into multiple candidate area maps according to the size of the actual depth data map is to pre-build a sliding window of the size of the actual depth data map, the sliding window size can be a×a, and the sliding window slides within the positioning area, and each movement The range defined by the initial position of the sliding window and the range defined after each movement are both used as candidate area maps.

[0062] Preferably, the sliding window slides from the upper left corner to the lower right corner of the positioning area or from the lower right corner to the upper left corner.

[0063] With the above solution, the sliding window movement range covers all positions in the positioning area, ensuring that the candidate area map includes images of any position in the positioning area.

[0064] Step S400: constructing a two-dimensional coordinate graph of the actual depth data graph and the candidate area graph based on the growth direction of the longitude and latitude data, inputting the two-dimensional coordinate graph into a preset first model, and converting each two-dimensional coordinate graph into a first vector;

[0065] In some embodiments of the present invention, a method for constructing a two-dimensional coordinate map for the actual depth data map and the candidate area map is to set the coordinates of the center of gravity of the grid cell with the smallest longitude and latitude value in the actual depth data map or the candidate area map to (0,0), set the coordinates of the center of gravity of the grid cell adjacent to the origin whose longitude is greater than the origin to (0,1), ..., set the coordinates of the center of gravity of the grid cell with the largest longitude and latitude value in the actual depth data map or the candidate area map to Where a represents the side length of the actual depth data map, and r represents the size of the first resolution.

[0066] In some embodiments of the present invention, the first model may be a MobileNetV3-small model trained using the ImageNet dataset. The actual depth data map or the two-dimensional coordinate map of the candidate region map is input into the MobileNetV3-small model, and the first vector is obtained from the third-to-last pooling layer of the model.

[0067] The two-dimensional coordinate map includes the center of gravity of each grid unit, and each center of gravity is marked with a horizontal coordinate value, a vertical coordinate value and a depth parameter value.

[0068] The first vector is a 576-dimensional vector.

[0069] Step S500: screening multiple candidate area maps based on the distance between the first vector corresponding to the actual depth data map and the first vector corresponding to the candidate area map, and expanding the range of the screened candidate area map within the positioning area to obtain a candidate area map;

[0070] In some embodiments of the present invention, the distance between the first vector corresponding to the actual depth data map and the first vector corresponding to each candidate area map is calculated, and M candidate area maps with smaller distances are selected;

[0071] The step of expanding the range of the screened candidate area map within the positioning area is to expand the side length of each candidate area map in the M candidate area maps to twice the original length, and the expansion direction is preferably expanded by half in the upward, downward, left and right directions. If it is expanded to the boundary of the positioning area in one direction, it is expanded in the other direction to obtain M candidate area maps.

[0072] In some embodiments of the present invention, if the size of the unexpanded candidate region is a×a, the size of the to-be-selected region map is 2a×2a.

[0073] Using the above scheme, this scheme first expands the candidate area to facilitate coarse positioning and prevent positioning errors.

[0074] Step S600: Based on a preset neighborhood size, obtain the covariance matrix of each grid cell in the selected area map and the actual depth data map, calculate the eigenvalues ​​of each covariance matrix, calculate the linearity value, flatness value, and scattering value of each grid cell based on the eigenvalues, and assign the linearity value, flatness value, and scattering value to the corresponding grid cell;

[0075] In some embodiments of the present invention, the neighborhood may extend 1, 2, or 10 grid cells to the top, bottom, left, and right of the grid cell, respectively. This is not specifically limited herein. The covariance matrix of each grid cell is calculated jointly by the grid cell itself and the grid cells in the neighborhood of the grid cell.

[0076] Furthermore, if it expands to the edge in one direction, it expands in the other direction;

[0077] In some embodiments of the present invention, a centroid point vector (abscissa, ordinate, depth value) is constructed for each grid cell in the selected area map and the actual depth data map according to the horizontal and vertical coordinates and depth value of the centroid point, and a covariance matrix is ​​constructed based on the centroid point vector of each grid cell and the neighborhood of the grid cell.

[0078] In some embodiments of the present invention, the linear value describes the degree of elongation of the neighborhood, while the flatness value evaluates the degree of fit of a plane, and the scattering value corresponds to an isotropic spherical neighborhood, which increases the dimension considered by this solution and improves positioning accuracy.

[0079] Step S700: Based on a preset extraction window, a plurality of matching area maps of the size of the extraction window are extracted from the candidate area map and the actual depth data map, wherein the matching area is provided with grid cells having a positive vertical coordinate value of a center of gravity, a depth parameter, a linear value, a flatness value, and a scattering type value;

[0080] In some embodiments of the present invention, the extraction window size can be 0.5a×0.5a, and a sliding window method is adopted to move the extraction window by 0.05a each time, and the range defined by the initial position of the extraction window and the range defined after each movement are used as matching area maps.

[0081] Step S800: Input the matching area map into a preset second model to obtain a second vector corresponding to each matching area map, calculate the distance sum between the matching area map of the actual depth data map and the matching area map of each candidate area map based on the second vector, and match the candidate area map based on the distance sum.

[0082] In some embodiments of the present invention, the second model may be a PN-Net model, which is trained based on a preset training data set using Triplet loss as a loss function.

[0083] In some embodiments of the present invention, the matching region map input to the second model includes abscissa values, ordinate values, depth parameters, linearity values, flatness values, and scattering type values ​​corresponding to each grid cell.

[0084] In some embodiments of the present invention, the second vector output by the second model is a 256-dimensional vector.

[0085] In some embodiments of the present invention, the distance between the first vectors or the distance between the second vectors can be calculated using a cosine distance formula.

[0086] Using the above scheme, this scheme first obtains multiple candidate area maps through a first vector including horizontal and vertical coordinates and depth parameters, and further adds three parameters, linear value, flatness value and scattering type value, to each grid cell through the covariance matrix. Among them, the linear value describes the degree of elongation of the neighborhood, while the flatness value evaluates the degree of a plane fitting, and the scattering value corresponds to an isotropic spherical neighborhood. The final candidate area map is further obtained through grid cells with positive vertical coordinate values, depth parameters, linear values, flatness values ​​and scattering type values ​​to complete precise underwater positioning.

[0087] like Figure 2 As shown, in some embodiments of the present invention, the method further includes step S900:

[0088] Obtaining a second vector corresponding to the matching area map of the center of gravity position of the actual depth data map;

[0089] Calculating the distance between the second vector and the second vector of the matching area graph in the matched area graph to be selected;

[0090] The position of the matching area map in the candidate area map with the smallest distance value is selected as the position of the actual depth data map.

[0091] By adopting the above scheme, the actual depth data map is further positioned in the map of the selected area to achieve precise positioning after rough positioning, completing the two-stage underwater terrain matching positioning strategy, namely coarse positioning and precise positioning, improving positioning accuracy and facilitating the positioning of specific locations.

[0092] In some embodiments of the present invention, the step of receiving the actual depth data map transmitted by the sonar device further includes the following steps:

[0093] Obtain multiple sounding points closest to each sounding point, and calculate the average value of the depth parameters of the sounding point and the multiple sounding points closest to it;

[0094] Determine whether the absolute value of the difference between the depth parameter of the sounding point and the average value is greater than a preset first threshold;

[0095] If it is greater, the sounding point is determined to be an outlier, and the depth parameter of the sounding point is modified to the average of the depth parameters of the multiple sounding points closest to the sounding point.

[0096] In some embodiments of the present invention, if it is not greater than, the depth parameter of the sounding point is not changed.

[0097] In some embodiments of the present invention, the first threshold is determined based on a preset environmental error δ, which may be 5δ.

[0098] When adopting the above scheme, due to the presence of underwater organisms such as fish, the sonar waves may be blocked by the underwater organisms, thereby causing abnormal points. This application determines the abnormal points based on the absolute value of the difference between the depth parameter of the sounding point and the average value to prevent the data of the abnormal points from affecting the subsequent terrain matching of this application, thereby affecting the positioning accuracy.

[0099] In some embodiments of the present invention, the step of calculating the depth parameter of the center of gravity based on the depth parameter of the first sounding point close to the center of gravity includes:

[0100] Calculate the distance between the center of gravity and each sounding point in the first sounding point respectively;

[0101] A weight is assigned to each of the first number of sounding points based on the distance value, and the weighted average value of the depth parameters of the first number of sounding points is calculated as the depth parameter of the centroid point.

[0102] In some embodiments of the present invention, the weighted average value of the depth parameter of the first sounding point is calculated using the following formula:

[0103]

[0104] s represents the weighted average value, n represents the value of the first number, i represents any of the sounding points of the first number, s i Indicates the depth parameter of the i sounding point, d i Indicates the distance between the sounding point i and the center of gravity.

[0105] Using the above scheme, the depth parameter is calculated for the centroid point of each grid cell in the actual depth data map.

[0106] In some embodiments of the present invention, the step of calculating the linearity value, the flatness value, and the scattering value of each grid cell based on the eigenvalue includes:

[0107] Sort the obtained multiple eigenvalues ​​from large to small according to their numerical values;

[0108] The three largest eigenvalues ​​are obtained, where eigenvalue one ≥ eigenvalue two ≥ eigenvalue three, and the linearity value, the flatness value, and the scattering type value are calculated based on the eigenvalue one, the eigenvalue two, and the eigenvalue three.

[0109] In some embodiments of the present invention, multiple eigenvalues ​​may be calculated for each covariance matrix.

[0110] In some embodiments of the present invention, the linearity value, the flatness value, and the scattering value are calculated based on the eigenvalue 1, the eigenvalue 2, and the eigenvalue 3 according to the following formula:

[0111]

[0112]

[0113]

[0114] In some embodiments of the present invention, the distance sum between the matching area map of the actual depth data map and the matching area map of each candidate area map is calculated based on the second vector, and the step of matching the candidate area map based on the distance sum is as follows:

[0115] Calculate the sum of the distances between the matching area map of the actual depth data map and the matching area map of each candidate area map respectively;

[0116] The distance sum of the actual depth data map for each matching area map is added to obtain the distance sum, and the matching area map with the smallest distance sum is the final matching candidate area map.

[0117] In some embodiments of the present invention, the actual depth data map includes multiple matching area maps, and the corresponding second vectors are a1, b1, and c1 respectively; the matched selected area map includes multiple matching area maps, and the corresponding second vectors are a2, b2, c2, d2, and e2 respectively;

[0118] Calculate the distance between a1 and a2, b2, c2, d2 and e2, and add them up to get the sum of the distances of a1;

[0119] Calculate the distance between b1 and a2, b2, c2, d2 and e2, and add them up to get the sum of the distances of b1;

[0120] Calculate the distance between c1 and a2, b2, c2, d2 and e2, and add them up to get the sum of the distances of c1;

[0121] Add the sum of distances a1, b1, and c1 to get the total distance.

[0122] The sum of distances calculated from the matched candidate region graph compared to other candidate region graphs is the smallest among all candidate region graphs.

[0123] The second aspect of the present invention provides an underwater terrain matching and positioning device, which includes a computer device, the computer device includes a processor and a memory, the memory stores computer instructions, the processor is used to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the device implements the steps of the above method.

[0124] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the aforementioned underwater terrain matching positioning method. The computer-readable storage medium can be a tangible storage medium, such as a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a floppy disk, a hard disk, a removable storage disk, a CD-ROM, or any other form of storage medium known in the art.

[0125] To address the low accuracy of previous terrain matching positioning algorithms, this approach first rasterizes the sonar bathymetric data based on the first-resolution pre-stored underwater terrain data. This includes operations such as denoising, interpolation, and merging. A two-stage terrain matching positioning strategy is then employed: large-scale coarse positioning obtains general location information, followed by small-scale precision positioning to obtain precise location information. To address the low efficiency of previous terrain matching algorithms, a pre-trained deep neural network and pre-stored terrain block calculation results are used to improve the efficiency of the positioning algorithm. This positioning algorithm can be run offline or online, ensuring the real-time operation of the underwater terrain matching positioning algorithm through deep learning-oriented servers or edge computing devices.

[0126] This solution adopts a feature extraction and matching positioning strategy based on deep learning. It does not require manual setting of feature extraction modes. It automatically extracts high-level features through pre-trained deep neural networks. This can avoid the situation where the traditional manually set feature extraction mode may cause the algorithm to fail due to factors such as environment and equipment. The two-stage positioning method can achieve a compromise between operating efficiency and positioning accuracy, and achieve rapid positioning while ensuring accurate positioning. The grid-based underwater terrain data matching and processing method can fully utilize network coordinates to find spatial neighbors. By correcting and rasterizing terrain data, the accuracy of terrain matching positioning can be improved. The parameters of this algorithm can be adjusted according to the accuracy of the acquisition equipment, the resolution of the pre-stored terrain, the running time limit, and the positioning accuracy requirements.

[0127] Improvements to this solution include:

[0128] 1. A two-stage underwater terrain matching positioning strategy, namely coarse positioning and fine positioning;

[0129] 2. Use a pre-trained deep neural network to process pre-stored local terrain area data and store the extracted feature vectors in advance to improve computational efficiency;

[0130] 3. Use rasterized data processing and matching positioning methods.

[0131] The two-step search and matching scheme is the most accurate and efficient algorithm step. Aspects that can be improved include:

[0132] Parameter adjustment adopts a more flexible approach, including some adaptive algorithms and machine learning algorithms;

[0133] The feature extraction and matching networks used in coarse and precise positioning can continue to be improved to enhance the algorithm execution effect.

[0134] It should be understood by those skilled in the art that the various exemplary components, systems and methods described in conjunction with the embodiments disclosed herein can be implemented in hardware, software or a combination of the two. Whether it is specifically performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present invention are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link via a data signal carried in a carrier.

[0135] It should be understood that the present invention is not limited to the specific configurations and processes described above and illustrated in the figures. For the sake of brevity, a detailed description of known methods is omitted. In the above embodiments, several specific steps are described and illustrated as examples. However, the method of the present invention is not limited to the specific steps described and illustrated. Those skilled in the art may make various changes, modifications, and additions, or change the order of the steps after understanding the spirit of the present invention.

[0136] In the present invention, features described and / or illustrated for one embodiment may be used in the same or similar manner in one or more other embodiments, and / or combined with or replace features of other embodiments.

[0137] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be within the scope of protection of the present invention.

Claims

1. An underwater terrain matching positioning method, characterized in that: The steps of the method include, Receiving an actual depth data map transmitted by a sonar device, wherein the actual depth data map is marked with a sounding point having latitude and longitude parameters and depth parameters; Dividing the actual depth data map into a plurality of grid units based on a preset first resolution to obtain a grid depth map, calculating the position of the centroid of each grid unit, and calculating the depth parameter of the centroid based on the depth parameter of the first depth measurement point close to the centroid; Divide the preset positioning area into multiple candidate area maps according to the size of the actual depth data map, each candidate area map is divided into grid cells of the first resolution size, and the center of gravity of each grid cell is marked with a depth parameter and a latitude and longitude parameter; Constructing a two-dimensional coordinate graph of the actual depth data graph and the candidate area graph based on the growth direction of the longitude and latitude data, inputting the two-dimensional coordinate graph into a preset first model, and converting each two-dimensional coordinate graph into a first vector; Filtering the plurality of candidate area maps based on a distance between a first vector corresponding to the actual depth data map and a first vector corresponding to the candidate area map, and expanding the range of the filtered candidate area maps within the positioning area to obtain a candidate area map; Based on the preset neighborhood size, the covariance matrix of each grid cell in the selected area map and the actual depth data map is obtained, the eigenvalues ​​of each covariance matrix are calculated, and the linearity value, flatness value and scattering value of each grid cell are calculated based on the eigenvalues, and the linearity value, flatness value and scattering value are assigned to the corresponding grid cell; Based on a preset extraction window, a plurality of matching area maps of the size of the extraction window are extracted from the candidate area map and the actual depth data map respectively, wherein grid cells having a positive vertical coordinate value of a center of gravity, a depth parameter, a linear value, a flatness value, and a scattering type value are set in the matching area; The matching area map is input into a preset second model to obtain a second vector corresponding to each matching area map, and the distance sum between the matching area map of the actual depth data map and the matching area map of each candidate area map is calculated based on the second vector, and the candidate area map is matched based on the distance sum.

2. The underwater terrain matching positioning method according to claim 1, characterized in that: The method further comprises the steps of: Obtaining a second vector corresponding to the matching area map of the center of gravity position of the actual depth data map; Calculating the distance between the second vector and the second vector of the matching area graph in the matched area graph to be selected; The position of the matching area map in the candidate area map with the smallest distance value is selected as the position of the actual depth data map.

3. The underwater terrain matching positioning method according to claim 1 or 2, characterized in that: The step of receiving the actual depth data map transmitted by the sonar device also includes the following steps: Obtain multiple sounding points closest to each sounding point, and calculate the average value of the depth parameters of the sounding point and the multiple sounding points closest to it; Determine whether the absolute value of the difference between the depth parameter of the sounding point and the average value is greater than a preset first threshold; If it is greater, the sounding point is determined to be an outlier, and the depth parameter of the sounding point is modified to the average of the depth parameters of the multiple sounding points closest to the sounding point.

4. The underwater terrain matching positioning method according to claim 1, characterized in that: The step of calculating the depth parameter of the center of gravity point based on the depth parameter of the first sounding point close to the center of gravity point comprises: Calculate the distance between the center of gravity and each sounding point in the first sounding point respectively; A weight is assigned to each of the first number of sounding points based on the distance value, and the weighted average value of the depth parameters of the first number of sounding points is calculated as the depth parameter of the centroid point.

5. The underwater terrain matching positioning method according to claim 4, characterized in that: The weighted average value of the depth parameter of the first sounding point is calculated using the following formula: s represents the weighted average value, n represents the value of the first number, i represents any of the sounding points of the first number, s i Indicates the depth parameter of the i sounding point, d i Indicates the distance between the sounding point i and the center of gravity.

6. The underwater terrain matching positioning method according to claim 1, characterized in that: The steps of calculating the linearity value, flatness value, and scattering value of each grid cell based on the eigenvalue include: Sort the obtained multiple eigenvalues ​​from large to small according to their numerical values; The three largest eigenvalues ​​are obtained, where eigenvalue one ≥ eigenvalue two ≥ eigenvalue three, and the linearity value, the flatness value, and the scattering type value are calculated based on the eigenvalue one, the eigenvalue two, and the eigenvalue three.

7. The underwater terrain matching positioning method according to claim 6, characterized in that: The linearity value, flatness value, and scattering value are calculated based on eigenvalue 1, eigenvalue 2, and eigenvalue 3 according to the following formula:

8. The underwater terrain matching positioning method according to claim 1, characterized in that: The distance sum between the matching area map of the actual depth data map and the matching area map of each candidate area map is calculated based on the second vector, and the step of matching the candidate area map based on the distance sum is as follows: Calculate the sum of the distances between the matching area map of the actual depth data map and the matching area map of each candidate area map respectively; The distance sum of the actual depth data map for each matching area map is added to obtain the distance sum, and the matching area map with the smallest distance sum is the final matching candidate area map.

9. An underwater terrain matching and positioning device, characterized in that: The apparatus comprises a computer device, the computer device comprising a processor and a memory, the memory storing computer instructions, the processor being configured to execute the computer instructions stored in the memory, and when the computer instructions are executed by the processor, the apparatus implements the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images