Intelligent cutting device for underwater canal slope lining plate and construction method

Abstract

The invention provides an intelligent cutting device for an underwater canal slope lining plate and a construction method. The device comprises an underwater self-adaptive walking platform, a three-dimensional visual recognition module, a cutting execution module, a buoyancy and pressure adjusting module and a remote control module. The intelligent cutting device for the underwater canal slope lining plate integrates three-dimensional laser scanning and algorithm recognition, the cutting range can be automatically recognized, manual judgment is not needed, the limitation of manual control of a simple robot is broken through, and the construction efficiency is improved; a triangular crawler belt of the underwater self-adaptive walking platform is matched with buoyancy adjustment of a buoyancy and pressure adjusting module, the underwater self-adaptive walking platform can be matched with a 1: 1.5-1: 3 variable-gradient channel slope and the water depth of 1-3 m, and the problem that a traditional robot slips and rolls over is solved. Closure pumping drainage is not needed, underwater cutting is directly carried out, the technical problem that water supply is affected by dry field repair is solved, and the pumping drainage cost is saved; no person enters the whole process, and safety accidents of divers are eliminated.

Description

Technical Field

[0001] This invention belongs to the technical field of underwater cutting equipment, specifically relating to an intelligent cutting device and construction method for underwater channel slope lining plates. Background Technology

[0002] Underwater canal slope lining slabs, mostly made of 10-30cm thick concrete, are the core protective structure of irrigation and water conveyance canals, preventing water leakage and slope erosion. After long-term operation, the lining slabs are prone to damage such as cracks, spalling, and hollowing due to freeze-thaw cycles, water flow impact, and foundation settlement. If not repaired in time, the damage will expand and may even cause the canal slope to collapse, affecting water conveyance efficiency and project safety.

[0003] Currently, the core challenge in repairing damaged underwater channel slope lining panels is achieving precise cutting of the damaged area. Existing cutting technologies are mainly divided into three categories:

[0004] Manual underwater cutting involves divers carrying handheld cutting tools such as pneumatic cutting saws to work underwater. They need to manually judge the extent of the damage and manually control the cutting trajectory. Cutting a single 2m×2m lining plate takes ≥2 hours. Underwater visibility is low, water flow is strong, cutting accuracy is poor, and there are risks of drowning and electric shock for divers.

[0005] Dry excavation and repair involves temporarily diverting water and pumping out the channel. Then, a ground cutting machine is used to cut the damaged area. Although this method is highly precise, it requires water to be shut off for 3 to 7 days, which seriously affects agricultural irrigation or urban water supply. In addition, the cost of pumping out water from the channel is high.

[0006] Small underwater robot cutting uses a simple tracked underwater robot equipped with a cutting blade, which is manually operated by a shore-based remote control. However, it lacks autonomous damage recognition capabilities and relies on shore personnel to determine the cutting range through a camera. Furthermore, the track has poor adaptability to canal slopes and is prone to slipping. The cutting depth cannot be automatically controlled, and it is easy to cut through the lining plate or make incomplete cuts.

[0007] Therefore, it is necessary to design an underwater cutting device to solve the current technical problems. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides an intelligent cutting device and construction method for underwater channel slope lining plates that can improve cutting accuracy and efficiency, enable safe and unmanned operation, and not affect water supply.

[0009] The technical solution of this invention is: an intelligent cutting device for underwater channel slope lining plates, comprising:

[0010] An underwater adaptive walking platform includes a platform body, with triangular tracked chassis on both sides of the platform body, and an attitude sensing unit mounted on the platform body;

[0011] A three-dimensional vision recognition module is installed on the main body of the platform, including an underwater camera with an artificial light source, a laser scanner and an image processing unit. The image processing unit is used to fuse laser point cloud data and optical image data to reconstruct a three-dimensional model of the damaged area of ​​the lining plate and identify the cutting boundary.

[0012] The cutting execution module is located at the end of the platform body and includes a vibration cutting assembly, a driving component, and a pressure sensor. The driving component is used to drive the vibration cutting assembly to move up and down to adjust the cutting depth, and the pressure sensor is used to detect the contact pressure between the vibration cutting assembly and the lining plate.

[0013] The buoyancy and pressure adjustment module includes buoyancy airbags evenly distributed on the platform body and an inflation / deflation control unit. The inflation / deflation control unit is used to adjust the buoyancy of the buoyancy airbags to control the positive pressure of the device on the canal slope surface.

[0014] The remote control module is connected to the underwater adaptive walking platform, the three-dimensional vision recognition module, the cutting execution module, and the buoyancy and pressure adjustment module, and is used to realize remote monitoring of the cutting operation status, cutting task planning, and construction data storage.

[0015] Furthermore, the image processing unit is configured as follows:

[0016] Based on the fused 3D point cloud model and optical image data, the damaged areas on the surface of the lining plate are automatically identified and the cutting boundaries are marked through a pre-trained image recognition model.

[0017] The remote control module is configured to receive data from the attitude sensing unit and transform the cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system of the underwater adaptive walking platform.

[0018] Furthermore, the remote control module includes a shore-based control terminal and an underwater equipment terminal; the shore-based control terminal is communicatively connected to the underwater equipment terminal mounted on the platform body via a wired or wireless data transmission module; the underwater equipment terminal is used to collect and upload data from various underwater modules, and simultaneously receive and execute instructions from the shore-based control terminal.

[0019] Furthermore, a suction component is provided next to the cutting execution module to remove debris generated during the cutting process in real time, so as to avoid debris interfering with the field of view of the underwater camera or affecting the cutting accuracy.

[0020] Furthermore, the triangular track chassis includes a track frame, with two support wheels symmetrically arranged at the bottom of the track frame and a drive wheel at the top. Tracks are fitted on the outer sides of the support wheels and the drive wheel, and rubber anti-slip teeth are provided on the outer side of the track.

[0021] Furthermore, the track is internally equipped with permanent magnets that are adsorbed onto steel sheets embedded in the canal slope lining plate.

[0022] The intelligent cutting construction method for underwater channel slope lining slabs, using the cutting device described in any of the above-mentioned methods, includes the following steps:

[0023] Step 1, Data Acquisition: Control the underwater adaptive walking platform to move in the target area, and simultaneously acquire laser point cloud data and optical image data of the lining plate surface through the three-dimensional vision recognition module, and reconstruct a three-dimensional model with texture information by the image processing unit.

[0024] Step 2, Data Processing: The image processing unit integrates the three-dimensional model with the optical image data, and automatically identifies the cracks, spalling and hollow areas on the surface of the lining plate through a pre-trained image recognition model, and marks the boundaries to be cut in the three-dimensional model;

[0025] Step 3, Path Planning: The remote control module receives the platform attitude data measured by the attitude sensing unit, transforms the marked cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system of the underwater adaptive walking platform, and automatically generates the platform walking path and cutting execution command that fits the canal slope surface.

[0026] Step 4, Cutting Construction: The remote control module controls the vibration cutting component to start and coordinates with the underwater adaptive walking platform to move along the generated walking path to perform precise cutting of the damaged area of ​​the lining plate.

[0027] Furthermore, in the data processing step, the pre-trained image recognition model is a semantic segmentation model based on a convolutional neural network; the model performs pixel-level classification on the fused data, outputs a segmentation map that identifies cracks, peeling, and hollow areas, and generates closed cutting boundaries based on the segmentation map.

[0028] Furthermore, in the path planning step, the remote control module combines the real-time data of the attitude sensing unit and transforms the cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system through coordinate transformation, generating a platform walking path composed of the platform position and attitude sequence and conforming to the canal slope.

[0029] Furthermore, in the cutting construction step, the remote control module controls the driving component to dynamically adjust the cutting depth based on the real-time contact pressure fed back by the pressure sensor, so that the real-time contact pressure is maintained within a preset pressure range.

[0030] When the pressure is above the upper limit of the range, the cutting depth is reduced; when the pressure is below the lower limit of the range, the cutting depth is increased.

[0031] The beneficial effects of this invention are:

[0032] (1) The underwater channel slope lining plate intelligent cutting device integrates three-dimensional laser scanning and algorithm recognition, which can automatically identify the cutting range without manual judgment, breaking through the limitations of simple robot manual operation and improving construction efficiency.

[0033] (2) The triangular tracks of the underwater adaptive walking platform, combined with the buoyancy adjustment module of the buoyancy and pressure adjustment module, can adapt to the 1:1.5 to 1:3 slope of the channel and the 1 to 3m water depth, solving the problem of traditional robots slipping and overturning.

[0034] (3) No need for interception and drainage, direct underwater cutting solves the technical problem of dry site repair affecting water supply and saves drainage costs; no one goes underwater throughout the process, eliminating the safety accidents of divers.

[0035] (3) The cutting depth is preset by laser three-dimensional scanning and adjusted in real time by pressure sensor feedback, reducing trajectory error and depth error, far exceeding the accuracy of manual and simple robot. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the intelligent cutting device for underwater channel slope lining plates in this invention.

[0037] Figure 2 This is a schematic diagram of the intelligent cutting device for underwater channel slope lining plates in this invention.

[0038] Figure 3 This is a flowchart of the intelligent cutting construction method for underwater channel slope lining plates in this invention. Detailed Implementation

[0039] Various exemplary embodiments of the invention will now be described in detail with reference to the accompanying drawings. The descriptions of the exemplary embodiments are merely illustrative and are in no way intended to limit the invention or its application or use. The invention can be embodied in many different forms and is not limited to the embodiments described herein. These embodiments are provided to make the invention thorough and complete, and to fully express the scope of the invention to those skilled in the art. It should be noted that, unless otherwise specifically stated, the relative arrangement of components and steps, the composition of materials, numerical expressions, and values ​​set forth in these embodiments should be interpreted as merely exemplary and not as limiting.

[0040] The terms "first," "second," and similar words used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different parts. Words such as "including" or "comprising" mean that the element preceding the word encompasses the element listed after it, without excluding the possibility of encompassing other elements. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0041] like Figure 1 and 2 As shown, an intelligent cutting device for underwater channel lining slabs is disclosed, comprising: an underwater adaptive walking platform 1, including a platform body 12, with triangular tracked chassis on both sides of the platform body 12, and an attitude sensing unit mounted on the platform body; a three-dimensional vision recognition module 2, disposed on the platform body 12, including an underwater camera 21 with an artificial light source, a laser scanner 22, and an image processing unit, wherein the laser scanner 22 is used to collect laser point cloud data, the underwater camera 21 is used to collect optical image data, and the image processing unit is used to fuse the laser point cloud data and the optical image data to reconstruct a three-dimensional model of the damaged area of ​​the lining slab and identify the cutting boundary; and a cutting execution module 3, disposed at the end of the platform body 12, including... The system includes a vibration cutting assembly 31, a drive component 32, and a pressure sensor. The drive component 32 drives the vibration cutting assembly 31 to rise and fall to adjust the cutting depth. The pressure sensor detects the contact pressure between the vibration cutting assembly 31 and the lining plate. A buoyancy and pressure adjustment module 4 includes buoyancy airbags 41 evenly distributed on the platform body and an inflation / deflation control unit. The inflation / deflation control unit adjusts the buoyancy of the buoyancy airbags to control the positive pressure exerted by the device on the channel slope surface. A remote control module 5 is connected to the underwater adaptive walking platform 1, the three-dimensional vision recognition module 2, the cutting execution module 3, and the buoyancy and pressure adjustment module 4. This module enables remote monitoring of the cutting operation status, planning of cutting tasks, and storage of construction data.

[0042] The underwater channel slope lining plate intelligent cutting device in the above embodiments integrates three-dimensional laser scanning and algorithm recognition, which can automatically identify the cutting range without manual judgment, breaking through the limitations of manual operation of simple robots and improving construction efficiency. The triangular tracks of the underwater adaptive walking platform 1, together with the buoyancy adjustment module 4, can adapt to channel slopes with varying gradients of 1:1.5 to 1:3 and water depths of 1 to 3 meters, solving the problem of slippage and rollover of traditional robots. It can cut directly underwater without diversion and drainage, solving the technical problem of dry site repair affecting water supply and saving drainage costs. The entire process is unmanned, eliminating the safety accidents of divers. The cutting depth is preset by laser three-dimensional scanning and adjusted in real time by pressure sensors, reducing trajectory error and depth error, far exceeding the accuracy of manual operation and simple robots.

[0043] In some embodiments, the image processing unit is configured to: automatically identify the damaged area on the surface of the lining plate and mark the cutting boundary based on the fused three-dimensional point cloud model and optical image data through a pre-trained image recognition model; the remote control module 5 is configured to: receive the data from the attitude sensing unit and transform the cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system of the underwater adaptive walking platform.

[0044] Specifically, the image processing unit fuses the 3D point cloud model with optical image data to create a 3D model with realistic texture. A pre-trained image recognition model analyzes this fused model and identifies damaged areas. Then, it extends a certain distance on both sides of the damaged area, marking a rectangular cutting boundary in the 3D model. This boundary exists in a virtual "3D model coordinate system." The remote control module receives data from the attitude sensor to determine the platform's current attitude on the slope. Using a coordinate transformation algorithm, it converts the virtual "cutting boundary" from the image processing unit to the robot's actual world coordinate system. After the transformation, a platform walking path is finally generated. This path ensures that when the underwater adaptive walking platform moves along this path, its end-mounted vibration cutting component can accurately complete the rectangular boundary, and because the underwater adaptive walking platform's attitude is taken into account, the cutting trajectory perfectly matches the canal slope.

[0045] In some embodiments, the remote control module 5 includes a shore control terminal and an underwater equipment terminal; the shore control terminal is communicatively connected to the underwater equipment terminal mounted on the platform body via a wired or wireless data transmission module; the underwater equipment terminal is used to collect and upload data from various underwater modules, and simultaneously receive and execute instructions from the shore control terminal.

[0046] As an example, the onshore control terminal uses an industrial tablet PC, while the underwater equipment terminal uses a data acquisition device. The onshore control terminal and the underwater equipment terminal communicate with each other via a 4G / 5G communication module.

[0047] In some embodiments, a suction component is provided next to the cutting execution module 3 to remove debris generated during the cutting process in real time, preventing debris from interfering with the underwater camera's field of vision or affecting the cutting accuracy. Specifically, the suction component has a suction force ≥10kPa and a flow rate of 5L / min.

[0048] In some embodiments, the triangular track chassis includes a track frame 14, with two support wheels 16 symmetrically arranged at the bottom of the track frame 14 and a drive wheel 15 arranged at the top. A track 11 is fitted on the outer side of the support wheels 16 and the drive wheel 15, and rubber anti-slip teeth 13 are provided on the outer side of the track 11. Specifically, the track 11 is 150mm wide, the rubber anti-slip teeth 13 have a tooth height of 5mm and a spacing of 10mm to adapt to a slope of 1:1.5 to 1:3 to prevent rollover. The drive wheel 15 is driven by a servo motor with a power of 200W, an adjustable speed of 0 to 30rpm, and an IP68 protection rating, which can meet the requirements of long-term operation in water depths of 1 to 3m.

[0049] In some embodiments, the track is provided with permanent magnets that are adsorbed onto steel sheets embedded in the lining plate of the channel slope to prevent slippage caused by water flow impact; specifically, the magnetic force of the permanent magnets is ≥500N.

[0050] In some embodiments, the platform body 12 is a 316L stainless steel box with a diameter of 1200mm × 800mm × 500mm, which is resistant to seawater corrosion and has an underwater service life of ≥5 years.

[0051] In some embodiments, the attitude sensing unit includes a three-axis gyroscope and a tilt sensor to monitor the contact angle between the platform and the canal slope in real time. By adjusting the difference in track speed, the platform is automatically kept level, ensuring that the vibration cutting assembly 31 is perpendicular to the lining plate.

[0052] In some embodiments, such as Figure 3 As shown, an intelligent cutting construction method for underwater channel slope lining slabs is disclosed, employing the cutting device as described in any of the above embodiments, and including the following steps:

[0053] Step 1, Data Acquisition: Control the underwater adaptive walking platform to move in the target area, and simultaneously acquire laser point cloud data and optical image data of the lining plate surface through the three-dimensional vision recognition module, and reconstruct a three-dimensional model with texture information by the image processing unit.

[0054] Step 2, Data Processing: The image processing unit integrates the three-dimensional model with the optical image data, and automatically identifies the cracks, spalling and hollow areas on the surface of the lining plate through a pre-trained image recognition model, and marks the boundaries to be cut in the three-dimensional model;

[0055] Step 3, Path Planning: The remote control module receives the platform attitude data measured by the attitude sensing unit, transforms the marked cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system of the underwater adaptive walking platform, and automatically generates the platform walking path and cutting execution command that fits the canal slope surface.

[0056] Step 4, Cutting Construction: The remote control module controls the vibration cutting component to start and coordinates with the underwater adaptive walking platform to move along the generated walking path to perform precise cutting of the damaged area of ​​the lining plate.

[0057] In some embodiments, in the data processing step, the pre-trained image recognition model is a semantic segmentation model based on a convolutional neural network; the model performs pixel-level classification on the fused data, outputs a segmentation map identifying cracks, peeling, and hollow areas, and generates closed cutting boundaries based on the segmentation map.

[0058] Specifically, the system will apply textures. Figure 3 The 3D model is input into a pre-trained convolutional neural network semantic segmentation model, which classifies the model surface: areas with dark linear features are classified as "cracks," areas with missing surface material and exposed internal rough structures are classified as "scratches," and intact concrete surfaces are classified as "background." After processing, the model outputs a segmentation map, which is no longer a color photograph but a "map" coded with different colors, where red lines represent identified cracks and blue blocks represent identified scraped areas. For cracks, the outline is regularly expanded on both sides to generate a rectangular cutting boundary with uniform width that completely encloses the crack. For scraped areas, the smallest bounding rectangle that can completely cover the irregular area is calculated as its cutting boundary. Finally, a set of cutting boundaries composed of several closed polygons is generated. These boundaries accurately define all the damaged concrete areas that need to be removed, providing precise input for subsequent automatic path planning.

[0059] In some embodiments, during the path planning step, the remote control module combines the real-time data from the attitude sensing unit and transforms the cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system through coordinate transformation, thereby generating a platform walking path consisting of the platform position and attitude sequence and conforming to the canal slope.

[0060] Specifically, the coordinate transformation steps are as follows:

[0061] Rotation transformation: Based on the platform pitch angle, roll angle and yaw angle provided by the attitude sensing unit, the transformation matrices for rotation around the X, Y and Z axes are calculated respectively; these three matrices are combined to form a total rotation matrix R; it defines how the "up" direction of the model corresponds to the "gravity and slope" direction in the real world;

[0062] Translation transformation: Based on the initial registration position of the 3D model in the world, a translation vector T is determined, which defines the specific position of the model's "origin" in the real world;

[0063] Perform matrix multiplication: for each cutting boundary point P defined in the 3D model model The algorithm calculates the corresponding point P in the walking coordinate system for the coordinate system (x, y, z, 1) using a single matrix multiplication. global (X,Y,Z,1); its core formula can be simplified to: P global =M×P model Here, M is the comprehensive transformation matrix, which is composed of the previously calculated rotation matrix R and translation vector T. This operation transforms the local coordinates of the point into global coordinates in one operation through rotation and translation.

[0064] Path sequence generation: The algorithm performs the above transformations on all path points on the cutting boundary one by one. Finally, the output is a control sequence consisting of these transformed global coordinate points and the attitude that the platform should have at that point.

[0065] In some embodiments, during the cutting process, the remote control module controls the drive component to dynamically adjust the cutting depth based on the real-time contact pressure fed back by the pressure sensor, so that the real-time contact pressure is maintained within a preset pressure range; wherein, when the pressure is higher than the upper limit of the range, the cutting depth is reduced, and when the pressure is lower than the lower limit of the range, the cutting depth is increased; as an example, the lower limit threshold of the contact pressure is 20N, and the upper limit threshold of the contact pressure is 50N.

[0066] The technical solution of this embodiment will be further described below through a specific engineering example.

[0067] Taking an agricultural irrigation canal (water depth 1.5m, canal slope 1:2, lining slab of C30 concrete, thickness 20cm, damaged area is 1.5m×1.8m spalling zone) as an example, the specific implementation is as follows:

[0068] 1. Device parameter configuration

[0069] Walking platform: triangular track distribution, track magnetic force 500N, servo motor speed 15rpm, platform weight 80kg, buoyancy bag inflation capacity 3L (suitable for water depth 1.5m);

[0070] Vision module: The 3D laser scanner has a scanning range of 1.2m×1.5m, and the AI ​​algorithm has an accuracy of ≥98% in identifying damaged areas, with a cutting boundary marking error of ≤1mm;

[0071] Cutting system: High-frequency cutting blade frequency 40kHz, preset cutting depth 20cm (consistent with the thickness of the lining board), pressure sensor threshold 20-50N, vacuum cleaner suction power 10kPa;

[0072] Control system: wireless transmission distance of 800m, lithium battery capacity of 200Ah, and can cut 60㎡ of lining board on a single charge.

[0073] 2. Complete workflow (operation without interrupting water supply)

[0074] Device deployment: The device is lifted to the bank of the channel by a crane, the inflation volume of the buoyancy bag is adjusted (3L), the device is placed in the water, the magnetic track is activated, and it is attached to the surface of the channel slope (the tilt sensor shows that the platform level is ±0.5°, which meets the requirements);

[0075] Damage identification:

[0076] A 3D laser scanner scans the target area (3m×3m) to generate a three-dimensional point cloud model of the lining slab;

[0077] The underwater camera captures images, and the AI ​​algorithm automatically identifies the 1.5m×1.8m peeling and damage area, marks the rectangular cutting boundary (coordinates: X0-Y0 to X1.5-Y1.8), and converts the boundary into the platform's walking path;

[0078] Path planning: The control system plans the platform's movement sequence based on the path (from the upper left corner of the damaged area → upper right corner → lower right corner → lower left corner, closing the cut), setting the cutting depth to 20cm and the movement speed to 5cm / s; Intelligent cutting:

[0079] The platform moves along the path, the cutting blade starts (40kHz vibration), and the electric push rod controls the cutting depth to 20cm. The pressure sensor monitors the contact pressure in real time (maintained at 30-40N). When the local lining plate protrudes and the pressure rises to 55N, the system automatically reduces the cutting depth by 0.5cm to avoid cutting through.

[0080] The vacuum cleaner works simultaneously to remove concrete debris, and the camera transmits the cutting footage back in real time to ensure that the trajectory is without deviation.

[0081] Cutting verification: After cutting, the 3D scanner scans the cutting area again to confirm that the cutting boundary error is ≤2mm, the depth error is ≤1mm, the damaged area is completely separated, and there is no residue.

[0082] Device recovery: The control platform returns to the shore, the magnetic track is shut off, the gas in the buoyancy bag is released, and the cutting operation is completed by the crane recovery device (the whole process takes 12 minutes, without water interruption or manual entry into the water).

[0083] The various embodiments of the present invention have now been described in detail. To avoid obscuring the concept of the invention, some details known in the art have not been described. Those skilled in the art will fully understand how to implement the technical solutions disclosed herein based on the above description.

[0084] The embodiments described above only illustrate some implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. An underwater channel slope lining plate intelligent cutting device, characterized in that, The underwater adaptive walking platform comprises a platform main body, and triangular track chassis are arranged on both sides of the platform main body; the platform main body is provided with a posture sensing unit; A three-dimensional visual identification module is arranged on the platform main body and comprises an underwater camera with an artificial light source, a laser scanner and an image processing unit; the image processing unit is used for fusing laser point cloud data and optical image data, reconstructing a three-dimensional model of a damaged area of a lining plate and identifying a cutting boundary; A cutting execution module is arranged at an end of the platform main body and comprises a vibration cutting assembly, a driving component and a pressure sensor; the driving component is used for driving the vibration cutting assembly to ascend and descend to realize cutting depth adjustment; and the pressure sensor is used for detecting contact pressure of the vibration cutting assembly and the lining plate; A buoyancy and pressure adjusting module comprises buoyancy air bags arranged uniformly on the platform main body and a gas charging and discharging control unit; the gas charging and discharging control unit is used for adjusting buoyancy of the buoyancy air bags to control normal pressure of the device on a channel slope surface; A remote control module is connected with the underwater adaptive walking platform, the three-dimensional visual identification module, the cutting execution module and the buoyancy and pressure adjusting module and is used for realizing remote monitoring of a cutting operation state, planning of a cutting task and storage of construction data. The image processing unit is configured to:

2. The underwater trench slope lining board intelligent cutting device according to claim 1, characterized in that: based on the fused three-dimensional point cloud model and optical image data, automatically identify a damaged area on a lining plate surface and mark a cutting boundary through a pre-trained image recognition model; the remote control module is configured to receive data of the posture sensing unit and convert the cutting boundary from a three-dimensional model coordinate system to a walking control coordinate system of the underwater adaptive walking platform. The remote control module comprises a shore control terminal and an underwater equipment terminal; the shore control terminal is in communication connection with the underwater equipment terminal arranged on the platform main body through a wired or wireless data transmission module; and the underwater equipment terminal is used for collecting data of underwater modules and uploading, and receiving and executing instructions from the shore control terminal.

3. The underwater trench slope lining board intelligent cutting device according to claim 1, characterized in that: An suction component is arranged beside the cutting execution module and is used for removing generated debris in real time during cutting to avoid interference of the debris with a field of view of the underwater camera or influence on cutting precision.

4. The underwater trench slope lining board intelligent cutting device according to claim 1, characterized in that: The triangular track chassis comprises a track frame, two supporting wheels are symmetrically arranged at the bottom of the track frame, a driving wheel is arranged at the top of the track frame, a track is sleeved outside the supporting wheel and the driving wheel, and rubber anti-skid teeth are arranged outside the track.

5. The underwater trench slope lining board intelligent cutting device according to claim 1, characterized in that: Permanent magnets are arranged inside the track and are adsorbed to embedded steel sheets in a channel slope lining plate.

6. The underwater trench slope lining board intelligent cutting device according to claim 5, characterized in that: The method comprises the following steps:

7. An intelligent cutting construction method for underwater channel slope lining plate, using the cutting device according to any one of claims 1 to 6, characterized in that, Step 1, data acquisition: the underwater adaptive walking platform is controlled to move in a target area, laser point cloud data and optical image data on a lining plate surface are synchronously collected through the three-dimensional visual identification module, and a three-dimensional model with texture information is reconstructed by the image processing unit; ​ Step 2, data processing: the image processing unit fuses the three-dimensional model and optical image data, automatically identifies cracks, peeling and hollowing damage areas on the lining plate surface through a pre-trained image recognition model, and marks the cutting boundary in the three-dimensional model; Step 3, path planning: the remote control module receives the platform attitude data measured by the attitude sensing unit, converts the marked cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system of the underwater adaptive walking platform, automatically generates a platform walking path and cutting execution instruction that conforms to the channel slope surface, and Step 4, cutting construction: the remote control module controls the vibration cutting assembly to start, and cooperatively controls the underwater adaptive walking platform to move along the generated walking path to execute precise cutting on the damaged area of the lining plate.

8. The intelligent cutting construction method of the underwater channel slope lining plate according to claim 7, characterized in that: In the data processing step, the pre-trained image recognition model is a semantic segmentation model based on a convolutional neural network; the model classifies the fused data at the pixel level, outputs a segmentation map with cracks, peeling and hollowing areas, and generates a closed cutting boundary based on the segmentation map.

9. The intelligent cutting construction method of the underwater channel slope lining plate according to claim 7, characterized in that: In the path planning step, the remote control module combines real-time data from the attitude sensing unit, converts the cutting boundary from the three-dimensional model coordinate system to the walking control coordinate system through coordinate transformation, and generates a platform walking path composed of platform position and attitude sequences that conforms to the channel slope surface.

10. The intelligent cutting construction method of the underwater channel slope lining plate according to claim 7, characterized in that: In the cutting construction step, the remote control module dynamically adjusts the cutting depth according to the real-time contact pressure feedback from the pressure sensor, so that the real-time contact pressure is maintained within a predetermined pressure range; When the pressure is higher than the upper limit of the range, the cutting depth is reduced, and when the pressure is lower than the lower limit of the range, the cutting depth is increased.

Images