Underwater intelligent netting cleaning robot application method and system
Through digital twin technology and convolutional neural network, the intelligent mesh clothing cleaning robot optimizes the cleaning path and posture, solving the problems of low efficiency, high cost and harsh environment underwater cage cleaning, achieving efficient and accurate cleaning results.
Patent Information
- Application Number
- CN202411169398.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-24
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-08-24
AI Technical Summary
The cleaning of existing underwater cages relies on manual operations, with low efficiency, high cost, harsh environment, poor quality of automation equipment, high energy consumption and incomplete cleaning.
An intelligent mesh clothing cleaning robot combining digital twin technology and convolutional neural networks is used to build a cage model through a computer system, and position and stain recognition are used to optimize cleaning paths, combining SCADAS data acquisition and IoT edge gateway for Real-time analysis to realize adaptive posture adjustment and efficient cleaning.
It improves cleaning efficiency and quality, reduces labor costs, avoids the risk of manual operation in harsh environments, achieves efficient and accurate stain removal, and reduces the risk of energy consumption and excessive cleaning.
Smart Images

Figure CN119004705A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of robot technology, and in particular to an application method and system of an underwater intelligent net cleaning robot. Background Art
[0002] At present, the cleaning of underwater net cages mainly relies on manual cleaning, which is not only inefficient but also poses a threat to the safety and health of operators. In addition, existing automated cleaning equipment often has problems such as large mass, high energy consumption, and incomplete cleaning. In view of these shortcomings, the present invention proposes a new application method and system of underwater intelligent net cleaning robot. Summary of the invention
[0003] Based on the above problems, the purpose of the present invention is to provide an application method and system of an underwater intelligent net cleaning robot to solve the problems of low manual operation efficiency, high cost, and harsh working environment in existing underwater net cage cleaning operations. Through automation and intelligent technology, the cleaning efficiency and quality can be improved while reducing labor costs.
[0004] To achieve the above object, the present invention adopts the following technical solutions: An application method of an underwater intelligent net cleaning robot comprises the following steps: S1. Use digital twin technology to build a cage model on a computer as a virtual environment for robot cleaning work; S2, placing the underwater intelligent net cleaning robot into the water, importing the robot model into the computer, positioning the robot through the sensor on the robot, and automatically adjusting the relative position between the robot and the net cage; S3, the robot performs underwater inspection tasks, uses underwater cameras to capture stain image data of the object to be cleaned, and uses convolutional neural networks to accurately identify and segment the stain area, combining the geometric shape and size information of the object to intelligently calibrate the cleaning area; S4. Collect field data through SCADAS data acquisition equipment, including video stream, seawater flow rate, underwater pressure and robot posture, etc., prepare the collected data into a data set for digital twin model training, use integral reinforcement learning to train the digital twin model, and construct the value function as well as the control strategy and disturbance input at the initial moment. Use the IoT edge gateway for analysis and processing to update the physical and motion states of the digital twin model in real time; S5. According to the calibration level and shape of the attachments in the cleaning area, the cleaning process is simulated in the digital twin model through the path planning method, the cleaning path is optimized and simulation tests are performed, and the test results are converted into actual cleaning instructions; S6. When the water flow and pressure change, the underwater robot takes action based on the received control strategy and disturbance input, updates its own status, and feeds back the position, attitude, linear velocity, and angular velocity status data information to the digital twin model. Further real-time comparison is performed based on the fed-back status data to correct the digital twin, and the thruster power output is adjusted by selecting the optimal system design solution to achieve adaptive posture adjustment of the robot, ensuring the continuity and efficiency of the cleaning process.
[0005] As a preferred solution of the underwater intelligent net cleaning robot application method of the present invention, during the cleaning process, the propeller power is dynamically adjusted through the net force sensor in combination with the type and density distribution of the net cage attachments and the sea condition data, to ensure that the cleaning disc maintains a constant contact force with the surface to be cleaned, thereby achieving efficient cleaning.
[0006] As a preferred solution of the application method of the underwater intelligent net cleaning robot of the present invention, the SCADAS data acquisition device collects field data, including video stream, seawater flow rate, underwater pressure and robot posture, etc., and prepares the collected data into a data set, which includes: Geometric data: The geometric data of the net and underwater environment are obtained through laser scanning, sonar or other 3D imaging technologies, such as the size characteristics of the net, the size characteristics and structural characteristics of the underwater robot, the depth of the water, the volume of the water area, etc. Dynamic data: Use force sensors and inertial measurement units (IMUs) to obtain dynamic data of the robot and water flow and other environments, such as hydrodynamics, system dynamics, buoyancy, propeller propulsion performance, structural force analysis, etc. Operational data: Obtain the operational data of the robot when performing cleaning tasks, such as trajectory, speed, torque, robot input and output characteristics, etc.
[0007] As a preferred solution of the underwater intelligent net cleaning robot application method of the present invention, a convolutional neural network is used to accurately classify image data, and the best candidate area is screened through a non-maximum suppression algorithm to eliminate redundant information and ensure the accuracy of the cleaning process.
[0008] An underwater intelligent net cleaning robot system, using the above-mentioned underwater intelligent net cleaning robot application method, comprises: an underwater camera, mounted on the side of the robot, for capturing stain image data; A scraper brush plate is provided at the bottom of the robot and is used for cleaning operations; Lighting lamp, used to provide auxiliary lighting for underwater cameras; Posture detection module, including gyroscope, accelerometer, compass, depth sensor, temperature sensor, pressure sensor, etc., used to detect the posture of the robot; The drive module, which consists of thrusters driven by electronic controllers, is used to control the movement of the robot; Umbilical cable, used for communication between the underwater robot and the surface control console; The above-water control console is used to build a digital twin model, conduct simulation tests, and control the robot to move along the cleaning path and perform cleaning operations based on the stain information collected by the camera; The beneficial effects of the present invention are: The application method and system of the underwater intelligent net cleaning robot provided by the present invention, firstly, the digital twin technology constructs a net cage model on a computer as a virtual environment for the robot cleaning work, and the underwater robot is placed in the actual water area and the robot model is also imported into the virtual environment to achieve digital twin; secondly, the underwater cleaning robot conducts inspection, the camera will collect images, obtain image data of the characteristic parts and stain areas of the object to be cleaned, and perform image preprocessing; then, a convolutional neural network is used to identify and segment the image data of the stain area of the object to be cleaned; at the same time, the field data is collected through the SCADAS data acquisition device, and the data collected in the physical environment is mapped to the digital twin model in real time, including physical, motion, environment and other attributes. Combined with the prior information of the shape and size of the object to be cleaned, the system can automatically plan the optimal cleaning path, and perform simulation demonstration on the three-dimensional model to verify the feasibility of the path. The application method and system of the underwater intelligent net cleaning robot provided by the present invention, on the one hand, eliminate the need for manual work in a harsh cleaning environment, and on the other hand, the existing cleaning equipment is bulky, inflexible and energy-intensive, and the present invention improves the cleaning efficiency and intelligence level; a cleaning robot based on visual guidance can use image information to clean key parts and stained areas in a targeted manner, which can more effectively remove stains; the present invention uses a scraping brush plate at the bottom to achieve contact cleaning; the present invention controls the size of the cleaning force based on force feedback to ensure the contact cleaning pressure between the brush plate and the surface of the cleaned object, which can achieve high-quality cleaning effects and avoid excessive cleaning of the cleaned area by the brush plate, causing damage to the net surface and energy waste; according to the inclination angle of the net cage net, the propeller of the present invention enables the robot to fit the net to maintain or change to the best cleaning posture, and during the cleaning process, continuously adjusts the posture of the underwater robot so that its brush plate and the surface of the net cage attachment have the largest cleaning coverage area, thereby improving the cleaning efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the contents of the embodiments of the present invention and these drawings without paying any creative work.
[0010] Figure 1 It is a flow chart of an application method of an underwater intelligent net cleaning robot provided in a specific embodiment of the present invention. DETAILED DESCRIPTION
[0011] In order to make the technical problems solved by the present invention, the technical solutions adopted and the technical effects achieved clearer, the technical solutions of the embodiments of the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.
[0012] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance. Among them, the terms "first position" and "second position" are two different positions.
[0013] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection or a detachable connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0014] like Figure 1 As shown, this embodiment provides an application method of an underwater intelligent net cleaning robot, which is used to solve the problems of harsh underwater net cleaning environment, heavy workload, low efficiency, low cleaning quality and low intelligence, and includes the following steps: S1. Use digital twin technology to build a cage model on a computer as a virtual environment for robot cleaning work; S2, placing the underwater intelligent net cleaning robot into the water, importing the robot model into the computer, positioning the robot through the sensor on the robot, and automatically adjusting the relative position between the robot and the net cage; S3, the robot performs underwater inspection tasks, uses underwater cameras to capture stain image data of the object to be cleaned, and uses convolutional neural networks to accurately identify and segment the stain area, combining the geometric shape and size information of the object to intelligently calibrate the cleaning area; S4. Collect field data through SCADAS data acquisition equipment, including video stream, seawater flow rate, underwater pressure and robot posture, etc., prepare the collected data into a data set for digital twin model training, use integral reinforcement learning to train the digital twin model, and construct the value function as well as the control strategy and disturbance input at the initial moment. Use the IoT edge gateway for analysis and processing to update the physical and motion states of the digital twin model in real time; S5. According to the calibration level and shape of the attachments in the cleaning area, the cleaning process is simulated in the digital twin model through the path planning method, the cleaning path is optimized and simulation tests are performed, and the test results are converted into actual cleaning instructions; S6. When the water flow and pressure change, the underwater robot takes action based on the received control strategy and disturbance input, updates its own status, and feeds back the position, attitude, linear velocity, and angular velocity status data information to the digital twin model. Further real-time comparison is performed based on the fed-back status data to correct the digital twin, and the thruster power output is adjusted by selecting the optimal system design solution to achieve adaptive posture adjustment of the robot, ensuring the continuity and efficiency of the cleaning process.
[0015] This paper uses advanced digital twin technology, combined with the professional Simcenter software platform, to build an accurate three-dimensional digital twin model based on the detailed parameters and performance indicators of the underwater net cleaning robot. The model not only includes the physical structure of the robot, but also covers the characteristics of its operating environment - the aquaculture cage. Through rigid-flexible multi-body dynamics simulation, the model can accurately simulate the physical interaction between the robot and the complex underwater environment.
[0016] In order to ensure the real-time and accuracy of the model, the present invention uses SCADAS data acquisition equipment to collect key field data, including video streams, seawater flow rate, underwater pressure, and the robot's real-time posture, etc. These data are quickly analyzed and processed by the IoT edge gateway, and then integrated into the digital twin model in real time, dynamically updating the model's physical and motion properties to reflect actual working conditions.
[0017] Digital twin model establishment process: Use CAD software or 3D modeling tools to build a geometric model of the net and underwater environment. The following methods can be used: Point cloud processing: Use the ICP (Iterative Closest Point) algorithm to align the point cloud data with the known model to generate an accurate environmental model.
[0018] Meshing: Convert the geometric model into the mesh model required for finite element analysis for subsequent simulation and analysis.
[0019] 1) Design of digital twin model: (1) In the formula, Represents the system design solution, Representative subsystem design solutions, and , They are the characteristic parameters of the underwater robot’s size, buoyancy, resultant external force, water velocity, water temperature, water volume, liquid density, underwater robot’s position, posture, linear velocity, and angular velocity; 2) Net buoyancy expression of underwater robot: (2) In the formula, B and G represent the buoyancy and gravity of the underwater robot respectively. , ; 3) , , They are x axis, y axis, z Water velocity in the axial direction; 4) The expression of the total external force acting on the underwater robot is: (3) In the formula, and are the repulsive force and ocean current force on the underwater robot, and through computational fluid dynamics, Roughly considered as , For the ocean current force; 5) Input and output characteristics of underwater robots x With output speed y The approximate relationship is: (4) In the formula, the output speed , and the above cubic function relationship is obtained through the actual measurement of underwater robot operation data, and They represent the angular velocity, front-rear linear velocity, and upper-lower linear velocity of the underwater robot respectively, and the input-output relationship can be approximated by a cubic function relationship.
[0020] The kinematic and dynamic models of the underwater robot are established, taking into account the influence of underwater fluid dynamics.
[0021] First, a coordinate system is established to describe the movement of the robot, and then the kinematic model is established by deriving the rotation matrix of the coordinate system. Then, the force and torque relationship between the coordinate systems of the underwater robot is analyzed. Finally, appropriate parameter variables are selected to establish a more accurate dynamic model to more realistically simulate the actual movement of the robot underwater.
[0022] The displacement of the underwater robot on three axes in the geographic coordinate system is described as the longitudinal displacement x , lateral displacement y and vertical displacement z , the posture of rotation around the three axes is described as , , , also called roll angle, pitch angle and heading angle. The speed of the underwater robot moving on the three axes in the carrier coordinate system is described as the longitudinal speed , lateral speed and vertical speed The angular velocity around the three axes is described as , , In the carrier coordinate system, the three-axis axial force on the robot is described as , , The moments around the three axes are described as , , .
[0023] The robot kinematic model is a description of the underwater robot in two coordinate systems. The position of the underwater robot in the geographic coordinate system is described as , the attitude angle is expressed as , position and attitude angle vector for: (5) The linear velocity of the robot in the carrier coordinates is expressed as , the angular velocity is expressed as , linear velocity and angular velocity vectors It is expressed as: (6) For the robot's linear speed , the transformation matrix from the carrier coordinate system to the geographic coordinate system for: (7) because is an orthogonal matrix, we have: (8) For angular velocity , the transformation matrix from the carrier coordinate system to the geographic coordinate system for: (9) The kinematic model from the robot carrier coordinate system to the geographic coordinate system is obtained as follows: (10) The kinematic model from the geographic coordinate system to the carrier coordinate system is: (11) The underwater robot is a rigid body, ignoring the influence of the buoyancy rope. The six-degree-of-freedom dynamic model of the robot in the carrier coordinates can be described as: (12) Where: M is the inertia matrix of the underwater robot, , is the robot rigid body inertia matrix, is the additional mass matrix; is the Coriolis force and centripetal force matrix, , and are the Coriolis force and centripetal force matrices generated by the robot and the additional mass, respectively; is the hydrodynamic damping matrix; is the restoring force and torque matrix composed of gravity and buoyancy; is the thrust of the propeller.
[0024] In terms of data fusion and model calibration, the digital twin model is calibrated by fusing the sensor data in actual operation with the virtual model. Kalman filtering or "extended Kalman filtering (EKF)" is used to fuse data from different sensors and correct model errors.
[0025] (13) Where: represents the estimated value of the state at time k; The predicted estimated value of the state at time k; represents the Kalman gain at time k; The measured value at time k; represents the measurement matrix.
[0026] In the face of deformation or tilt of the net caused by changes in water flow speed and pressure, the present invention can achieve adaptive posture adjustment of the robot by adjusting the power output of the thruster to ensure the continuity and efficiency of the cleaning process. In order to improve the reliability of the system, the present invention also optimizes the fault detection and prevention mechanism to ensure that the robot can respond quickly when encountering abnormal situations. Through the combination of the digital twin model and the HEEDS fast optimization iterative algorithm, the system can continuously analyze and compare the collected operation data, automatically adjust and optimize the model, and enhance the robot's ability to adapt to complex environmental changes.
[0027] In terms of motion coordination control, the present invention uses virtual-reality mapping technology to reflect the changes in the posture of the net caused by environmental changes in real time in the digital model. Using the digital twin HEEDS algorithm and machine learning algorithm, the system can analyze and predict the best dynamic adjustment strategy and optimize the posture adjustment of the robot. By finely controlling the thrusters, the robot can adapt to changes in water flow and pressure, maintain close contact with the net, and ensure the cleaning effect. Use simulation platforms (such as Gazebo, ROS, etc.) to run synchronously with the digital twin model to simulate every step of the underwater robot in performing the cleaning task. Make real-time adjustments based on policy results and sensor data to ensure the accuracy of the robot in actual operation, using PID control or model predictive control (MPC) (14) in, is the error between the expected trajectory and the actual trajectory; Control input, which usually means any form of control signal applied to the system; is the proportional gain, which is used to adjust the controller's response to the current error; The integral gain is used to adjust the controller's response to error accumulation.
[0028] The underwater intelligent net cleaning robot system of the present invention can use high-definition image information to clean key parts and stained areas in a targeted manner by adopting a cleaning robot based on vision guidance. The scraping brush plate design at the bottom realizes contact cleaning, and the cleaning force control technology based on force feedback ensures the appropriate contact pressure between the brush plate and the surface of the object being cleaned, which not only ensures the cleaning quality, but also avoids excessive cleaning and potential damage to the cleaned area.
[0029] The underwater robot's propeller is not only used to adjust the robot's posture, but also responsible for power distribution to achieve the best cleaning effect. According to the tilt angle of the cage net and the characteristics of the attachments, the system can dynamically adjust the propeller's power output, so that the robot always maintains the best cleaning posture during the cleaning process, maximizes the cleaning coverage area, and thus improves the overall cleaning efficiency.
[0030] The design of the cleaning module takes into account the intelligent distribution of power of the brush disc. Combining the type characteristics and density distribution of the cage attachments, as well as real-time sea conditions, the system can study and implement adaptive cleaning strategies. Through the data feedback of the net force sensor, the system can dynamically adjust the power output of the propeller to ensure that the cleaning disc maintains a constant scraping force, achieving efficient and thorough cleaning results.
[0031] The contact force with the net during cleaning is obtained through the force sensor, and the cleaning force is adjusted to avoid damaging the net. The method of admittance control is used to achieve: (15) in, is the external force; M is the virtual mass; B is the virtual damping; K is the virtual spring stiffness.
[0032] According to the cleaning effect, adjust the robot's motion trajectory to ensure the maximum cleaning coverage and cleaning effect, and use genetic algorithm or particle swarm optimization algorithm for global optimization: (16) in, It refers to the maximum cleaning efficiency; It means minimizing energy consumption.
[0033] The control method and system of the underwater intelligent net cleaning robot provided by the present invention, on the one hand, solve the problems of low manual operation efficiency, high cost, and bad working environment existing in the existing underwater net cage cleaning operation, and improve the cleaning efficiency and quality through automation and intelligent technology, while reducing the labor cost; on the other hand, the existing cleaning equipment is bulky and inflexible, and has high energy consumption, and the present invention improves the cleaning efficiency and intelligence level; the cleaning robot based on visual guidance can use image information to clean key parts and stain areas in a targeted manner, and can remove stains more effectively; the present invention uses a scraping brush plate at the bottom to achieve contact cleaning; the present invention controls the size of the cleaning force based on the force feedback method to ensure the contact cleaning pressure between the brush plate and the surface of the cleaned object, which can achieve high-quality cleaning effect and avoid the brush plate from excessively cleaning the cleaned area to cause damage to the net surface and waste of energy; according to the inclination angle of the net of the net cage, the propeller of the present invention makes the robot fit the net to maintain or change to the best cleaning posture, and in the cleaning process, continuously adjusts the posture of the underwater robot so that the brush plate and the surface of the net cage attachment have the largest cleaning coverage area, thereby improving the cleaning efficiency.
[0034] Vision is used to identify and locate stains on the surface of the object to be cleaned and the key cleaning areas. Then, based on the visually identified object and the prior information of its shape and size, the corresponding cleaning path is selected for cleaning, thereby achieving accurate and high-quality cleaning operations.
[0035] The cage attachment images obtained by the camera are preprocessed, and the retinex algorithm is used to enhance the image to obtain the feature map and the target sample candidate frame. After the region of interest pooling layer and the trained fully connected classification network, target recognition is performed to obtain the attachment classification of the target image to be identified.
[0036] Identification process of areas with many attachments: First, use the camera to collect visual images, use median filtering and Gaussian filtering to reduce the noise of the image, and then use the single-scale retinex algorithm (SSR) to eliminate the brightness difference to obtain the pre-processed image; in order to improve the accuracy of target recognition and segmentation, a method based on convolutional neural network is used for target recognition and segmentation.
[0037] The stained area image obtained by the camera is preprocessed, and a deep convolutional neural network is used to obtain the feature map and the target sample candidate frame. After the region of interest pooling layer and the trained fully connected classification network, target detection is performed to obtain the stained area position information of the component object.
[0038] Perform median filtering on the target image. Median filtering is a nonlinear smoothing technique that sets the grayscale value of each pixel to the median of all the pixel grayscale values in a certain neighborhood window of the point. Median filtering is a nonlinear signal processing technique based on sorting statistics theory that can effectively suppress noise. The basic principle of median filtering is to replace the value of a point in a digital image or digital sequence with the median of the values of each point in a neighborhood of the point, so that the surrounding pixel values are close to the true value, thereby eliminating isolated noise points.
[0039] The specific method is to use a two-dimensional sliding template of a certain structure to sort the pixels in the board according to the size of the pixel value, and generate a monotonically rising (or falling) two-dimensional data sequence. The output of the two-dimensional median filter is g(x,y)=med{f(xk,yl),(k,l∈W)}, where f(x,y) and g(x,y) are the original image and the processed image respectively. W is a two-dimensional template, usually a 3*3, 5*5 area, and can also be different shapes, such as linear, circular, cross, circular, etc. The median filter has a good filtering effect on impulse noise, especially while filtering noise, it can protect the edge of the signal so that it is not blurred. The median filter method is very effective in eliminating salt and pepper noise.
[0040] Gaussian filtering is a linear smoothing filter that is suitable for eliminating Gaussian noise and is widely used in the noise reduction process of image processing. In layman's terms, Gaussian filtering is the process of weighted averaging the entire image. The value of each pixel is obtained by weighted averaging the value of itself and other pixels in the neighborhood. The specific operation of Gaussian filtering is: use a template (or convolution, mask) to scan each pixel in the image, and replace the value of the central pixel of the template with the weighted average gray value of the pixels in the neighborhood determined by the template. Most of the noise in the image is Gaussian noise, so Gaussian filters are also widely used. Gaussian filtering is a linear smoothing filter that is suitable for eliminating Gaussian noise. The specific operation of Gaussian filtering is: use a user-specified template (or convolution, mask) to scan each pixel in the image, and replace the value of the central pixel of the template with the weighted average gray value of the pixels in the neighborhood determined by the template.
[0041] One-dimensional Gaussian distribution:
[0042] Two-dimensional Gaussian distribution:
[0043] The Single Scale Retinex (SSR) algorithm is an image enhancement technology. The Retinex theory is based on the working principle of the human visual system, especially the processing mechanism of the retina on light and reflectivity. The main purpose of this algorithm is to improve the visual effect of the image, especially under low light conditions. The core idea of the Single Scale Retinex algorithm is to separate the brightness and reflectivity of the image. In natural scenes, the pixel value of the image contains not only the reflectivity information of the object, but also the lighting information. The Retinex algorithm attempts to restore the true reflectivity of the object from the image, thereby enhancing the visual effect of the image. It is particularly suitable for enhancing images taken under low light conditions and can significantly improve the brightness and contrast of the image.
[0044] Compute the log reflectance plot: R(x,y) = log(I(x,y)) Background subtraction:
[0045] Reflectivity restoration: Since the training based on convolutional neural network depends on high-quality images, and the image quality directly obtained by the camera is poor due to the complex and changeable cleaning scene environment, it is necessary to preprocess the image before inputting it into the neural network for training to obtain better recognition results. Image preprocessing includes: using median filtering and Gaussian filtering to reduce the noise of the image, and then using the single-scale Retinex algorithm to enhance the image brightness and saturation to obtain the preprocessed image. The process of identifying the target based on the convolutional neural network method mainly includes: first generating a series of target candidate areas, then using the convolutional neural network to extract the convolution features in the target image, and finally classifying the target image data, using non-maximum suppression on the adjusted sample candidate frame, removing the sample candidate frame with too much overlap, and obtaining the sample target candidate frame.
[0046] Convolutional neural networks have the characteristics of local perception, shared weights and multi-core convolution. The basic structure of a convolutional neural network consists of an input layer, a hidden layer, a fully connected layer and an objective function. The hidden layer includes several convolutional layers, activation function layers and pooling layers.
[0047] In computer vision, the input layer provides the network with raw image data. The convolution layer in the hidden layer extracts feature information from the image by performing convolution operations on the raw image data provided. The weights of the elements in the convolution kernel matrix are determined after the network is trained.
[0048] The activation function layer in the hidden layer takes the result of the convolution operation as input, with the purpose of introducing nonlinear factors to make the network more expressive. Common activation functions include sigmoid function, Tanh function and Relu function. The pooling layer is also called the downsampling layer, which is used to reduce the dimension of the feature map and the number of network calculation parameters, while allowing the network to extract more important features. In the pooling layer, the commonly used pooling operation methods are maximum pooling and average pooling.
[0049] The fully connected layer fuses the local feature information extracted by the hidden layer and outputs the final result to the classifier. The objective function is also the loss function, which is used to calculate the error between the predicted value output by the network and the true value provided. In regression tasks and classification tasks, the most commonly used objective functions are the quadratic loss function and the cross entropy loss function, respectively. Among them, formula (a) represents the quadratic loss function, and formula (b) represents the cross entropy loss function.
[0050] (a) Among them, W represents the weight vector of the network, b is the bias vector of the network, M represents the number of input sample data, x represents the input sample data, y is the true value of the sample data, L is the maximum number of network layers, and a LRepresents the predicted value of the network.
[0051] (b) Among them, represents the predicted value of the network output, and y (i) represents the sample category label 0 or 1 corresponding to the i-th group.
[0052] First, the convolutional neural network layer extracts convolutional features from the original image, then inputs the feature map into the RPN network for candidate box extraction, and finally classifies and regresses the candidate boxes to achieve target detection.
[0053] Cleaning path designs include but are not limited to straight lines and concentric circles to adapt to different environments and cleaning requirements, and achieve more stable and efficient robot control.
[0054] In a virtual environment, use Algorithm or Dijkstra algorithm generates the best cleaning path for the robot:
[0055] in, is the actual cost from the starting point to node n, is the estimated cost from node n to the goal point.
[0056] In practice, the path is adjusted dynamically to avoid obstacles using the Dynamic Windowing (DWA):
[0057] in It refers to the velocity component of the robot along the x-axis; It refers to the velocity component of the robot along the y-axis.
[0058] Note that the above are only preferred embodiments of the present invention and the technical principles used. Those skilled in the art will understand that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments and substitutions can be made by those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in more detail through the above embodiments, the present invention is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of the present invention, and the scope of the present invention is determined by the scope of the appended claims.
Claims
1. An application method of an underwater intelligent net cleaning robot, characterized in that: The following steps are involved: S1. Use digital twin technology to build a cage model on a computer as a virtual environment for robot cleaning work; S2, placing the underwater intelligent net cleaning robot into the water, importing the robot model into the computer, positioning the robot through the sensor on the robot, and automatically adjusting the relative position between the robot and the net cage; S3, the robot performs underwater inspection tasks, uses underwater cameras to capture stain image data of the object to be cleaned, and uses convolutional neural networks to accurately identify and segment the stain area, combining the geometric shape and size information of the object to intelligently calibrate the cleaning area; S4. Collect field data through SCADAS data acquisition equipment, including video stream, seawater flow rate, underwater pressure and robot posture, etc., prepare the collected data into a data set for digital twin model training, use integral reinforcement learning to train the digital twin model, and construct the value function as well as the control strategy and disturbance input at the initial moment. Use the IoT edge gateway for analysis and processing to update the physical and motion states of the digital twin model in real time; S5. According to the calibration level and shape of the attachments in the cleaning area, the cleaning process is simulated in the digital twin model through the path planning method, the cleaning path is optimized and simulation tests are performed, and the test results are converted into actual cleaning instructions; S6. When the water flow and pressure change, the underwater robot takes action based on the received control strategy and disturbance input, updates its own status, and feeds back the position, attitude, linear velocity, and angular velocity status data information to the digital twin model. Further real-time comparison is performed based on the fed-back status data to correct the digital twin, and the thruster power output is adjusted by selecting the optimal system design solution to achieve adaptive posture adjustment of the robot, ensuring the continuity and efficiency of the cleaning process.
2. The application method of the underwater intelligent net cleaning robot according to claim 1, characterized in that: Digital twin technology is used for parametric modeling, and three-dimensional digital models of the net cleaning robot and aquaculture cages are constructed. The physical interaction between the robot and the cages is accurately simulated, and underwater high-definition cameras are used to capture specific landmarks to assist in positioning. The cleaning path is intelligently planned, and the feasibility of the path is verified through simulation.
3. The application method of the underwater intelligent net cleaning robot according to claim 1, characterized in that: During the cleaning process, the propeller power is dynamically adjusted through the net force sensor based on the type and density distribution of the cage attachments and sea condition data to ensure a constant contact force between the cleaning disc and the surface to be cleaned, thereby achieving efficient cleaning.
4. The application method of the underwater intelligent net cleaning robot according to claim 1, characterized in that: The dataset includes: Geometric data: The geometric data of the net and underwater environment, such as the size characteristics of the net, the size characteristics of the underwater robot, the structural characteristics, the water depth, the water volume, etc., are obtained through laser scanning, sonar or other 3D imaging technologies; Dynamics data: Use force sensors and inertial measurement units to obtain dynamics data of robots and water flow environments, such as hydrodynamics, system dynamics, buoyancy, propeller propulsion performance, and structural force analysis; Operational data: Obtain the trajectory, speed, torque, robot input and output characteristics and other operational data of the robot when performing cleaning tasks.
5. The application method of the underwater intelligent net cleaning robot according to claim 1, characterized in that: Convolutional neural networks are used to accurately classify image data, and the best candidate areas are screened through non-maximum suppression algorithms to eliminate redundant information and ensure the accuracy of the cleaning process.
6. An intelligent cleaning robot system, characterized in that: The application method of the intelligent net clothing cleaning robot according to any one of claims 1 to 5 comprises: an underwater camera, mounted on the side of the robot, for capturing stain image data; A scraper brush plate is provided at the bottom of the robot and is used for cleaning operations; Lighting lamp, used to provide auxiliary lighting for underwater cameras; Posture detection module, including gyroscope, accelerometer, compass, depth sensor, temperature sensor, pressure sensor, etc., used to detect the posture of the robot; The drive module, which consists of thrusters driven by electronic controllers, is used to control the movement of the robot; Umbilical cable, used for communication between the underwater robot and the surface control console; The water control console is used to build a digital twin model, conduct simulation tests, and control the robot to move along the cleaning path and perform cleaning operations based on the stain information collected by the camera.
Citation Information
Patent Citations
Intelligent cleaning robot path planning method and system
CN113012149A
Method for building digital twinborn monitoring system of automobile engine cylinder block production line
CN116300741A
Underwater robot formation control method based on digital twinning and reinforcement learning
CN116679733A
Automatic cleaning device and cleaning method for high-rise building facade
CN117356991A
Cited By
Bulkhead cleaning method
CN119389381A
Intelligent drilling deflection detection robot and drilling deflection correction method applied by same
CN120350948A
Method and system for evaluating anus cleaning effect
CN120761617A
Energy control system and method of underwater robot
CN120972600A
Stainless steel pipe fitting cleaning method and device based on automatic control
CN121165556A
