A collaborative mapping system for multiple unmanned boats

Abstract

The present disclosure provides a multi-unmanned boat collaborative mapping system, the system includes: an unmanned boat environment intelligent perception system, used to obtain point cloud data of the target area, state information and positioning information of the unmanned boat; an unmanned boat control system, used to control the navigation state and operation planning of the unmanned boat; a multi-unmanned boat collaborative planning formation system, used to control the navigation route of each unmanned boat and the data interaction between unmanned boats; an image data processing system, used to perform rigid and non-rigid registration on the point cloud data obtained by the unmanned boat environment intelligent perception system to generate mapping results; an unmanned boat communication system, used to communicate between unmanned boats and between unmanned boats and ground stations, and receive operating instructions for the unmanned boats. The application of this system effectively expands the mapping scope in complex marine environments, improves the efficiency of the collaborative mapping process, and reduces the error of image registration.

Description

Technical Field

[0001] The present disclosure relates to the field of image processing, and in particular to a collaborative mapping system for multiple unmanned boats. Background Art

[0002] As a new technology, unmanned boat collaboration can be widely used in a variety of tasks and play an important role due to its strong adaptability and flexible deployment, such as marine surveying and mapping, environmental monitoring, resource exploration, etc. In order to realize the construction of environmental maps in complex marine scenes, a single unmanned boat is often used to perceive the environmental map and perform non-rigid splicing between frames for multi-frame point cloud mapping. However, if only a single unmanned boat is used for operation, the operation time cannot be guaranteed, and the mapping task of a specific sea area cannot be completed efficiently. In addition, the accuracy of mapping only through non-rigid splicing is low. Summary of the invention

[0003] The present disclosure provides a collaborative mapping system for multiple unmanned boats to at least solve the above technical problems existing in the prior art.

[0004] According to a first aspect of the present disclosure, a multi-unmanned boat collaborative mapping system is provided, the system comprising: an unmanned boat environment intelligent perception system, used to obtain point cloud data of a target area, status information of the unmanned boat and positioning information; an unmanned boat control system, used to control the navigation status and operation planning of the unmanned boat; a multi-unmanned boat collaborative planning formation system, used to control the navigation route of each unmanned boat and the data interaction between the unmanned boats; an image data processing system, used to perform rigid and non-rigid registration on the point cloud data obtained by the unmanned boat environment intelligent perception system to generate a mapping result; an unmanned boat communication system, used to communicate between unmanned boats and between unmanned boats and ground stations and receive operation instructions for the unmanned boats; the unmanned boat communication system comprises: a multi-unmanned boat self-organizing network communication system, used to ensure real-time update and correction of its own positioning during the multi-unmanned boat mapping process, and to transmit data between unmanned boats and between unmanned boats and ground stations; a data link communication system, used to ensure the real-time allocation of multi-unmanned boat mapping tasks to each unmanned boat strategy, and to transmit the collected data to a host.

[0005] In one possible implementation manner, the unmanned boat environment intelligent perception system includes: a point cloud laser scanning radar sensor module, used to construct point cloud data for the target area map; a gyroscope and stabilization sensor module, used to detect the posture changes and motion state of the unmanned boat, and measure the motion state information of the acceleration and angular velocity of the unmanned boat; a positioning module, used to obtain the position information and navigation information of the unmanned boat.

[0006] In one possible implementation manner, the unmanned boat control system includes: an unmanned boat central control module, used to control the motion state of each unmanned boat and the operation plan of the unmanned boat when performing a mission; an unmanned boat power control module, used to control the change of the heading state of the unmanned boat; and an unmanned boat stabilization system control module, used to control the point cloud laser scanning radar sensor module and the navigation posture of the unmanned boat based on the motion state information of the unmanned boat.

[0007] In one possible implementation, the multi-unmanned boat collaborative planning formation system includes: an unmanned boat task planning module, which is used to assign tasks and plan paths for unmanned boats during the collaborative mapping process of multiple unmanned boats; an inter-boat communication module, which is used to exchange data and share information between unmanned boats when performing designated navigation tasks according to the unmanned boat task planning module; and a multi-unmanned boat formation control and coordination module, which is used to ensure the relative position and speed synchronization of the unmanned boats in the formation, and to dynamically adjust the formation of the unmanned boats in the formation.

[0008] In one possible implementation, the image data processing system includes: an inter-frame algorithm registration module, which is used to perform similarity analysis between consecutive frames and rigidly and non-rigidly register the target with the terrain based on the video sequence obtained by the unmanned boat environment intelligent perception system during the unmanned boat mapping process, to construct a three-dimensional scene model and obtain a first mapping result; a multi-machine image stitching algorithm module is used to integrate the image data of multiple unmanned boats according to the first mapping result to generate a second mapping result.

[0009] In one possible implementation mode, the image data processing system is specifically used to obtain point cloud data of different frames of the same unmanned boat to be registered or images of different unmanned boats; extract boundary features, normal vector features and surface neighborhood curvature features of the point cloud data; determine rigid registration weights and non-rigid registration weights based on the boundary features, the normal vector features and the surface neighborhood curvature features; perform rigid registration and non-rigid registration on the point cloud data respectively to obtain rigid registration results and non-rigid registration results; generate a mapping result based on the rigid registration result, the non-rigid registration result, the rigid registration weight and the non-rigid registration weight.

[0010] In one possible implementation manner, the point cloud data is subjected to rigid registration and non-rigid registration respectively to obtain a rigid registration result and a non-rigid registration result, including: performing rigid registration on the point cloud data according to an iterative closest point algorithm and a singular value decomposition algorithm to obtain a rigid registration result; performing non-rigid registration on the point cloud data according to a Gaussian mixture model and a maximum expectation algorithm to obtain a non-rigid registration result.

[0011] In one possible implementation, the image data processing system is further configured to update the rigid registration weight and the non-rigid registration weight based on the mapping result.

[0012] In one possible implementation manner, the system further includes: a central station, which is used to process the data acquired by the unmanned boat environment intelligent perception system and upload the data to the cloud.

[0013] The disclosed multi-unmanned boat collaborative mapping system includes an unmanned boat environment intelligent perception system for obtaining point cloud data of the target area, state information of the unmanned boat and positioning information; an unmanned boat control system for controlling the navigation state and operation planning of the unmanned boat; a multi-unmanned boat collaborative planning formation system for controlling the navigation route of each unmanned boat and the data interaction between the unmanned boats; an image data processing system for rigidly registering and non-rigidly registering the point cloud data obtained by the unmanned boat environment intelligent perception system to generate a mapping result; and an unmanned boat communication system for communicating between unmanned boats and between unmanned boats and ground stations and receiving operation instructions for unmanned boats. The application of this system effectively expands the mapping scope in complex marine environments, effectively and quickly realizes the coverage of specific areas, improves the efficiency of the collaborative mapping process, and reduces the error of image registration through the hybrid registration process of rigid registration and non-rigid registration. It has the advantages of strong adaptability, high flexibility, high mapping accuracy, low communication burden, and multi-boat collaborative cluster operation.

[0014] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The above and other objects, features and advantages of the exemplary embodiments of the present disclosure will become readily understood by reading the detailed description below with reference to the accompanying drawings. In the accompanying drawings, several embodiments of the present disclosure are shown in an exemplary and non-limiting manner, in which:

[0016] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts.

[0017] Figure 1 The module schematic diagram of a multi-unmanned boat collaborative mapping system according to an embodiment of the present disclosure is shown. Figure 1 ;

[0018] Figure 2 A schematic diagram of the process of generating mapping results by performing rigid registration and non-rigid registration by the image data processing system according to an embodiment of the present disclosure is shown;

[0019] Figure 3 The module schematic diagram of a multi-unmanned boat collaborative mapping system according to an embodiment of the present disclosure is shown. Figure 2 . DETAILED DESCRIPTION

[0020] In order to make the purpose, features, and advantages of the present disclosure more obvious and easy to understand, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present disclosure.

[0021] Figure 1 The module schematic diagram of a multi-unmanned boat collaborative mapping system according to an embodiment of the present disclosure is shown. Figure 1 .

[0022] like Figure 1 As shown, the multi-unmanned boat collaborative mapping system 10 includes: an unmanned boat environment intelligent perception system 11, which is used to obtain point cloud data of the target area, state information of the unmanned boat and positioning information; an unmanned boat control system 12, which is used to control the navigation state and operation planning of the unmanned boat; a multi-unmanned boat collaborative planning formation system 13, which is used to control the navigation route of each unmanned boat and the data interaction between unmanned boats; an image data processing system 14, which is used to perform rigid and non-rigid registration on the point cloud data obtained by the unmanned boat environment intelligent perception system 11 to generate a mapping result; rigid registration refers to a registration method in which the spatial transformation is a rigid transformation, that is, the spatial transformation includes a translation transformation and a rotation transformation, so as to keep the distance between any two points unchanged after the transformation; non-rigid registration refers to restoring the corresponding relationship between the image with non-rigid deformation in spatial position and anatomical structure by finding a certain nonlinear spatial transformation. An unmanned boat communication system 15 is used to communicate between unmanned boats and between unmanned boats and ground stations and receive operation instructions for unmanned boats. Among them, the unmanned boat communication system 15 includes: a multi-unmanned boat self-organizing network communication system, which is used to ensure the real-time update and correction of its own positioning during the multi-unmanned boat mapping process, and to transmit data between the unmanned boat and other unmanned boats or between the unmanned boat and the ground station; a data link communication system, which is used to ensure the real-time allocation of strategies for each unmanned boat in the multi-unmanned boat mapping task, and can efficiently and quickly transmit the data collected in the ocean mapping task to the host to support the unmanned boat's task execution and real-time remote decision-making. In this application, the 5G data link communication system is used preferentially.

[0023] The present application obtains relevant data and information of multiple unmanned boats for water mapping through an unmanned boat environment intelligent perception system 11, controls the navigation status and operation plan of each unmanned boat through an unmanned boat control system 12, realizes collaborative control of the unmanned boat formation through a multiple unmanned boat collaborative planning formation system 13, and transmits the point cloud data obtained by the unmanned boat environment intelligent perception system 11 to an image data processing system 14. The image data processing system 14 performs rigid and non-rigid registration on the point cloud data to generate corresponding mapping results. Among them, the communication between unmanned boats and between unmanned boats and ground stations, as well as the reception of unmanned boat operation instructions are realized by the unmanned boat communication system 15. The multiple unmanned boat collaborative mapping system 10 effectively expands the mapping scope in complex marine environments, improves the efficiency of the collaborative mapping process, and reduces image registration errors.

[0024] In one possible implementation, the unmanned boat environment intelligent perception system 11 includes:

[0025] The point cloud laser scanning radar sensor module is used to construct point cloud data for the target area map. By forming a formation of several unmanned boats equipped with point cloud laser scanning radar sensors, the unmanned boats can obtain point cloud data of their corresponding areas, so that collaborative surface mapping operations can be completed in a certain formation.

[0026] The gyroscope and stabilization sensor module is used to detect the posture changes and motion status of the unmanned boat. During the point cloud data construction process, the acceleration, angular velocity and other motion status information of the unmanned boat are measured to ensure that the unmanned boat reaches its designated position and that the cameras, sensors and other equipment carried by the unmanned boat remain stable during the movement of the unmanned boat, so as to ensure good mapping effects of the unmanned boat.

[0027] The positioning module is used to obtain the location information and heading information of the unmanned boat. In order to make the data location obtained by the system in navigation and map construction more accurate, the positioning module can adopt a high-precision global positioning module.

[0028] In one embodiment, the unmanned boat control system 12 includes:

[0029] The unmanned boat central control module is used to control the motion state of each unmanned boat in the system when performing tasks, and to control the unmanned boat's collaborative mapping, obstacle avoidance, cluster formation and other operational planning.

[0030] The unmanned boat power control module is used to control the changes in the heading state of the unmanned boat, such as the navigation speed and heading angle of the unmanned boat.

[0031] The unmanned boat stabilization system control module is used to control the point cloud laser scanning radar sensor module and the navigation posture of the unmanned boat based on the motion state information of the unmanned boat, so as to reduce the disturbance to the hull caused by waves, etc., ensure the stability of real-time mapping, effectively control roll, bow pitch, etc., and ensure stable and accurate heading when performing tasks. The motion state information of the unmanned boat is obtained through the data feedback from the gyroscope and stabilization sensor module.

[0032] In one possible implementation, the multi-unmanned boat collaborative planning formation system 13 includes:

[0033] The unmanned boat mission planning module is used to assign tasks and plan paths for unmanned boats during the collaborative mapping process of multiple unmanned boats, formulate mission objectives and plans within the entire domain, and ensure that the mapping of the entire domain can be covered. The unmanned boat mission planning module can calculate the statistical credibility of the maximum detection distance of a single point cloud laser scanning radar sensor based on the parameters of the point cloud laser scanning radar sensor carried by the unmanned boat, and obtain the corresponding detection distance based on the confidence level, where the confidence level can be selected according to the actual situation, such as 90%, 95%, etc. In order to ensure a complete and accurate mapping effect, the corresponding planning task of each unmanned boat can be assigned according to the detection distance under a certain confidence level. The distance between adjacent unmanned boats should not be less than the detection distance to ensure high-precision mapping coverage in the entire area. For example, when performing tasks that do not require high mapping accuracy, such as hydrological measurement, the confidence level can be set to 90%, and the formation of the unmanned boat cluster can be set according to the detection distance corresponding to the confidence level of 90%; when performing tasks that require high mapping accuracy, such as marine surveying or search and rescue missions, the confidence level can be set to 95%, and the formation of the unmanned boat cluster can be set more densely to ensure navigation safety and the accuracy of measurement and exploration. The specific formation and parameter settings can be adjusted in real time according to the actual sea area and application.

[0034] The inter-boat communication module is used to exchange data and share information between unmanned boats when performing designated navigation tasks according to the unmanned boat mission planning module, ensuring that the unmanned boats in the formation can communicate in real time and that the corresponding control parameters are updated in a timely manner.

[0035] Among all the unmanned boats in the formation, at least one unmanned boat serves as the leader. Distributed control is adopted between the remaining unmanned boats and the leader unmanned boat, and communication and data exchange are realized through the inter-boat communication module to complete the update of control parameters and ensure the flexibility and stability of the formation. When the leader unmanned boat fails or the signal is interrupted, the unmanned boat central control module can re-control the remaining unmanned boats in the formation, establish a new leader unmanned boat in the formation and complete subsequent instructions.

[0036] The multi-unmanned boat formation control and coordination module is used to ensure the relative position and speed synchronization of the unmanned boats in the formation through event-triggered distributed control rate, and to achieve dynamic adjustment of the formation of unmanned boats in the formation to adapt to mapping tasks in different sea areas. It reduces energy consumption and communication burden while reducing communication frequency, and reduces calculation amount and adjustment parameters to achieve multi-machine consistency.

[0037] In one embodiment, the image data processing system 14 includes:

[0038] The inter-frame algorithm registration module is used in the process of unmanned boat mapping. Based on the intelligent perception system of the unmanned boat environment, further, the similarity analysis between the continuous frames of the video sequence obtained by the point cloud laser scanning radar sensor module and the rigid and non-rigid mixed registration of the target and the terrain are performed to build a more complete and detailed three-dimensional scene model. The inter-frame algorithm registration module determines the relative position relationship between the point cloud data and the overlapping area between the images by comparing the feature points or areas between consecutive frames, and obtains the first mapping result.

[0039] The multi-machine image stitching algorithm module is used to integrate the image data from multiple unmanned boats according to the first mapping result to obtain a second mapping result. The second mapping result is a target image and terrain map with a larger range and higher resolution.

[0040] That is to say, the inter-frame algorithm registration module first performs inter-frame registration based on different frames of the same unmanned boat. Based on the results of the inter-frame registration, the multi-machine image stitching algorithm module then aligns and fuses the image data collected from different unmanned boats to form a larger and more continuous image or map.

[0041] In one possible implementation, Figure 2 As shown, the image data processing system 14 performs rigid registration and non-rigid registration to generate a mapping result in the following process:

[0042] Step 201, obtaining point cloud data of different frames of the same unmanned boat or images of different unmanned boats to be registered;

[0043] Step 202, extracting boundary features, normal vector features and surface neighborhood curvature features of the point cloud data;

[0044] Step 203, determining a rigid registration weight and a non-rigid registration weight according to the boundary features, the normal vector features and the surface neighborhood curvature features;

[0045] Step 204, performing rigid registration and non-rigid registration on the point cloud data to obtain rigid registration results and non-rigid registration results;

[0046] Step 205: Generate a mapping result according to the rigid registration result, the non-rigid registration result, the rigid registration weight and the non-rigid registration weight.

[0047] After collecting point cloud data, the point cloud laser scanning radar sensor module uploads the point cloud data to the cloud, and the image data processing system performs mixed mapping of rigid registration and non-rigid registration to achieve fast and accurate large-scale map construction and target registration.

[0048] According to the distribution characteristics of point cloud, the feature information of point cloud is enhanced by clustering, and edge detection is performed by angle criterion algorithm (AC). For each point P, the K points nearest to it are searched as its neighborhood Np, and the neighborhood Np is projected onto its tangent plane to obtain the projected neighboring points. The neighboring points are connected with point P in pairs in a certain order with point P as the intersection point to form a series of angles , set the angle threshold k, when the maximum angle When it is larger than the angle threshold k, point P is determined as a boundary point, and then the boundary feature is obtained.

[0049] Since the point cloud plane is perpendicular to the normal vector, according to the spatial coordinate characteristics, the local neighborhood of each point in the point cloud can be fitted with a plane. For each point in the point cloud, the covariance matrix is ​​calculated using the neighborhood points, and the eigenvalues ​​are obtained by eigenvalue decomposition of the covariance matrix. The eigenvector corresponding to the minimum eigenvalue is the normal vector of point P, and the normal vector feature is obtained.

[0050] The covariance matrix is ​​decomposed into eigenvalues ​​to obtain eigenvalues, and the surface neighborhood curvature characteristics are obtained through the eigenvalues. The formula is: ,in, is the surface neighborhood curvature feature, , , are the three eigenvalues ​​of the covariance matrix of point P.

[0051] The weights of rigid registration and non-rigid registration are assigned according to the deformation degree or density distribution of boundary features and normal vector features. For example, areas with large local deformation or high point cloud density are given higher non-rigid registration weights; areas with small deformation are more suitable for rigid registration and are given higher rigid registration weights. It is understandable that the weights can be flexibly adjusted according to the actual mapping situation to ensure the best mapping effect. According to the weights assigned to the preset areas, different degrees of rigid registration and non-rigid registration are performed respectively, and the best mapping effect is obtained through iteration.

[0052] The rigid registration part, i.e., the rigid registration result and the rigid registration weight, and the non-rigid registration part, i.e., the non-rigid registration result and the non-rigid registration weight, are weightedly calculated to generate a mapping result.

[0053] In one possible implementation, rigid registration and non-rigid registration are performed on the point cloud data to obtain a rigid registration result and a non-rigid registration result, including:

[0054] The point cloud data is rigidly registered according to the iterative closest point algorithm and the singular value decomposition algorithm to obtain the rigid registration result;

[0055] The point cloud data is non-rigidly registered according to the Gaussian mixture model and maximum expectation algorithm to obtain the non-rigid registration result.

[0056] The rigid registration algorithm in this application uses a combination of the Iterative Closest Point (ICP) algorithm and the Singular Value Decomposition (SVD) algorithm to quickly and accurately achieve inter-frame registration of three-dimensional point clouds. The ICP algorithm is suitable for the registration of three-dimensional point cloud data. The optimal alignment between two frames of point clouds is confirmed by the ICP algorithm, and the transformation matrix of the ICP algorithm is solved by the SVD algorithm to improve the accuracy and stability of the registration. Select one point cloud as a reference point cloud and another point cloud as a source point cloud. For each source point cloud, find its nearest reference point cloud. Combined with the SVD algorithm, calculate the optimal transformation matrix of the source point cloud relative to the reference point cloud. The combination of the ICP algorithm and the SVD algorithm improves the robustness and accuracy of the registration algorithm. The input of the ICP algorithm is two groups of point clouds, the source point cloud and the reference point cloud. For each point in the source point cloud, the algorithm will find the point closest to the source point cloud in the reference point cloud, that is, the target point cloud, and establish a corresponding relationship between the source point cloud and the target point cloud to form a one-to-one point set. Furthermore, the KD tree can be used to speed up the search process.

[0057] The error function of the rigid transformation is: ,in, represents the total number of points in the source point cloud, represents the point corresponding to the source point cloud, represents the point corresponding to the target point cloud, is the rotation matrix, and t is the translation transformation. The rotation matrix R and the translation transformation t are solved by the SVD algorithm to minimize the error function. Finally, R and t are integrated into the transformation matrix and applied to the source point cloud for rigid transformation to obtain the transformed result point cloud. At the same time, the coordinates of each source point cloud after rigid transformation are obtained. .

[0058] The non-rigid registration algorithm uses a Gaussian mixture model (GMM) to describe the point set and identify local structures. Its input is two sets of point clouds, which are fitted into a GMM model. The fitted GMM is used as the initial state of the non-rigid registration. The EM algorithm is used to optimize the GMM combined with continuous simulated annealing. In the calculation method of the EM algorithm, each iteration is divided into two steps, one is the expectation step (E step), and the other is the maximum step (M step). In the E step, the expected value of the latent variable is estimated by the continuous simulated annealing method. In the M step, the likelihood function is maximized to update the model parameters to iteratively find the optimal transformation parameter v(y). The transformation formula is applied: , the parameter y is the three-dimensional coordinate of each point in the source point cloud. To achieve the best alignment of the image, the coordinates P of each source point cloud after non-rigid transformation are obtained. N , achieving efficient and accurate large-scale mapping tasks.

[0059] In one embodiment, the image data processing system 14 is further configured to update the rigid registration weight and the non-rigid registration weight based on the mapping result.

[0060] After performing rigid registration and non-rigid registration on the point cloud data respectively, this application integrates their respective transformation parameters and multiplies them with the corresponding weights, applies the final integration result to the original point cloud data to be registered for rigid and non-rigid mixed registration, and calculates the registration error and similarity error. The registration error and similarity error are used to update the transformation parameters and iterate continuously until the registration error is less than a certain threshold, which is considered to be converged, so as to achieve accurate and stable inter-frame registration and multi-machine image stitching.

[0061] Specifically, the integration steps of rigid registration and non-rigid registration are: according to the coordinates of each point in the source point cloud, the coordinates of the source point cloud after rigid transformation and the coordinates of the source point cloud after non-rigid transformation, the transformation vector is calculated, multiplied by the corresponding transformation weight, and the scaled vector is added, and finally applied to the source point cloud coordinates to obtain the transformed coordinates. The formula is as follows: ),in, is the source point cloud coordinate, is the coordinate of the source point cloud after rigid transformation, is the coordinate of the source point cloud after non-rigid transformation, is the weight corresponding to the non-rigid transformation, and its value is set by the similarity error, 0< <1.

[0062] In this application, the registration error uses the root mean square error (RMSE) as the error to measure the registration accuracy, and its formula is: , where N is the total number of points in the source point cloud, is the point in the target point cloud corresponding to the source point cloud, is a point in the source point cloud, is the corresponding transformation for each point in the source point cloud. When the registration error is less than a certain threshold, the registration is considered to be completed and the optimal registration effect is achieved. The key points are extracted by combining the boundary features and normal direction of the point cloud using the Normal Aligned Radial Feature (NARF) key point detector, and the weights of rigid registration and non-rigid registration are determined according to the curvature difference of the key points. The point cloud similarity error uses the surface neighborhood curvature difference of the key point as a measure of the deformation similarity between the transformed point cloud and the target point cloud. The formula is: ,in, represents the scaling parameter of the i-th key point, , Represents the surface curvature value of the i-th key point of the target point cloud, Indicates the curvature value of the key point corresponding to the source point cloud after hybrid registration transformation. The smaller the difference in curvature value, the smaller the deformation distance between the target point cloud and the transformed point cloud, the weight of rigid registration can be reduced, and the weight of non-rigid registration can be increased to improve robustness; the larger the difference in curvature value, the larger the deformation distance between the target point cloud and the transformed point cloud, the weight of non-rigid registration can be increased, and the weight of rigid registration can be reduced. After several rounds of iterations, the point cloud similarity error is integrated and calculated, and the weights of rigid registration and non-rigid registration are updated according to the point cloud similarity error until the registration error is lower than the preset threshold.

[0063] It is understandable that the weights of rigid registration and non-rigid registration can be adjusted in real time according to the actual environment and mapping effect to ensure the best mapping effect.

[0064] In one possible implementation, the system also includes a central station, which is located at the center of the unmanned boat cluster formation planned by the unmanned boat mission planning module. The central station can build an edge computing module in the data link communication system to perform preliminary processing on the data measured by the unmanned boat environment intelligent perception system and upload it to the cloud for subsequent processing. At the same time, it can also make decisions such as obstacle avoidance and planning based on real-time operation conditions to ensure the stability and security of the system when there are problems such as poor cloud network.

[0065] Figure 3 The module schematic diagram of a multi-unmanned boat collaborative mapping system according to an embodiment of the present disclosure is shown. Figure 2 .

[0066] like Figure 3As shown, the multi-unmanned boat collaborative mapping system of the present application includes an unmanned boat environment intelligent perception system 31, an unmanned boat control system 32, a multi-unmanned boat collaborative planning formation system 33, an image data processing system 34 and an unmanned boat communication system 35. The unmanned boat environment intelligent perception system 31 includes: a point cloud laser scanning radar sensor module 311, a gyroscope and stabilization sensor module 312 and a positioning module 313; the unmanned boat control system 32 includes: an unmanned boat central control module 321, an unmanned boat power control module 322 and an unmanned boat stabilization system control module 323; the multi-unmanned boat collaborative planning formation system 33 includes: an unmanned boat task planning module 331, an inter-boat communication module 332 and a multi-unmanned boat formation control and coordination module 333; the image data processing system 34 includes: an inter-frame algorithm registration module 341 and a multi-machine image stitching algorithm module 342; the unmanned boat communication system 35 includes: a multi-unmanned boat self-organizing network communication system 351 and a data link communication system 352.

[0067] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0068] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of the present disclosure, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0069] The above is only a specific embodiment of the present disclosure, but the protection scope of the present disclosure is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present disclosure, which should be included in the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure should be based on the protection scope of the claims.

Claims

1. A multi-unmanned boat collaborative mapping system, characterized in that: The system comprises: Unmanned boat environment intelligent perception system, used to obtain point cloud data of the target area, status information and positioning information of the unmanned boat; An unmanned boat control system, used to control the navigation status and operation planning of the unmanned boat; A multi-unmanned boat collaborative planning and formation system is used to control the navigation route of each unmanned boat and the data interaction between unmanned boats; An image data processing system is used to perform rigid registration and non-rigid registration on the point cloud data obtained by the unmanned boat environment intelligent perception system to generate a mapping result; The image data processing system is specifically used to obtain point cloud data of different frames of the same unmanned boat to be registered or images of different unmanned boats; extract boundary features, normal vector features and surface neighborhood curvature features of the point cloud data; determine rigid registration weights and non-rigid registration weights according to the boundary features, the normal vector features and the surface neighborhood curvature features; perform rigid registration and non-rigid registration on the point cloud data respectively to obtain rigid registration results and non-rigid registration results; generate a mapping result according to the rigid registration result, the non-rigid registration result, the rigid registration weight and the non-rigid registration weight; The point cloud data are subjected to rigid registration and non-rigid registration respectively to obtain rigid registration results and non-rigid registration results, including: performing rigid registration on the point cloud data according to an iterative nearest point algorithm and a singular value decomposition algorithm to obtain a rigid registration result; performing non-rigid registration on the point cloud data according to a Gaussian mixture model and a maximum expectation algorithm to obtain a non-rigid registration result; the non-rigid registration algorithm part uses a Gaussian mixture model to describe the point set and identify the local structure, and its input is two groups of point clouds, the point clouds are fitted into a Gaussian mixture model, and the fitted Gaussian mixture model is used as the initial state of non-rigid registration, and the Gaussian mixture model is optimized by the maximum expectation algorithm combined with continuous simulated annealing to register the Gaussian mixture model. In the calculation method of the maximum expectation algorithm, each iteration is divided into two steps: an expectation step and a maximum step. In the expectation step, the continuous simulated annealing method is used to estimate the expected value of the hidden variable, and in the maximum step, the maximum likelihood function is used to update the model parameters to iteratively find the optimal transformation parameters.

2. The system according to claim 1, characterized in that The unmanned boat environment intelligent perception system comprises: Point cloud laser scanning radar sensor module, used to construct point cloud data for target area maps; The gyroscope and stabilization sensor module is used to detect the attitude change and motion state of the unmanned boat, and measure the motion state information of the acceleration and angular velocity of the unmanned boat; The positioning module is used to obtain the position information and navigation information of the unmanned boat.

3. The system according to claim 2, characterized in that The unmanned boat control system comprises: The unmanned boat central control module is used to control the motion status of each unmanned boat and the operation plan of the unmanned boat when performing tasks; An unmanned boat power control module, used to control the change of the heading state of the unmanned boat; The unmanned boat stabilization system control module is used to control the point cloud laser scanning radar sensor module and the navigation posture of the unmanned boat based on the motion state information of the unmanned boat.

4. The system according to claim 1, characterized in that The multi-unmanned boat collaborative planning formation system includes: The unmanned boat mission planning module is used to assign tasks and plan paths for unmanned boats during collaborative mapping of multiple unmanned boats; An inter-boat communication module is used to exchange data and share information between unmanned boats when performing a designated navigation mission according to the unmanned boat mission planning module; The multi-unmanned boat formation control and coordination module is used to ensure the relative position and speed synchronization of the unmanned boats in the formation, and to dynamically adjust the formation of the unmanned boats in the formation.

5. The system according to claim 1, characterized in that The image data processing system comprises: An inter-frame algorithm registration module is used to perform similarity analysis between consecutive frames and rigidly and non-rigidly register the target with the terrain during the unmanned boat mapping process, based on the video sequence obtained by the unmanned boat environment intelligent perception system, to construct a three-dimensional scene model and obtain a first mapping result; The multi-machine image stitching algorithm module is used to integrate the image data of multiple unmanned boats according to the first mapping result to generate a second mapping result.

6. The system according to claim 1, characterized in that The image data processing system is further configured to update the rigid registration weight and the non-rigid registration weight based on the mapping result.

7. The system according to claim 1, characterized in that The system further comprises: The unmanned boat communication system is used for communication between unmanned boats and between unmanned boats and ground stations and for receiving operation instructions for the unmanned boats.

8. The system according to claim 7, characterized in that The unmanned boat communication system comprises: The multi-unmanned boat self-organizing network communication system is used to ensure the real-time update and correction of the positioning of the multi-unmanned boats during the mapping process, and to transmit data between the unmanned boats and between the unmanned boats and the ground station; The data link communication system is used to ensure the real-time allocation of strategies for each unmanned boat in the multi-unmanned boat mapping task, and to transmit the collected data to the host.

9. The system according to claim 1, characterized in that The system further comprises: The central station is used to process the data obtained by the unmanned boat environment intelligent perception system and upload it to the cloud.

Images