Underwater inspection robot and inspection method thereof

Abstract

The invention discloses an underwater inspection robot and an inspection method thereof. The underwater inspection robot comprises a rack; the driving assembly is arranged on the rack and used for driving the rack to move underwater; the detection cover is mounted on the rack, and a detection groove with an opening in one end is formed in the detection cover; the transparent sealing cover is arranged on the side, provided with the opening, of the detection cover and used for blocking the detection groove; the image pick-up device is arranged on the detection cover and located in the detection groove; the processor is arranged on the detection cover and located in the detection groove. And through the combination of the detection cover and the transparent sealing cover, water can be effectively isolated, and a better working environment is provided for the camera and the processor. And the processor reconstructs a three-dimensional model of the hydraulic structure according to the image obtained by the camera, so that the actual state of the hydraulic structure is analyzed, the safety degree of the hydraulic structure is evaluated, and the safety of the hydraulic structure is ensured.

Description

Technical Field

[0001] The invention relates to the technical field of underwater inspection, in particular to an underwater inspection robot and an inspection method thereof. Background Art

[0002] Underwater inspection robots are automated devices designed specifically for inspection, maintenance and monitoring in underwater environments. These robots are capable of performing a variety of tasks in oceans, rivers, lakes and other waters, including but not limited to inspection of submarine cables and pipelines, health monitoring of hydraulic structures, management of marine ranches, and search and rescue operations.

[0003] When using underwater inspection robots to monitor the health of hydraulic structures, it refers to the process of using various monitoring technologies and equipment to collect real-time or regular data on hydraulic structures, reconstructing the three-dimensional model of the hydraulic structure through a computer, and analyzing the actual state of the hydraulic structure and evaluating its safety. The existing three-dimensional model reconstruction process of hydraulic structures has low accuracy, which leads to low precision of the three-dimensional model of the hydraulic structure, affecting the analysis results. Summary of the invention

[0004] The technical problem to be solved by the present invention is: the present invention provides an underwater inspection robot and an inspection method thereof to solve the problem of low accuracy of the reconstructed three-dimensional model of hydraulic structures.

[0005] The technical solution adopted by the present invention to solve the technical problem is: an underwater inspection robot, comprising: a frame; a driving component, the driving component is arranged on the frame, and is used to drive the frame to move underwater; a detection cover, the detection cover is installed on the frame, and a detection slot with an opening at one end is arranged in the detection cover; a transparent cover, the transparent cover is arranged on one side of the opening of the detection cover, and is used to block the detection slot; a camera, the camera is arranged on the detection cover, and is located in the detection slot; a processor, the processor is arranged on the detection cover, and is located in the detection slot.

[0006] The beneficial effects of the present invention are as follows: the frame is used to carry structures such as the drive assembly and the detection cover. The drive assembly is used to drive the frame to move, thereby driving the camera to move, so as to comprehensively photograph the hydraulic structure located underwater. The combination of the detection cover and the transparent cover can effectively isolate water and provide a better working environment for the camera and the processor. The processor reconstructs the three-dimensional model of the hydraulic structure based on the image obtained by the camera, thereby analyzing the actual state of the hydraulic structure and evaluating its safety level to ensure the safety of the hydraulic structure.

[0007] Preferably, the drive assembly comprises:

[0008] A plurality of vertical power sources are installed at intervals on the frame, and the vertical power sources are electrically connected to the processor;

[0009] A plurality of lateral power sources are installed on the frame at intervals, and the lateral power sources are electrically connected to the processor.

[0010] Preferably, it also includes:

[0011] The illuminators are evenly spaced and arranged on the transparent cover along the circumference, and the illuminators are electrically connected to the processor.

[0012] Preferably, it also includes:

[0013] A fastener, which is arranged between the transparent cover and the detection cover;

[0014] A sealing ring, which is arranged between the transparent cover and the detection cover;

[0015] The pressure seal is made of a flexible material. A cavity is arranged inside the pressure seal. The cavity is connected to a controller. The controller is electrically connected to the processor.

[0016] Based on another object of the present invention, the present invention provides the following technical solution: an inspection method of an underwater inspection robot, the inspection method uses the underwater inspection robot as described above, and the inspection method comprises the following steps:

[0017] S1. Obtain images of hydraulic structures within the inspection area through a camera;

[0018] S2, obtaining the corresponding MVS point cloud of the hydraulic structure according to the hydraulic structure image;

[0019] S3, obtaining corresponding multiple regional point clouds according to the MVS point cloud image of the hydraulic structure;

[0020] S4, segmenting the regional point cloud into multiple fine point clouds by a segmentation method;

[0021] S5. Generate a Delaunay triangulation based on multiple fine point clouds, splice the Delaunay triangulations, and fuse them to generate a three-dimensional model of the hydraulic structure.

[0022] Preferably, the method for obtaining the MVS point cloud in the above S2 is: processing the hydraulic structure image according to Context Capture photography software to obtain the hydraulic structure MVS point cloud.

[0023] Preferably, the above S3 specifically includes the following steps:

[0024] S31, drawing a regional outline on the surface of the hydraulic structure under water;

[0025] S32. Obtaining the regional point cloud according to the regional outline and the MVS point cloud.

[0026] Preferably, the above S4 specifically includes the following steps:

[0027] S41, preprocessing the regional point cloud;

[0028] S42, dividing the regional point cloud into n layers with equal width along the direction of the normal vector of the surface of the hydraulic structure;

[0029] S43, traversing the points in the regional point cloud to obtain point cloud sets P at different levels;

[0030] S44, determining the judgment value K of the point cloud sets P at different levels, and when K≤d, d is a threshold, merging the two-level point cloud sets;

[0031] When K>d, the outermost point cloud set is saved separately. After traversing all point cloud sets P, the regional point cloud is accurately divided into several fine point clouds.

[0032] Preferably, the above S41 includes the following steps:

[0033] S411, obtain the maximum brightness value R in the regional point cloud max ;

[0034] S412, adjusting the brightness value of the regional point cloud, the adjustment method is as follows:

[0035]

[0036] In the formula, R i ′ is the adjusted brightness value of the point in the regional point cloud, R i The brightness value of the points in the area point cloud before adjustment.

[0037] Preferably, the calculation method of the judgment value K in the above S44 is:

[0038]

[0039] In the formula, t represents the number of points in the outermost point cloud set, o represents the number of points in the point cloud set adjacent to the outermost point cloud set, and R j is the brightness value of the outermost point cloud set, R l is the brightness value of the point cloud set adjacent to the outermost point cloud set, p(x, y, z) is the spatial position of the outermost point cloud set, q(x, y, z) is the spatial position of the point cloud set adjacent to the outermost layer, and α is the weighting factor of the normalized image distance and geometric distance. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The figure is a three-dimensional structural diagram of an underwater inspection robot;

[0041] Figure 2 A top view of an underwater inspection robot;

[0042] Figure 3It is a cross-sectional view of an underwater inspection robot;

[0043] Figure 4 for Figure 3 A partial enlarged view of point A in the middle.

[0044] In the figure: a frame 1, a detection cover 2, a vertical power source 3, a fastener 4, a transparent cover 5, a lateral power source 6, an illuminator 7, a pressure seal 8, a sealing ring 9, a camera 10, and a mounting plate 11. DETAILED DESCRIPTION

[0045] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0046] In this article, the terms "upper, lower, inside, outside" and the like are established based on the positional relationships shown in the drawings. The corresponding positional relationships may also change depending on the drawings, and therefore cannot be understood as absolute limitations on the scope of protection; moreover, relational terms such as "first" and "second" and the like are merely used to distinguish one component from another with the same name, and do not necessarily require or imply any actual relationship or order between these components.

[0047] Embodiment 1

[0048] like Figure 1-4 As shown, this embodiment provides an underwater inspection robot, including: a frame 1; a driving component, which is arranged on the frame 1 and is used to drive the frame 1 to move underwater; a detection cover 2, which is installed on the frame 1, and a detection slot with an opening at one end is provided in the detection cover 2; a transparent cover 5, which is arranged on a side of the detection cover 2 where the opening is provided, and is used to block the detection slot; a camera 10, which is arranged on the detection cover 2 and is located in the detection slot; a processor, which is arranged on the detection cover 2 and is located in the detection slot.

[0049] Specifically, the frame 1 is used to carry structures such as the drive assembly and the detection cover 2. The drive assembly is used to drive the frame 1 to move, thereby driving the camera 10 to move, so as to comprehensively photograph the hydraulic structure located underwater. The combination of the detection cover 2 and the transparent cover 5 can effectively isolate water and provide a better working environment for the camera 10 and the processor. The processor reconstructs the three-dimensional model of the hydraulic structure based on the image obtained by the camera 10, thereby analyzing the actual state of the hydraulic structure and evaluating its safety level to ensure the safety of the hydraulic structure.

[0050] The hydraulic structure includes a dam and a partially underwater building. The processor is preferably a CPU. A mounting plate 11 is provided in the detection cover 2 for mounting the processor, the camera 10 and other structures.

[0051] Further, the driving assembly includes: a plurality of vertical power sources 3, the plurality of vertical power sources 3 are installed at intervals on the frame 1, and the vertical power sources 3 are electrically connected to the processor;

[0052] A plurality of lateral power sources 6 are installed at intervals on the frame 1, and the lateral power sources 6 are electrically connected to the processor. The vertical power source 3 is used to control the depth of the underwater inspection robot, and the lateral power source 6 is used to control the movement of the underwater inspection robot on the horizontal plane and the posture of the underwater inspection robot. By moving the underwater inspection robot along the height direction and the horizontal plane direction, the camera 10 of the underwater inspection robot can fully capture the image of the hydraulic structure. By changing the posture of the underwater inspection robot in real time, the underwater inspection robot can obtain the image of the hydraulic structure from multiple angles, so that the subsequent point cloud construction is closer to the actual hydraulic structure, and the reconstruction accuracy of the subsequent three-dimensional model of the hydraulic structure is improved. Preferably, there are four vertical power sources 3 arranged at intervals in a square shape on the frame 1, and there are also four lateral power sources 6. The four lateral power sources 6 include two pairs of lateral power sources 6. The two lateral power sources 6 of the same pair are arranged opposite to each other, and the two pairs of lateral power sources 6 are arranged crosswise. There are also multiple cameras 10, and they are installed at different angles to capture images of the hydraulic structure from more angles and improve the reconstruction accuracy of the subsequent three-dimensional model of the hydraulic structure. The specific structure of the lateral power source 6 and the vertical power source 3 is a mature existing foundation, and no excessive restrictions are made here. It is sufficient to generate thrust. Preferably, the lateral power source 6 and the vertical power source 3 both include a motor and a propeller connected to the motor.

[0053] Furthermore, it also includes: illuminators 7, which are evenly spaced along the circumference of the transparent cover 5 and are electrically connected to the processor. The illuminators 7 are used to enhance the brightness to avoid too low brightness underwater, which may cause the point cloud of some areas to be unable to be obtained, affecting the subsequent reconstruction process of the hydraulic structure.

[0054] Further, it also includes: a fastener 4, which is arranged between the transparent cover 5 and the detection cover 2; a sealing ring 9, which is arranged between the transparent cover 5 and the detection cover 2; and a pressure seal 8, which is made of a flexible material, and a cavity is arranged in the pressure seal 8, and the cavity is connected to a controller, and the controller is electrically connected to the processor. The fastener 4 is used to detachably install the transparent cover 5 on the detection cover 2, and the fastener 4 is preferably a bolt, the sealing ring 9 is preferably an o-ring, and the pressure seal 8 is a sealing capsule, and its material is preferably nylon. The pressure seal 8 is arranged between the transparent cover 5 and the detection cover 2, preferably, the fastener 4 is arranged at the outermost side, the pressure seal 8 is arranged in the middle, and the sealing capsule is arranged at the innermost side. By arranging the fastener 4, the pressure seal 8 and the sealing ring 9 in the above order, the sealing effect can be better improved. The pressure of the internal cavity of the pressure seal 8 is controlled by a controller, which is preferably a small compressor. The processor determines the water pressure according to the depth of the underwater inspection robot, and increases the pressure inside the pressure seal 8 through the controller to offset the water pressure, thereby preventing the sealing structure from being destroyed under the water pressure. Preferably, an adhesive is also provided at the sealing ring 9 of the transparent cover 5 and the detection cover 2 to better improve the sealing performance.

[0055] Embodiment 2

[0056] The same parts in this embodiment as those in the first embodiment are given the same figure marks, and the same text descriptions are omitted.

[0057] This embodiment discloses an inspection method of an underwater inspection robot. The inspection method uses the underwater inspection robot in Embodiment 1. The inspection method includes the following steps:

[0058] S1. Acquire the image of the hydraulic structure in the inspection area through the camera 10; first determine the inspection area according to the size of the hydraulic structure located underwater, and then determine the inspection route according to the image acquisition range of the camera 10 and the size of the hydraulic structure. The preferred distance between the camera 10 and the surface of the hydraulic structure is 30cm-50cm to obtain a clearer image. The underwater inspection robot moves along the inspection route to drive the camera 10 to fully acquire the image of the hydraulic structure; S2. Acquire the corresponding MVS point cloud of the hydraulic structure according to the image of the hydraulic structure; S3. Acquire the corresponding multiple regional point clouds according to the MVS point cloud image of the hydraulic structure; S4. Segment the regional point cloud into multiple fine point clouds through a segmentation method; S5. Generate a Delaunay triangulation based on multiple fine point clouds, splice the Delaunay triangulation, and fuse them to generate a three-dimensional model of the hydraulic structure.

[0059] Specifically, the MVS point cloud of the hydraulic structure is segmented, and multiple fine point clouds are separated. Finally, they are spliced ​​through the Delaunay triangulation network and fused to generate a three-dimensional model of the hydraulic structure to ensure the molding effect of the three-dimensional model of the hydraulic structure, thereby making the three-dimensional model of the hydraulic structure more refined, and better analyzing the actual status of the hydraulic structure, and evaluating its safety level to ensure the safety of the hydraulic structure.

[0060] Furthermore, the method for obtaining the MVS point cloud in the above S2 is: processing the hydraulic structure image according to the Context Capture photography software to obtain the MVS point cloud of the hydraulic structure.

[0061] Furthermore, the above S3 specifically includes the following steps: S31, drawing a regional outline on the surface of the hydraulic structure located underwater; the regional outline can be drawn manually, or the processor can divide the surface of the hydraulic structure located underwater by averaging; S32, obtaining a regional point cloud according to the regional outline and the MVS point cloud. With the regional outline as the boundary and the normal vector of the surface of the hydraulic structure as the direction, the points located inside the regional outline on the normal vector of the surface of the hydraulic structure are obtained to form a regional point cloud.

[0062] Furthermore, the above S4 specifically includes the following steps:

[0063] S41, preprocessing the regional point cloud;

[0064] S42, dividing the regional point cloud into n layers with equal width along the direction of the normal vector of the surface of the hydraulic structure;

[0065] S43, traversing the points in the regional point cloud to obtain point cloud sets P at different levels;

[0066] S44, determine the judgment value K of different levels of point cloud sets P, when, is the threshold, merge the two layers of point cloud sets; preferably, is 8cm 2 .

[0067] When the outermost point cloud set is saved separately, after traversing all point cloud sets P, the regional point cloud is precisely divided into several fine point clouds. Specifically, the judgment starts from the outermost point cloud set, and the judgment value K is used to determine whether the outermost point cloud set and the inner point cloud set adjacent to it are the same structure. If they are the same structure, they are merged, and the merged point cloud sets form a new outermost point cloud set. If they are not the same structure, the outermost point cloud set is saved separately, and the inner point cloud set adjacent to it is used as the new outermost point cloud set, so as to separate the point clouds that do not belong to the same structure, and finally splice them through the Delaunay triangulation to improve the fineness of the reconstructed three-dimensional model of the hydraulic structure.

[0068] Furthermore, the above S41 includes the following steps:

[0069] S411, obtaining the maximum brightness value in the regional point cloud; here, it refers to the brightness value of the point with the highest brightness;

[0070] S412, adjusting the brightness value of the regional point cloud, the adjustment method is as follows:

[0071]

[0072] In the formula, R i ′ is the adjusted brightness value of the point in the regional point cloud, R i The brightness value of the points in the regional point cloud before adjustment. In order to prevent the brightness from affecting the judgment value, the brightness is adjusted to the range of 0-255 to avoid the temperature of the brightness from causing inaccurate calculation of the judgment value.

[0073] Furthermore, the calculation method of the judgment value K in the above S44 is:

[0074]

[0075] In the formula, t represents the number of points in the outermost point cloud set, o represents the number of points in the point cloud set adjacent to the outermost layer, and R j is the brightness value of the outermost point cloud set, R j Specifically represents the brightness value of the jth point in the outermost point cloud, R l is the brightness value of the point cloud set adjacent to the outermost point cloud set, R l Specifically, it represents the brightness value of the lth point in the point cloud set adjacent to the outermost point cloud set, p(x, y, z) is the spatial position of the jth point in the outermost point cloud set, q(x, y, z) is the spatial position of the lth point in the point cloud set adjacent to the outermost point cloud set, and α is the weighting factor of the normalized image distance and geometric distance.

[0076] The above-mentioned technical features can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

Claims

1. An underwater inspection robot, characterized in that: include: Rack (1); A driving assembly, the driving assembly being arranged on the frame (1) and being used for driving the frame (1) to move underwater; A detection cover (2), the detection cover (2) being mounted on the frame (1), and a detection slot with an open end being arranged in the detection cover (2); A transparent cover (5), the transparent cover (5) being arranged on a side of the detection cover (2) where the opening is arranged, and being used to block the detection slot; A camera (10), the camera (10) being arranged on the detection cover (2) and located in the detection slot; A processor is arranged on the detection cover (2) and is located in the detection slot.

2. The underwater inspection robot according to claim 1, characterized in that: The drive assembly comprises: A plurality of vertical power sources (3), wherein the plurality of vertical power sources (3) are installed at intervals on the frame (1), and the vertical power sources (3) are electrically connected to the processor; A plurality of lateral power sources (6), wherein the lateral power sources (6) are installed at intervals on the frame (1), and the lateral power sources (6) are electrically connected to the processor.

3. The underwater inspection robot according to claim 1, characterized in that: Also includes: Illuminators (7), the illuminators (7) are evenly spaced and arranged on the transparent cover (5) along the circumferential direction, and the illuminators (7) are electrically connected to the processor.

4. The underwater inspection robot according to claim 1, characterized in that: Also includes: A fastener (4), wherein the fastener (4) is arranged between the transparent cover (5) and the detection cover (2); A sealing ring (9), wherein the sealing ring (9) is arranged between the transparent cover (5) and the detection cover (2); A pressure seal (8), wherein the pressure seal (8) is made of a flexible material, a cavity is arranged inside the pressure seal (8), the cavity is connected to a controller, and the controller is electrically connected to the processor.

5. A patrol method for an underwater patrol robot, the patrol method using the underwater patrol robot according to any one of claims 1 to 4, characterized in that: The inspection method comprises the following steps: S1. Obtain images of hydraulic structures within the inspection area through a camera; S2. Acquire the corresponding MVS point cloud of the hydraulic structure according to the hydraulic structure image; S3, acquiring corresponding multiple regional point clouds according to the MVS point cloud image of the hydraulic structure; S4, segmenting the regional point cloud into multiple fine point clouds by a segmentation method; S5. Based on the multiple fine point clouds, a Delaunay triangulation network is generated, and the Delaunay triangulation networks are spliced ​​and fused to generate a three-dimensional model of the hydraulic structure.

6. The inspection method of an underwater inspection robot according to claim 5, characterized in that: The method for obtaining the MVS point cloud in the above S2 is: processing the hydraulic structure image according to Context Capture photography software to obtain the hydraulic structure MVS point cloud.

7. The inspection method of an underwater inspection robot according to claim 5, characterized in that: The above S3 specifically includes the following steps: S31, drawing a regional outline on the surface of the hydraulic structure under water; S32. Obtaining the regional point cloud according to the regional outline and the MVS point cloud.

8. The inspection method of an underwater inspection robot according to claim 5, characterized in that: The above S4 specifically includes the following steps: S41, preprocessing the regional point cloud; S42, dividing the regional point cloud into n layers with equal width along the direction of the normal vector of the surface of the hydraulic structure; S43, traversing the points in the regional point cloud to obtain point cloud sets P at different levels; S44, determining the judgment value K of the point cloud sets P at different levels, and when K≤d, d is a threshold, merging the two-level point cloud sets; When K>d, the outermost point cloud set is saved separately, and after traversing all the point cloud sets P, the regional point cloud is accurately divided into a number of the fine point clouds.

9. The inspection method of an underwater inspection robot according to claim 8, characterized in that: The above S41 includes the following steps: S411, obtaining the maximum brightness value R in the regional point cloud max ; S412: adjusting the brightness value of the regional point cloud, the adjustment method is as follows: In the formula, R i ′ is the adjusted brightness value of the point in the regional point cloud, R i The brightness value of the points in the area point cloud before adjustment.

10. The inspection method of an underwater inspection robot according to claim 8, characterized in that: The calculation method of the judgment value K in the above S44 is: In the formula, t represents the number of points in the outermost point cloud set, o represents the number of points in the point cloud set adjacent to the outermost point cloud set, and R j is the brightness value of the outermost point cloud set, R l is the brightness value of the point cloud set adjacent to the outermost point cloud set, p(x, y, z) is the spatial position of the outermost point cloud set, q(x, y, z) is the spatial position of the point cloud set adjacent to the outermost layer, and α is the weighting factor of the normalized image distance and geometric distance.

Images