Automatic cleaning method based on netting cleaning device

By using a combination of net washing robot, network camera, lighting, water pressure sensor and ultra-short baseline positioning system in cage farming, the automatic cleaning of net clothes is achieved, solving the problems of cleaning omissions and manual intervention in the prior art. , improves cleaning efficiency and coverage.

CN118988801AActive Publication Date: 2024-11-22QINGDAO SENKETE INTELLIGENT INSTR CO LTD
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202410815971.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-24
Publication Date
2024-11-22
Estimated Expiration
2044-06-24

AI Technical Summary

Technical Problem

The lack of automated mesh cleaning solutions in existing cage farming leads to long-term concentration of operators, which is prone to cleaning omissions, and the existing semi-automatic cleaning methods still require manual intervention.

Method used

Using automatic cleaning methods including network washing robots, network cameras, lighting, water pressure sensors and ultra-short baseline positioning system, the robot can be automatically cleaned through predetermined cleaning paths, depth control modes and image tracking modes, combining closed-loop negative feedback control and Three-dimensional vision technology to reduce manual intervention.

Benefits of technology

Automatic cleaning of mesh clothing is realized, cleaning coverage and efficiency are improved, operator requirements are reduced, manpower is saved, and cleaning quality is ensured through warning mechanisms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118988801A_ABST
    Figure CN118988801A_ABST
Patent Text Reader

Abstract

The invention provides an automatic cleaning method based on a netting cleaning device.The netting cleaning device comprises a netting cleaning robot, network cameras, illuminating lamps, a water pressure sensor and an ultra-short baseline positioning system, the network cameras are located in front of and behind the netting cleaning robot respectively, and the illuminating lamps are located in front of and behind the netting cleaning robot respectively; the water pressure sensor is located on a sealed cabin of the net washing robot, the ultra-short baseline positioning system comprises a base station and a beacon, the base station is fixed to a frame of the net cage, the beacon is located on the net washing robot, and a net is installed on the frame of the net cage. According to the technical scheme, the automatic cleaning method is provided, and the requirement for automatic netting cleaning operation is met. The automatic net washing method of the net washing robot is achieved by combining the ultra-short baseline positioning technology, the machine vision technology and the airborne water pressure sensor, the requirements for operators are reduced, manpower is saved, and the coverage rate of washing is increased.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of seawater cage aquaculture, and in particular to an automatic cleaning method based on a net cleaning device. Background Art

[0002] In modern cage aquaculture, nets are used to enclose a certain underwater enclosed space. As time goes by, the nets provide a suitable environment for the growth of various algae, shellfish and other organisms. The permeability of the mesh is a key factor affecting water exchange and is related to whether the aquaculture products can grow normally. Therefore, net cleaning has become one of the important tasks in cage aquaculture. The traditional method of operating a net cleaning robot is to configure an operating station for the robot. The operator observes the monitoring video footage captured by the robot's camera and manipulates the operating lever to manually control the robot's cleaning operation. This method requires the operator to concentrate for a long time, and inexperienced operators are prone to missing cleaning.

[0003] At present, many companies at home and abroad have developed net cleaning products, and the cleaning technology is also developing rapidly. Mainstay Manta supports semi-automatic and remote control cleaning strategies. In semi-automatic mode, the robot runs at a constant speed. When encountering certain key points (such as turning points or edges), the operator can use the remote control to fine-tune the position and direction of the robot's operation, similar to the cruise control of a car. This method saves a certain amount of manpower. Domestic net cleaning is still mainly manual cleaning, and there is no mature commercial automatic cleaning solution.

[0004] As various technologies and design concepts are constantly being tried and evolved, intelligent and automated cleaning technology is an inevitable development trend in the industry.

[0005] The net-washing robot usually has a network camera, a lighting lamp, tracks and a propeller, and can crawl on the net or slide on the propeller. The net-washing robot usually has a net-washing device such as a brush or a high-pressure water washing disc, and cleans the net after the robot is close to the net.

[0006] The net-washing robot is equipped with an operation station to perform net-washing operations in the net box. The operation station includes a remote monitoring system, a host computer system for the net-washing robot, and a command system such as a rocker, buttons, and knobs for controlling the movement of the robot. The monitoring system can return the video of the network camera in time to see the underwater cleaning situation in real time.

[0007] During the net washing operation, the command system of the operation station issues a control command to rotate the propeller of the net washing robot and make the robot stick to the net. The monitoring screen can be used to determine whether the robot has stuck to the net. After the robot sticks to the net, the operator controls the robot to move forward through the command system of the operation station. The video screen can clearly show the traces of cleaning. Then, according to the established cleaning path, the command system is used to command the robot to clean in sequence until the cleaning is completed.

[0008] During the cleaning operation, firstly, due to the complex underwater environment, the robot is prone to go astray. Secondly, once the robot slips, it can only clean the net close to the net. This requires the operator to always monitor the video screen, stay alert at all times, and drive the robot on the predetermined route to cover the net as much as possible. The existing literature does not discuss in detail the method of automatic cleaning of the net-washing robot in the net cage. Remote control cleaning is the mainstream of cleaning operations now, but it is necessary to stay focused and constantly adjust the cleaning direction and speed. Although semi-automatic cleaning operations have appeared abroad, they still require intermittent manual intervention. There are no cases and methods of automatic cleaning in China. Summary of the invention

[0009] In order to make up for the deficiencies of the prior art, the present invention provides an automatic cleaning method based on a net cleaning device.

[0010] The present invention is realized by the following technical scheme: an automatic cleaning method based on a net cleaning device, the net cleaning device comprising a net washing robot, a network camera, a lighting lamp, a water pressure sensor and an ultra-short baseline positioning system, the network camera is respectively located in front and behind the net washing robot, the lighting lamp is respectively located in front and behind the net washing robot, the water pressure sensor is located on the sealed cabin of the net washing robot, and the ultra-short baseline positioning system comprises a base station and a beacon, and the beacon is located on the net washing robot;

[0011] The specific steps include:

[0012] Step S1, an ultra-short baseline base station is fixedly installed in the net box, and the base station is fixed on the frame of the net box;

[0013] Step S2: The program predetermines the cleaning path, completes the circular cleaning of one layer, moves diagonally downward, and increases the depth H of the position of the net washing robot, H≤H clean , H clean It is the actual cleaning width of the robot. When it is running, it will wash one layer clockwise and one layer counterclockwise, and then cycle in sequence until the whole net is cleaned;

[0014] Step S3: After the net-washing robot is close to the net near the water surface, the automatic cleaning mode is started, and the current position given by the ultra-short baseline positioning is recorded as the initial position P. s , at this time, the target depth H depthSet to 1, that is, maintain a water depth of 1 meter, perform the first layer of cleaning horizontally, use the depth control mode, use the closed-loop negative feedback control algorithm to control the robot's depth, and maintain it at H depth Within ±ΔH;

[0015] Step S4: Depth control mode specifically includes the following steps: current depth and target depth H depth Error of |H error |>ΔH, depth adjustment is performed, L now is the current head orientation of the net washing robot, according to H error Calculate the direction L that the head of the net washing robot needs to be adjusted to obj and the horizontal forward direction L 0 Angle θ∈(-θ linit ,θ lim it ), S 1 is the coefficient weighting, which is obtained based on engineering experience debugging, V is the estimated speed of the robot, and θ is limited to the maximum allowable angle ±θ limit The robot head needs to rotate in the range of Δθ, and the robot head will be facing L after rotating Δθ. obj , and then follow the new nose direction L obj Run T depth After a certain time, determine whether the error satisfies |H error |>ΔH and calculate the new θ and Δθ, that is, T depth The depth control cycle makes the robot move along H depth ±ΔH operation; depth control duration T continuous Will be in |H error |≤ΔH is established or the image tracking mode is effective, and the value is cleared if the specified maximum depth control time T is continuously run. limit After that, the depth still cannot reach |H error |<ΔH, indicating that the depth control fails, at this time it switches to the fixed mode and alarms and waits for manual takeover;

[0016] Step S5: Continuously query the ultra-short baseline positioning information, only determine the horizontal coordinates x and y of the positioning data, without determining the vertical coordinate z, determine whether a layer of cleaning has been completed, and return to the starting position P s ;

[0017] Step S6: After determining that one layer of cleaning has been completed, perform ultra-short baseline positioning data processing, remove ultra-short baseline positioning abnormal values, and calculate the variance R of the cleaning depth, setting R Threshold is the depth error threshold. After each layer is cleaned, the variance R is calculated. When R>R ThresholdWhen the cleaning is considered to be missed, a warning of missed cleaning risk is issued for this layer, and the depth of this layer is assigned to H depth , the program automatically guides the robot to re-clean the current layer according to the depth control; at the same time, the upper computer uses a three-dimensional method to draw a cleaning path diagram of the net washing robot in the net box for the user to view. The user can also decide to manually select a certain water layer based on the three-dimensional view and control the net washing robot to repeatedly clean the water layer;

[0018] Step S7: When one layer of cleaning has been completed and there is no risk warning of missed cleaning, when cleaning a deeper layer, first use the depth control mode to move diagonally downward to increase the depth H of the position of the net washing robot. The robot switches the forward direction and uses the direction opposite to the previous layer as the heading, and records the starting position P of the current layer of cleaning. s , image tracking mode is adopted starting from the second layer;

[0019] Step S8, the image tracking mode specifically includes the following steps: identifying the dividing line between the uncleaned area and the cleaned area, drawing a reference line at a k% position in the image, if the attached area is on the left, finding the reference line from the left edge of the image to the right by k%, if the attached area is on the right, finding the reference line from the right edge of the image to the left by k%; calculating the lateral error ΔE between the extracted dividing line and the reference line, and calculating the angle between the dividing line and the horizontal edge of the image as α, calculating the heading error Δα=α-90, calculating the angle at which the robot needs to be adjusted in steering, and limiting the angle within the allowable range, Where S 2 is the coefficient, V is the control vector composed of the estimated speed, steering angle and speed Calculate and update in real time based on the video;

[0020] Step S9: If the segmentation line cannot be identified, the image tracking mode will be temporarily switched to the depth control mode, and the robot depth H will be maintained. depth ±ΔH operation, that is, H depth The depth of the net-washing robot when the image tracking mode fails and it runs in depth control mode;

[0021] Step S10, repeating steps S5-S9 to automatically clean the entire net, and replaying the cleaning work of the entire net box according to the three-dimensional cleaning trajectory.

[0022] As a preferred solution, when the first layer is cleaned in step S3, the image tracking mode is not effective.

[0023] As a preferred solution, the value of k in step S8 can be set and adjusted, and the value of k is usually greater than 70.

[0024] As a preferred solution, there is a hazard handling mechanism during the automatic cleaning process:

[0025] a. If the net disappears in the video, the boundary between the cleaned and uncleaned areas cannot be extracted;

[0026] b. The robot's posture suddenly changes and is greater than the threshold Right now

[0027] If both conditions a and b are satisfied at the same time, it is considered that the net-washing robot (1) is off the net. At this time, the net-washing robot (1) is kept stable at the current water depth, and an alarm is issued to wait for manual takeover.

[0028] When further cleaning the first layer, if condition a cannot be met, condition b can still be judged to implement hazard handling.

[0029] The present invention adopts the above technical solution, and has the following beneficial effects compared with the prior art: the present invention provides an automatic cleaning method to solve the demand for automatic operation of net cleaning. The present invention combines ultra-short baseline positioning technology, machine vision technology, and airborne water pressure sensors to realize the automatic net washing method of the net washing robot, reduce the requirements for operators, save manpower, and improve the coverage of cleaning.

[0030] Additional aspects and advantages of the present invention will become apparent from the following description or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0032] Figure 1 This is the structure diagram of the net washing robot;

[0033] Figure 2 This is a schematic diagram of the net-washing robot in the net cage;

[0034] Figure 3 It is a schematic diagram of the cleaning path strategy of the net-washing robot;

[0035] Figure 4 is a diagram of the net-washing robot in the unwashed and cleaned areas;

[0036] Figure 5 This is a diagram of the depth maintenance method of the net washing robot;

[0037] Figure 6 This is the cleaning robot's circular automatic cleaning flow chart.

[0038] in, Figures 1 to 5 The corresponding relationship between the reference numerals and the components is as follows:

[0039] 1. Net-cleaning robot; 2. Network camera; 3. Lighting; 4. Ultra-short baseline positioning base station; 5. Ultra-short baseline positioning beacon; 6. Water pressure sensor; 7. Net cage; 8. Net; 9. Predetermined cleaning trajectory; 10. Dividing line; 11. Currently required cleaning depth; 12. Reference line. DETAILED DESCRIPTION

[0040] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0041] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0042] Combine the following Figures 1 to 6 The automatic cleaning method based on the net cleaning device according to the embodiment of the present invention is specifically described.

[0043] The present invention proposes an automatic cleaning method based on a net cleaning device, such as Figure 1 As shown, the net cleaning device of the present method includes a net-washing robot 1, a network camera 2, a lighting lamp 3, a water pressure sensor 6 and an ultra-short baseline positioning system. The network camera 2 is respectively located in front and behind the net-washing robot 1, and is used to collect underwater images for image processing; the lighting lamp 3 is respectively located in front and behind the net-washing robot 1, and the lighting lamp 3 is used to fill light for the network camera 2; the water pressure sensor 6 is located on the sealed cabin of the net-washing robot 1, and is used to measure the water pressure, and then calculate the depth of the robot; the ultra-short baseline positioning system includes a base station 4 and a beacon 5, and the beacon 5 is located on the net-washing robot 1. The ultra-short baseline system is used to realize acoustic positioning, and the accurate position of the robot can be obtained. A net 8 is installed on the frame of the net box 7;

[0044] The robot's approximate position in the net cage is determined by acoustic positioning. This positioning has a certain error, usually within 1 meter, and there may be a positioning delay of several seconds to tens of seconds when blocked by fish schools. Under unobstructed conditions, a positioning information can usually be obtained in 1-2 seconds. The camera collects the "road information" in front of the robot, and uses image recognition technology to identify the boundary between the cleaned area and the uncleaned area of ​​the net as a reference for the forward trajectory. The acoustic positioning information, visual information, and the depth meter information carried by the robot are integrated to achieve autonomous operation and net cleaning of the robot in most scenarios.

[0045] The specific steps include:

[0046] Step S1: Figure 2 As shown, an ultra-short baseline base station 4 is fixedly installed in the net box 7, and the base station 4 is fixed on the frame of the net box 7;

[0047] Step S2: Figure 3 As shown, the program predetermines the cleaning path 9, completes the circular cleaning of one layer, moves diagonally downward, and increases the depth H of the position where the net washing robot 1 is located. H is predetermined by the staff on site based on experience, and H≤H clean , H clean It is the actual cleaning width of the robot washing plate, ensuring that each round of cleaning slightly overlaps the cleaning track of the previous layer to prevent missing cleaning; during operation, it cycles clockwise one layer, counterclockwise one layer, in sequence, until the entire net is cleaned. This is done to ensure that the robot's umbilical cord will not be damaged due to excessive twisting, thereby achieving layered cleaning;

[0048] Step S3: After the net-washing robot 1 is close to the net 8 near the water surface, the automatic cleaning mode is started, and the current position given by the ultra-short baseline positioning is recorded as the initial position P. s , at this time, the target depth H depth Set to 1, that is, maintain a water depth of 1 meter, perform the first layer of cleaning horizontally, use the depth control mode, use the closed-loop negative feedback control algorithm to control the robot's depth, and maintain it at H depth Within ±ΔH;

[0049] Step S4: Depth control mode specifically includes the following steps: current depth and target depth H depth Error of |H error |>ΔH, perform depth adjustment, such as Figure 5 As shown, L now is the current head orientation of the net washing robot 1, according to H error Calculate the direction L to which the head of the net washing robot 1 needs to be adjusted obj and the horizontal forward direction L 0 Angle θ∈(-θ linit ,θ lim it ), S 1 is the coefficient weighting, which is obtained based on engineering experience debugging, V is the estimated speed of the robot, and θ is limited to the maximum allowable angle ±θ limit The robot head needs to rotate in the range of Δθ, and the robot head will be facing L after rotating Δθ. obj , and then follow the new nose direction L obj Run T depth After a certain time, determine again whether the error satisfies |H error |>ΔH and calculate the new θ and Δθ, that is, T depthThe depth control cycle makes the robot move along H depth ±ΔH operation; depth control duration T continuous Will be in |H error |≤ΔH is established or the image tracking mode is effective, and it is cleared to zero. If the maximum depth control time T is continuously run, limit After that, the depth still cannot reach |H error |<ΔH, indicating that the depth control fails. At this time, it switches to the fixed mode and alarms and waits for manual takeover. In particular, the image tracking mode cannot take effect during the first layer of cleaning;

[0050] Step S5: Continuously query the ultra-short baseline positioning information (only determine the horizontal coordinates x and y of the positioning data, without determining the vertical coordinate z), determine whether the cleaning of one layer has been completed, and return to the starting position P s , marking the completion of one layer of cleaning, thus achieving layered cleaning;

[0051] Step S6: After determining that one layer of cleaning has been completed, perform ultra-short baseline positioning data processing, remove ultra-short baseline positioning abnormal values, and calculate the variance R of the cleaning depth, setting R Threshold is the depth error threshold. After each layer is cleaned, the variance R is calculated. When the R of this layer is greater than R Threshold When the cleaning is considered to be missed, a warning of missed cleaning risk is issued for this layer, and the depth of this layer is assigned to H depth , the program automatically guides the robot to re-clean the current layer according to the depth control; at the same time, the upper computer uses a three-dimensional method to draw a cleaning path diagram of the net washing robot 1 in the net box 7 for the user's reference. The user can combine the three-dimensional view and manually select a water layer as needed to control the net washing robot 1 to repeatedly clean the water layer;

[0052] Step S7: When one layer of cleaning has been completed and there is no risk warning of missed cleaning, when cleaning a deeper layer, the robot 1 first moves diagonally downward using the depth control mode to increase the depth H of the position where the net-washing robot 1 is located. The robot switches its forward direction and uses the direction opposite to the previous layer as the heading direction to record the starting position P of the cleaning of this layer. s , starting from the second layer, the image tracking mode can be adopted;

[0053] Step S8, the image tracking mode specifically includes the following steps: identifying the dividing line 10 between the uncleaned area and the cleaned area, such as Figure 4 As shown, the left side of the dividing line 10 is the mussel contaminated area, and the right side is the clean net area. Figure 412 is the reference line. The reference line 12 is drawn at a k% position in the image. k is preset by the staff based on experience. The value of k is usually greater than 70, which can ensure that the area to be repeatedly cleaned does not exceed 30%. If the attached area is on the left, the reference line 12 is found by counting k% from the left edge of the image to the right. If the attached area is on the right, the reference line 12 is found by counting k% from the right edge of the image to the left. The lateral error ΔE between the extracted segmentation line 10 and the reference line 12 is calculated, and the angle between the segmentation line 10 and the horizontal edge of the image is calculated as α. The heading error Δα=α-90 is calculated, and the angle at which the robot needs to be adjusted is calculated, and the angle is limited within the allowable range. Where S 2 is the coefficient, which is obtained by debugging based on engineering experience, and V is the control vector composed of the estimated speed, steering angle and speed Calculate and update in real time based on the video;

[0054] Step S9: For problems such as turbid water or obstructions that may occur occasionally, the dividing line 10 may not be recognized. In this case, the image tracking mode is temporarily switched to the depth control mode, and the robot depth H is maintained. depth ±ΔH operation, that is, H depth The depth at which the net-washing robot 1 is located when the image tracking mode fails and the robot operates in the depth control mode;

[0055] There are dangerous handling mechanisms during the automatic cleaning process:

[0056] a. If the net disappears in the video, the boundary between the cleaned area and the uncleaned area cannot be extracted 10;

[0057] b. The robot's posture changes suddenly and too much, greater than the threshold Right now

[0058] If both conditions a and b are met at the same time, it is considered that the net-washing robot 1 is off the net. At this time, the net-washing robot (1) is kept stable at the current water depth, and an alarm is issued to wait for manual takeover. When cleaning the first layer, if condition a cannot be met, condition b can still be judged to achieve danger handling.

[0059] Step S10: Repeat steps S5-S9 to automatically clean the entire net. The staff will review the cleaning work of the entire net box according to the three-dimensional cleaning trajectory.

[0060] In the description of the present invention, the term "plurality" refers to two or more than two. Unless otherwise clearly defined, the orientation or positional relationship indicated by the terms "upper" and "lower" is based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention; the terms "connection", "installation", "fixation", etc. should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0061] In the description of this specification, the description of the terms "one embodiment", "some embodiments", "specific embodiments", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0062] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An automatic cleaning method based on a net cleaning device, characterized in that The net washing device comprises a net washing robot (1), a network camera (2), a lighting lamp (3), a water pressure sensor (6) and an ultra-short baseline positioning system, wherein the network camera (2) is respectively located at the front and rear of the net washing robot (1), the lighting lamp (3) is respectively located at the front and rear of the net washing robot (1), the water pressure sensor (6) is located on the sealed cabin of the net washing robot (1), and the ultra-short baseline positioning system comprises a base station (4) and a beacon (5), and the beacon (5) is located on the net washing robot (1); The specific steps include: Step S1, an ultra-short baseline base station (4) is fixedly installed in a net box (7), and the base station (4) is fixed on a frame (7) of the net box 7; Step S2: The program predetermines the cleaning path (9), completes the circular cleaning of one layer, moves diagonally downward, and increases the depth H of the position of the net washing robot (1), H≤H clean , H clean is the actual cleaning width of the robot, which washes the net one layer clockwise and one layer counterclockwise in sequence until the entire net is cleaned (8); Step S3: After the net-washing robot (1) is close to the net (8) of the net box (7) near the water surface, the automatic cleaning mode is started, and the current position given by the ultra-short baseline positioning is recorded as the initial position P. s , at this time, the target depth H depth Set to 1, that is, maintain a water depth of 1 meter, perform the first layer of cleaning horizontally, use the depth control mode, use the closed-loop negative feedback control algorithm to control the robot's depth, and maintain it at H depth Within ±ΔH; Step S4: Depth control mode specifically includes the following steps: current depth and target depth H depth Error of |H err or|>ΔH, depth adjustment is performed, L now is the current head orientation of the net-washing robot (1), according to H err or calculate the direction L to which the head of the net washing robot (1) needs to be adjusted obj The angle between the horizontal and forward direction L0 θ∈(-θ linit ,θ limit ), S1 is the coefficient weighting, which is obtained based on engineering experience, V is the estimated speed of the robot, and θ is limited to the maximum allowable angle ±θ limit The robot head needs to rotate in the range of Δθ, and the robot head will be facing L after rotating Δθ. obj , and then follow the new nose direction L obj Run T depth After a certain time, determine again whether the error satisfies |H error |>ΔH and calculate the new θ and Δθ, that is, T depth The depth control cycle makes the robot move along H depth ±ΔH operation; depth control duration T continuous Will be in |H error |≤ΔH is established or the image tracking mode is effective, and it is cleared to zero. If the maximum depth control time T is continuously run, limit After that, the depth still cannot reach |H error |<ΔH, indicating that the depth control fails, at this time it switches to the fixed mode and alarms and waits for manual takeover; Step S5: Continuously query the ultra-short baseline positioning information, only determine the horizontal coordinates x and y of the positioning data, without determining the vertical coordinate z, determine whether a layer of cleaning has been completed, and return to the starting position P s ; Step S6: After determining that one layer of cleaning has been completed, perform ultra-short baseline positioning data processing, remove ultra-short baseline positioning abnormal values, and calculate the variance R of the cleaning depth, setting R Threshold is the threshold depth error value. After each layer is cleaned, the variance R is calculated. When the R of this layer is greater than R Threshold When the cleaning is considered to be missed, a warning of missed cleaning risk is issued for this layer, and the depth of this layer is assigned to H depth , the program automatically guides the robot to re-clean the current layer according to the depth control; at the same time, a three-dimensional method is used in the upper computer to draw a cleaning path diagram of the net washing robot (1) in the net box (7), allowing manual selection of a certain water layer and controlling the net washing robot (1) to repeatedly clean the water layer; Step S7: When one layer of cleaning has been completed and there is no risk of missed cleaning, when cleaning a deeper layer, the robot (1) is first moved diagonally downward using the depth control mode to increase the depth H of the position where the robot (1) is located. The robot switches its forward direction and uses the direction opposite to the previous layer as the heading direction to record the starting position P of the cleaning layer. s , starting from the second layer, the image tracking mode can be adopted; Step S8, the image tracking mode specifically includes the following steps: identifying the dividing line (10) between the uncleaned area and the cleaned area, drawing a reference line (12) at a k% position in the image, if the attached area is on the left, finding the reference line (12) from the left edge of the image to the right by k%, if the attached area is on the right, finding the reference line (12) from the right edge of the image to the left by k%; calculating the lateral error ΔE between the extracted dividing line (10) and the reference line (12), and calculating the angle between the dividing line (10) and the horizontal edge of the image as α, calculating the heading error Δα=α-90, calculating the angle at which the robot needs to be adjusted in steering, and limiting the angle within an allowable range, Where S2 is the coefficient, V is the control vector composed of the estimated speed, steering angle and speed Calculate and update in real time based on the video; Step S9: If the dividing line (10) cannot be identified, the image tracking mode will be temporarily switched to the depth control mode, and the robot depth H will be maintained. depth ±ΔH operation, that is, H depth The depth at which the net-washing robot (1) is located when the image tracking mode fails and the robot operates in the depth control mode; Step S10, repeating steps S5-S9 to automatically clean the entire net, and replaying the cleaning work of the entire net box according to the three-dimensional cleaning trajectory.

2. The automatic cleaning method based on the net cleaning device according to claim 1 is characterized in that ,In the first layer cleaning in step S3, the depth control mode is used and the ,image tracking mode is ineffective.

3. The automatic cleaning method based on the net cleaning device according to claim 1 is characterized in that ,The value of k in step S8 can be set and adjusted, and the value of k is usually greater than 70.

4. The automatic cleaning method based on the net cleaning device according to claim 1 is characterized in that , the dangerous handling mechanism exists in the automatic cleaning process: a. If the net disappears in the video, the boundary between the cleaned area and the uncleaned area cannot be extracted 10; b. The robot's posture suddenly changes and is greater than the threshold Right now If both conditions a and b are satisfied at the same time, it is considered that the net-washing robot (1) is off the net. At this time, the net-washing robot (1) is kept stable at the current water depth, and an alarm is issued to wait for manual takeover.

5. The automatic cleaning method based on the net cleaning device according to claim 4 is characterized in that ,When cleaning the first layer, if condition a cannot be met, condition b can still be judged to implement hazard handling.

Citation Information

Patent Citations

  • Equipment and method for cleaning current-resistance submerged cage netting

    CN104798703A

  • Cleaning method and system for cleaning robot

    CN108113583A

  • Depth setting control method and device for underwater robot and corresponding underwater robot

    CN111857166A

  • Crawler-type netting cleaning robot for seawater cage culture and netting cleaning method

    CN113275295A

  • Multi-wall-surface operation path planning method and system for underwater wall-climbing cleaning robot

    CN116974279A