Robot control system and control method based on data model drive

Abstract

The present invention relates to the field of robot control technology, and in particular to a robot control system driven by a data model and a control method thereof. The method comprises the following steps: obtaining a visual photograph of seafood; matching the seafood visual image with seafood types to generate seafood matching identification data; marking the seafood status of the seafood segmentation image according to the seafood matching identification data to generate seafood status marking data; generating processing control instructions for a processing and loading robot of a breeding ship based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions. The present invention improves the robot control accuracy and adaptability through accurate seafood identification, dynamic energy consumption management, real-time cleaning adjustment, environmental perception and obstacle avoidance, and the construction of a multi-robot collaborative network.

Description

Technical Field

[0001] The present invention relates to the field of robot control technology, and in particular to a robot control system driven by a data model and a control method thereof. Background Art

[0002] Data model-driven robot control methods are derived from the development of traditional robot control technology. Early robot control systems mainly relied on pre-programmed instruction sets and simple feedback mechanisms, lacking flexibility and adaptability. With the advancement of computer technology and sensor technology, model-based control methods have gradually emerged. This method predicts and controls the behavior of the robot system by establishing a mathematical model, significantly improving the control accuracy and stability. With the development of artificial intelligence and machine learning technology, data model-driven control methods have begun to dominate. It combines data-driven learning algorithms with traditional physical models, and uses big data analysis, deep learning and other technologies to extract laws and patterns from historical data to optimize control strategies. In recent years, advanced methods such as reinforcement learning and adaptive control have further enhanced the control capabilities of data model-driven control, enabling robots to autonomously learn and adapt in complex and dynamic environments. However, in actual applications, the energy consumption management of robots when performing operations such as sorting, cutting, and cleaning is often neglected. At the same time, traditional control methods are difficult to adjust and optimize operations in real time, resulting in low robot control accuracy and adaptability. Summary of the invention

[0003] Based on this, it is necessary to provide a robot control system and a control method thereof driven by a data model to solve at least one of the above technical problems.

[0004] To achieve the above object, a robot control method based on data model driving is provided, the method comprising the following steps:

[0005] Step S1: obtaining a visual photograph of seafood; matching the seafood visual image with seafood types to generate seafood matching identification data; marking the seafood segmentation image with seafood status according to the seafood matching identification data to generate seafood status marking data; generating processing control instructions for a processing and loading robot for aquaculture vessels based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions, and cleaning instructions;

[0006] Step S2: When the aquaculture vessel processing and loading robot receives a sorting instruction, it calculates the sorting energy consumption of the seafood matching and identification data to obtain the sorting energy consumption data; the sorting action control feedback of the robot arm is performed on the aquaculture vessel processing and loading robot through the sorting energy consumption data, thereby generating the robot sorting instruction feedback data; based on the cutting instruction, the fixed cutting instruction feedback is performed on the aquaculture vessel processing and loading robot, thereby generating the robot cutting instruction feedback data;

[0007] Step S3: When the aquaculture vessel processing and loading robot receives a cleaning instruction, the aquaculture vessel processing and loading robot is dynamically adjusted based on the cleaning instruction to generate cleaning dynamic adjustment data; the aquaculture vessel processing and loading robot is controlled to perform a cleaning action of a mechanical arm according to the cleaning dynamic adjustment data, thereby generating robot cleaning mechanical arm control data; and the cleaning instruction is fed back to the robot cleaning mechanical arm control data, thereby generating robot cleaning instruction feedback data;

[0008] Step S4: Acquire robot environmental perception data; extract position information from the robot environmental perception data to obtain robot location data and loading target location data; perform regional obstacle avoidance path planning based on the robot location data and loading target location data to generate loading area obstacle avoidance path planning data; perform availability evaluation on the loading target location data to generate loading target location availability evaluation data; perform loading transportation control feedback on the loading area obstacle avoidance path planning data based on the loading target location availability evaluation data to generate robot loading transportation control feedback data;

[0009] Step S5: construct a multi-robot collaborative network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0010] The present invention matches the type of seafood through visual shooting images, and can accurately identify the types and states of various seafood. According to the matching identification data, detailed processing control instructions are generated, including sorting, cutting, and cleaning, to ensure that the robot can operate according to the specific requirements of each seafood. The automatically generated instructions reduce manual intervention and improve the efficiency and consistency of the overall processing process. The energy consumption of sorting is calculated and fed back, which can effectively optimize the energy consumption of the robot and reduce energy waste. The robot arm motion control is carried out through the feedback of the sorting energy consumption data to ensure the accuracy of the sorting action and reduce the possibility of misoperation. The fixed cutting instruction feedback improves the stability of cutting, reduces the error and unstable factors in the cutting process, and ensures the consistency of the cutting effect. Dynamic adjustment based on the cleaning instruction can respond to different cleaning requirements in real time, and improve the flexibility and adaptability of the cleaning process. The cleaning operation is optimized through the robot arm control and instruction feedback, which ensures the thorough cleaning of the seafood, reduces the residue and pollution, and the generated cleaning instruction feedback data helps to adjust the cleaning strategy in time, which improves the efficiency and effect of the cleaning process. Acquiring and analyzing environmental perception data can more accurately understand the position of the robot and the feeding target, and improve the accuracy of the operation. The regional obstacle avoidance path planning data ensures the safe driving of the robot in the working area, reducing the impact of collisions and obstacles. The availability of the target position is evaluated to ensure that the target position of the loading operation is suitable for the actual operation, thereby improving the success rate and efficiency of the operation. By building a multi-robot collaborative network, coordination and cooperation between robots are achieved, and processing efficiency and operation quality are improved. The real-time control monitoring and early warning system can detect and handle abnormal situations in robot control in a timely manner, reducing the occurrence rate of failures and downtime. The stability and reliability of the entire system are enhanced, the efficiency and consistency of robot operation are ensured, and the work efficiency of the overall system is improved. Therefore, the present invention improves the robot control accuracy and adaptability through accurate seafood identification, dynamic energy consumption management, real-time cleaning adjustment, environmental perception and obstacle avoidance, and the construction of a multi-robot collaborative network.

[0011] Preferably, step S1 comprises the following steps:

[0012] Step S11: using the camera on the processing and loading robot of the aquaculture vessel to obtain a visual picture of the seafood;

[0013] Step S12: performing image preprocessing on the seafood visual image to generate a standard seafood visual shooting image, wherein the image preprocessing includes image denoising, image filtering, image brightness enhancement and image resolution improvement;

[0014] Step S13: segmenting the edge contour of the standard seafood visual photograph to generate a seafood segmentation map; matching the seafood segmentation map with the type of seafood based on a preset seafood database to generate seafood matching recognition data;

[0015] Step S14: Mark the seafood status of the seafood segmentation map according to the seafood matching and identification data to generate seafood status marking data; generate processing control instructions for the aquaculture ship processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions.

[0016] After the present invention obtains the visual image of seafood through the camera, the system can pre-process the image and generate a standardized visual image of seafood. This standardized processing ensures the accuracy of subsequent edge contour segmentation and type matching, reduces the influence of ambient light and image quality on the recognition result, and improves the accuracy of seafood recognition. By performing type matching and state marking on the seafood segmentation image, the system can generate accurate processing control instructions for the processing and feeding robot. Seafood of different types and states will be sorted, cut or cleaned according to their characteristics, thereby realizing an automated and intelligent processing flow and improving production efficiency. The system performs type matching based on a preset seafood database, and can dynamically generate processing control instructions based on matching recognition data. This means that the system has strong flexibility and adaptability, can handle various types of seafood, and adapt to different processing requirements. In the image preprocessing process, denoising, filtering, brightness enhancement and resolution enhancement are performed. These optimization steps can improve the quality of the original image and ensure that the effects of subsequent image segmentation and recognition are more stable and accurate. By combining visual technology with image processing algorithms, intelligent technical support is provided for the seafood processing process of aquaculture vessels, reducing the need for manual intervention, lowering labor costs, and improving the overall efficiency and quality of production.

[0017] Preferably, step S2 comprises the following steps:

[0018] Step S21: When the processing and loading robot of the aquaculture vessel receives the sorting instruction, it performs a seafood characteristic analysis on the seafood matching identification data to generate seafood characteristic attribute data, wherein the seafood characteristic analysis includes type analysis, size analysis and weight analysis;

[0019] Step S22: Calculate the sorting energy consumption of the aquaculture vessel processing and loading robot according to the seafood characteristic attribute data to obtain the sorting energy consumption data; sort the seafood characteristic attribute data according to the sorting energy consumption data to generate the seafood sorting priority data;

[0020] Step S23: performing sorting path planning based on the seafood sorting priority ranking data to generate seafood sorting planning path data; controlling the robot arm sorting action of the aquaculture vessel processing and loading robot through the seafood sorting planning path data, thereby generating robot sorting robot arm control data;

[0021] Step S24: Control feedback is performed on the control data of the robot sorting mechanical arm, and robot sorting instruction feedback data is generated; based on the cutting instruction, the fixed cutting instruction feedback is performed on the aquaculture vessel processing and loading robot, thereby generating robot cutting instruction feedback data.

[0022] The present invention calculates energy consumption and prioritizes the attribute data of seafood characteristics, and the system can optimize the sorting order according to the characteristics of each seafood, such as type, size and weight. This priority sorting can not only reduce the energy consumption of the robot arm, but also speed up the sorting process, thereby improving the overall sorting efficiency. Before sorting, energy consumption calculation will be performed according to the attribute data of seafood characteristics. By giving priority to the sorting path with the lowest energy consumption, the system effectively reduces the energy consumption in the sorting process, thereby achieving the goal of energy saving and emission reduction, and helping to reduce the operating cost of aquaculture vessels. During the sorting process, based on the priority sorting data, the system generates the optimal sorting path for the processing and feeding robot. This path planning not only takes into account energy consumption, but also integrates the motion efficiency and sorting time of the robot arm to ensure that the sorting task can be completed quickly and accurately. Through the analysis and feedback control of the attribute data of seafood characteristics, the system can adjust the sorting action of the robot arm in real time and generate cutting instruction feedback data. This intelligent feedback mechanism enables the robot to dynamically adjust the operation according to the actual situation, improving the accuracy and stability of sorting and cutting. By combining sorting priority sorting with path planning, precise control of automated sorting operations is achieved. The robot can not only sort efficiently according to the preset path, but also continuously optimize its own operation through the feedback mechanism to ensure that every sorting and cutting action can achieve the best effect. Relying on data-driven to perform sorting and cutting operations, efficient control of sorting robot arms and cutting tools is achieved through data analysis and feedback. This method can significantly improve the accuracy and reliability of sorting and processing, thereby improving the efficiency and quality of the entire seafood processing process.

[0023] Preferably, the feedback of the fixed cutting instruction to the aquaculture vessel processing and feeding robot based on the cutting instruction includes:

[0024] When the processing and loading robot of the aquaculture vessel receives a cutting instruction, it updates the standard seafood visual image through the robot sorting instruction feedback data to generate a seafood sorting update image; it parses the cutting instruction to confirm the cutting position, cutting depth and cutting shape of the seafood;

[0025] Perform seafood location positioning on seafood sorting update photos to generate seafood location positioning data; perform seafood location fixation stability analysis based on seafood location positioning data to generate seafood retention stability analysis data;

[0026] Based on the cutting position, cutting depth and cutting shape of seafood, the cutting tool of the aquaculture vessel processing and loading robot is selected to generate cutting tool selection data; the cutting path is planned for the cutting tool selection data using the seafood retention stability analysis data to generate seafood cutting planning path data;

[0027] The seafood cutting planning path data is used to control the cutting action of the robot arm of the aquaculture vessel processing and loading robot, thereby generating robot cutting robot arm control data; the robot cutting robot arm control data is fed back with fixed cutting instructions, thereby generating robot cutting instruction feedback data.

[0028] The present invention can accurately confirm the cutting position, depth and shape of seafood by analyzing the cutting instructions in detail. Based on these parameters, the system plans the cutting path and adjusts the cutting tool in real time to ensure the accuracy of the cutting operation. This highly customized cutting process effectively avoids cutting errors and improves the quality of the finished product. During the cutting process, the system can monitor the position and status of seafood in real time by updating the captured images and positioning data. This dynamic view update mechanism ensures the accurate execution of the cutting operation, and can adjust the cutting path and tool selection in time according to the feedback data, avoiding cutting deviations caused by changes in the position of seafood. Before executing the cutting, the system will perform a positioning and fixing stability analysis of the seafood to ensure the stability of the seafood during the cutting process. This analysis can effectively reduce the displacement and shaking of seafood during the cutting process, ensure the smooth progress of the cutting operation, and avoid damaging the seafood or producing bad incisions. Based on the requirements of the cutting position, depth and shape, the system automatically selects the most suitable cutting tool and performs path planning in combination with the retention stability data of the seafood. This intelligent tool selection and path optimization can significantly improve the cutting efficiency and ensure that each cutting can achieve the best effect. Through precise path planning data, the system can intelligently control the cutting action of the robot arm to ensure that the robot arm can perform cutting operations according to the predetermined path and angle. This control mechanism not only improves the cutting accuracy, but also effectively extends the service life of the robot arm and cutting tools. After each cutting operation, the system will generate cutting instruction feedback data through the fixed cutting instruction feedback mechanism. This closed-loop feedback mechanism can continuously optimize the cutting operation process, making subsequent cutting more accurate and efficient, and helping to improve the automation and intelligence level of the entire seafood processing process.

[0029] Preferably, step S3 comprises the following steps:

[0030] Step S31: when the aquaculture vessel processing and loading robot receives a cleaning instruction, it performs cleaning requirement instruction analysis on the cleaning instruction to generate cleaning instruction analysis data, wherein the cleaning requirement instruction analysis includes cleaning time analysis, cleaning medium analysis and cleaning intensity analysis;

[0031] Step S32: updating the view of the seafood sorting update shot image based on the robot cutting instruction feedback data to generate the seafood cutting update shot image; performing surface stain recognition on the seafood cutting update shot image to generate surface stain recognition data;

[0032] Step S33: selecting a cleaning tool for the aquaculture vessel processing and loading robot based on the surface stain recognition data to generate cleaning tool selection data; dynamically adjusting the cleaning of the aquaculture vessel processing and loading robot based on the cleaning tool selection data to generate cleaning dynamic adjustment data;

[0033] Step S34: Controlling the cleaning action of the robot arm of the aquaculture vessel processing and loading robot according to the dynamic cleaning adjustment data, thereby generating robot cleaning robot arm control data; providing cleaning instruction feedback to the robot cleaning robot arm control data, thereby generating robot cleaning instruction feedback data.

[0034] The present invention analyzes the cleaning instructions, and the system can accurately perform the cleaning operation according to the specific cleaning time, medium and intensity requirements. This customized cleaning process ensures the cleanliness of seafood, reduces the situation of insufficient or excessive cleaning, and protects the quality of seafood. During the cleaning process, the system monitors the surface condition of seafood in real time based on cutting feedback data and updated shooting images. Through this dynamic view update, the system can quickly identify stains and adjust the cleaning strategy in time to ensure the pertinence and effectiveness of the cleaning operation. The system uses the updated seafood image to identify surface stains and can quickly and accurately locate the stain area. This intelligent recognition function enables the cleaning tool to focus on the seriously polluted area, improving the cleaning efficiency and effect. The system automatically selects the appropriate cleaning tool according to the type and distribution of surface stains and makes dynamic adjustments. This intelligent selection and adjustment mechanism ensures that each stain can be treated most appropriately and avoids unnecessary waste of energy and resources. By cleaning the dynamic adjustment data, the system accurately controls the cleaning action of the robot arm to ensure that the robot arm can clean with the best path and strength. This precise control not only improves the cleaning effect, but also reduces the risk of damage to seafood. After the cleaning operation, the system generates feedback data through the cleaning instruction feedback mechanism to continuously optimize the cleaning process. This closed-loop feedback mechanism can gradually improve the efficiency and accuracy of the cleaning process, making each cleaning operation more intelligent and efficient. Through the comprehensive application of cleaning instruction analysis, tool selection, dynamic adjustment and motion control, the cleaning process of the aquaculture vessel is automated and intelligent, which greatly reduces the need for manual intervention and improves production efficiency and product quality.

[0035] Preferably, using the cleaning tool selection data to dynamically adjust the cleaning of the aquaculture vessel processing and feeding robot includes:

[0036] The cleaning image of the processing and loading robot of the aquaculture vessel is taken in real time using the cleaning tool selection data to generate a seafood cleaning image set; the seafood cleaning image set is segmented to generate a non-stained area image set and a stained area image set;

[0037] Perform stain area change analysis on the stain area image set to generate stain area change data; perform surface stain residual rate calculation on the stain area image set based on the stain area change data to obtain surface stain residual rate data on the stain area; perform surface finish analysis on the non-stain area image set to generate non-stain area surface finish data;

[0038] The cleaning completion degree is analyzed by the surface stain residual rate data of the stained area and the surface finish data of the non-stained area to generate the cleaning completion degree data; the cleaning completion degree data is compared with the preset cleaning completion degree threshold value. When the cleaning completion degree data is greater than or equal to the preset cleaning completion degree threshold value, the cleaning water pressure of the aquaculture vessel processing and feeding robot is reduced to generate low completion degree cleaning adjustment data;

[0039] When the cleaning completion data is less than the preset cleaning completion threshold, the cleaning time of the aquaculture ship processing and loading robot is extended to generate high-completion cleaning adjustment data; the low-completion cleaning adjustment data and the high-completion cleaning adjustment data are integrated to generate dynamic cleaning adjustment data.

[0040] The present invention can continuously monitor the surface state of seafood during the cleaning process by taking real-time images of seafood cleaning. By segmenting the image and analyzing the area change of the stain area, the cleaning progress can be accurately judged to ensure the accuracy of the cleaning operation. By analyzing the area change of the stain area, the system can calculate the residual rate of the stain. This intelligent calculation function helps the system to understand the cleaning effect in real time, ensuring that every stain can be thoroughly cleaned, thereby improving the cleaning quality. Analyzing the surface finish of the non-stain area helps to judge the impact of cleaning on the overall cleanliness of the seafood surface. This detailed analysis can avoid the reduction of the surface finish of the non-stain area due to excessive cleaning, thereby protecting the appearance and quality of the seafood. The system intelligently determines whether to extend the cleaning time or reduce the water pressure by comparing the cleaning completion data with the preset threshold. In this way, not only can the stain area be thoroughly cleaned, but also excessive cleaning of the cleaned area can be avoided, saving water resources and energy consumption. The dynamic cleaning adjustment data realizes adaptive control of the cleaning operation by integrating the cleaning adjustment data of low completion and high completion. The system can adjust the cleaning parameters in time according to the actual cleaning effect, thereby improving the cleaning efficiency and accuracy. Through real-time monitoring and dynamic adjustment, the system can minimize the waste of water pressure and cleaning time while ensuring the cleaning quality. This efficient use of resources not only reduces the operating costs of the aquaculture vessel, but also reduces the impact on the environment. The process achieves full automation of the cleaning process through a series of intelligent algorithms, which not only improves the cleaning efficiency, but also greatly reduces the need for manual intervention, making the cleaning operation of the entire aquaculture vessel more intelligent and efficient.

[0041] Preferably, step S4 comprises the following steps:

[0042] Step S41: Acquire robot environment perception data;

[0043] Step S42: extracting position information from the robot's environmental perception data to obtain the robot's position data and the loading target position data;

[0044] Step S43: performing path area obstacle identification on the robot location data and the loading target location data to generate path area obstacle identification data; performing regional obstacle avoidance path planning on the robot location data and the loading target location data using the path area obstacle identification data to generate loading area obstacle avoidance path planning data;

[0045] Step S44: perform availability evaluation on the loading target position data to generate loading target position availability evaluation data; perform loading transport control feedback on the loading area obstacle avoidance path planning data through the loading target position availability evaluation data to generate robot loading transport control feedback data.

[0046] The present invention obtains robot environmental perception data and extracts position information, so that the system can accurately locate the position of the robot and the feeding target. This precise positioning lays a solid foundation for subsequent path planning and obstacle avoidance operations, ensuring that the robot can smoothly perform tasks in complex environments. The system can identify obstacles in the path area in real time, and perform regional obstacle avoidance path planning based on this information. This dynamic obstacle avoidance function enables the robot to flexibly respond to changes and obstacles in the environment, ensuring the smooth completion of the feeding task. The system evaluates the availability of the feeding target position to ensure that the feeding position is effective and feasible during the operation. This evaluation function helps to avoid task failure or delay caused by inappropriate target position, thereby improving the reliability of the system. By combining obstacle avoidance path planning and availability evaluation of the feeding target position, the system can generate an optimized feeding path and provide real-time feedback to the robot for feeding transportation control. This intelligent path optimization and feedback control mechanism improves the work efficiency and execution accuracy of the robot. Through real-time data analysis of the robot's location and the feeding target position, the system can reduce errors in path planning. This high-precision path planning not only improves the accuracy of robot navigation, but also reduces deviations that occur during execution. The entire process of step S4 realizes a high degree of automation in the loading operation process, from environmental perception to path planning to transportation control, all of which are completed efficiently without human intervention. This automation improves operational efficiency while reducing labor costs. Through full-process perception, identification, planning and feedback, the system can ensure the safety and reliability of the robot when performing loading tasks, and reduce task failures or safety accidents caused by environmental factors or path obstacles.

[0047] Preferably, step S44 includes the following steps:

[0048] Step S441: using sensors to collect data of the feeding area based on the feeding target position data, and obtaining monitoring data of the feeding area, wherein the monitoring data of the feeding area includes available space data of the feeding area, temperature status data of the feeding area, and equipment operation status data of the feeding area;

[0049] Step S442: performing a loading position availability status analysis based on the loading area available space data, the loading area temperature status data, and the loading area equipment operation status data, and generating loading target position availability evaluation data, wherein the loading target position availability evaluation data includes regional available status data and regional unavailable status data; when the loading position availability status is regional unavailable status data, the aquaculture vessel processing loading robot is controlled to wait in situ until the loading position availability status is regional available status data;

[0050] Step S443: when the availability status of the loading position is regional available status data, a spatial stereoscopic area analysis is performed on the available space data of the loading area to generate available space stereoscopic data, wherein the available space stereoscopic data includes two-dimensional plane area data and three-dimensional height stereoscopic data;

[0051] Step S444: Compare the two-dimensional plane area data and the three-dimensional height stereo data. When the two-dimensional plane area data is larger than the three-dimensional height stereo data, the aquaculture vessel processing loading robot is placed and loaded according to the loading area obstacle avoidance path planning data, and the robot loading and transportation control feedback data is generated; when the two-dimensional plane area data is smaller than the three-dimensional height stereo data, the aquaculture vessel processing loading robot is stacked and loaded according to the loading area obstacle avoidance path planning data, and the robot loading and transportation control feedback data is still generated.

[0052] The present invention collects the available space, temperature status and equipment operation status data of the loading target position in real time through sensors, and the system can accurately understand the real-time situation of the loading area. This precise monitoring function ensures a comprehensive understanding and evaluation of the environment before performing the loading operation. According to the monitoring data of the loading area, the system can dynamically analyze and generate the availability evaluation data of the loading target position, including the available state and unavailable state data of the area. When the area is in an unavailable state, the robot can wait intelligently to avoid misoperation caused by the environment being unsuitable for loading, thereby improving the safety and reliability of the operation. By analyzing and comparing the two-dimensional plane area and three-dimensional height stereo data of the available space in the loading area, the system can intelligently judge and select the most suitable loading method. This dynamic spatial analysis ensures the accuracy of the loading process and avoids placement failures caused by insufficient or unsuitable space. When the system detects that the two-dimensional plane area data is greater than the three-dimensional height stereo data, the robot will choose the method of directly placing the loading; and when the space restriction is large, the system will intelligently choose stacking loading. This intelligent decision-making and execution based on real-time data improves the flexibility and adaptability of the operation and ensures the success rate of the loading process. Through precise monitoring, dynamic availability assessment, intelligent spatial analysis and flexible loading execution, the system greatly improves the efficiency of loading operations and reduces delays caused by environmental or spatial issues. Through intelligent waiting for unavailable states and reasonable analysis of spatial data, the system can effectively avoid potential operational risks. This enhanced reliability and safety ensures the efficient execution of the robot in complex environments.

[0053] Preferably, step S5 comprises the following steps:

[0054] Step S51: integrating the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to generate robot comprehensive control feedback data;

[0055] Step S52: performing multi-robot coordinated collaborative task allocation on the robot integrated control feedback data through data model technology to generate a multi-robot collaborative control matrix;

[0056] Step S53: constructing a robot collaborative network based on the multi-robot collaborative control matrix to obtain a robot control network; performing real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0057] The present invention integrates the control feedback data of sorting, cutting, cleaning and feeding to generate comprehensive control feedback data of the robot. This integration can summarize all operation instructions and feedback, improve the overall control accuracy and coordination of the robot, and ensure the orderly progress of various tasks. The comprehensive control feedback data is assigned to multi-robot coordinated and collaborative tasks through data model technology to generate a multi-robot collaborative control matrix. This method can effectively assign tasks, improve the collaborative efficiency between different robots, and reduce duplication of labor and waste of resources. Based on the multi-robot collaborative control matrix, a robot collaborative network is constructed. This network structure can improve the overall performance of the robot system, ensure that different robots can work together efficiently, and optimize the entire production process. The robot control network is controlled, monitored and warned in real time to perform robot control operations. This real-time monitoring function can timely discover and respond to possible problems, and reduce operation interruptions or safety hazards caused by faults or abnormal situations. Through intelligent data integration and task allocation, the automation level of the robot system is improved, so that the collaborative ability of the robot in complex tasks is significantly enhanced, and the operation becomes more intelligent and efficient. Through multi-robot collaboration and optimized control networks, the system can complete various tasks more effectively, thereby improving production efficiency. At the same time, the real-time monitoring and early warning mechanism enhances the stability of the system and reduces production stagnation caused by operational errors or equipment failures. Through reasonable task allocation and collaborative network optimization, the system can more effectively utilize robot resources, reduce resource waste, and improve overall production efficiency.

[0058] In this specification, a robot control system based on data model driving is provided, which is used to execute the above-mentioned robot control method based on data model driving. The robot control system based on data model driving includes:

[0059] The instruction division module is used to obtain visual images of seafood; match the seafood visual images with seafood types to generate seafood matching identification data; mark the seafood status of the seafood segmentation images according to the seafood matching identification data to generate seafood status marking data; generate processing control instructions for the aquaculture ship processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions;

[0060] The sorting and cutting control module is used to calculate the sorting energy consumption of the seafood matching identification data when the aquaculture ship processing and loading robot receives the sorting instruction, and obtain the sorting energy consumption data; to control the sorting action of the robot arm of the aquaculture ship processing and loading robot through the sorting energy consumption data, so as to generate the robot sorting instruction feedback data; to feedback the fixed cutting instruction to the aquaculture ship processing and loading robot based on the cutting instruction, so as to generate the robot cutting instruction feedback data;

[0061] A cleaning control module is used for dynamically adjusting the cleaning of the aquaculture vessel processing and loading robot based on the cleaning instruction when the aquaculture vessel processing and loading robot receives a cleaning instruction, thereby generating cleaning dynamic adjustment data; controlling the cleaning action of the robot arm of the aquaculture vessel processing and loading robot according to the cleaning dynamic adjustment data, thereby generating robot cleaning robot arm control data; and providing cleaning instruction feedback to the robot cleaning robot arm control data, thereby generating robot cleaning instruction feedback data;

[0062] The feeding path planning module is used to obtain the robot's environmental perception data; extract the position information of the robot's environmental perception data to obtain the robot's position data and the feeding target position data; perform regional obstacle avoidance path planning based on the robot's position data and the feeding target position data to generate the feeding area obstacle avoidance path planning data; perform availability evaluation on the feeding target position data to generate the feeding target position availability evaluation data; perform feeding transportation control feedback on the feeding area obstacle avoidance path planning data based on the feeding target position availability evaluation data to generate the robot feeding transportation control feedback data;

[0063] The robot collaboration monitoring module is used to construct a multi-robot collaboration network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; and to perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0064] The beneficial effect of the present invention is that by acquiring a visual image of seafood and performing type matching and state marking, seafood of different types and states can be accurately identified, and corresponding processing control instructions (including sorting, cutting, and cleaning instructions) can be generated, thereby ensuring that the robot can accurately process the specific seafood according to its characteristics, thereby improving the accuracy and efficiency of processing. By calculating and feedback controlling the sorting energy consumption, the robot's sorting action is optimized, energy consumption is reduced, and sorting efficiency is improved. In addition, the feedback of the fixed cutting instruction ensures the stability and accuracy of the cutting process, thereby improving the cutting quality. By dynamically adjusting the cleaning process and controlling the robotic arm, flexible optimization of the cleaning operation is achieved, the cleaning effect and efficiency are improved, and the accuracy of the cleaning process is ensured by the feedback mechanism, avoiding the problem of excessive or insufficient cleaning. By acquiring and analyzing environmental perception data, position extraction, obstacle avoidance path planning, and target position availability evaluation can be effectively performed, ensuring that the robot can safely and effectively perform loading operations in complex environments, reducing collision risks and transportation errors, and improving operational safety and efficiency. By building a multi-robot collaborative network and a real-time monitoring and early warning system, the coordination and collaboration capabilities between multiple robots are enhanced, ensuring the overall efficiency and stability of the system, and being able to discover and solve potential problems in a timely manner, thereby improving the reliability and work efficiency of the entire system. Therefore, the present invention improves the robot control accuracy and adaptability through accurate seafood identification, dynamic energy consumption management, real-time cleaning adjustment, environmental perception and obstacle avoidance, and the construction of a multi-robot collaborative network. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A schematic diagram of a process flow of a robot control method driven by a data model;

[0066] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0067] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0068] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0069] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0070] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0071] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0072] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0073] To achieve this, please refer to Figures 1 to 4 , a robot control method based on data model driving, the method comprising the following steps:

[0074] Step S1: obtaining a visual photograph of seafood; matching the seafood visual image with seafood types to generate seafood matching identification data; marking the seafood segmentation image with seafood status according to the seafood matching identification data to generate seafood status marking data; generating processing control instructions for a processing and loading robot for aquaculture vessels based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions, and cleaning instructions;

[0075] Step S2: When the aquaculture vessel processing and loading robot receives a sorting instruction, it calculates the sorting energy consumption of the seafood matching and identification data to obtain the sorting energy consumption data; the sorting action control feedback of the robot arm is performed on the aquaculture vessel processing and loading robot through the sorting energy consumption data, thereby generating the robot sorting instruction feedback data; based on the cutting instruction, the fixed cutting instruction feedback is performed on the aquaculture vessel processing and loading robot, thereby generating the robot cutting instruction feedback data;

[0076] Step S3: When the aquaculture vessel processing and loading robot receives a cleaning instruction, the aquaculture vessel processing and loading robot is dynamically adjusted based on the cleaning instruction to generate cleaning dynamic adjustment data; the aquaculture vessel processing and loading robot is controlled to perform a cleaning action of a mechanical arm according to the cleaning dynamic adjustment data, thereby generating robot cleaning mechanical arm control data; and the cleaning instruction is fed back to the robot cleaning mechanical arm control data, thereby generating robot cleaning instruction feedback data;

[0077] Step S4: Acquire robot environmental perception data; extract position information from the robot environmental perception data to obtain robot location data and loading target location data; perform regional obstacle avoidance path planning based on the robot location data and loading target location data to generate loading area obstacle avoidance path planning data; perform availability evaluation on the loading target location data to generate loading target location availability evaluation data; perform loading transportation control feedback on the loading area obstacle avoidance path planning data based on the loading target location availability evaluation data to generate robot loading transportation control feedback data;

[0078] Step S5: construct a multi-robot collaborative network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0079] The present invention matches the type of seafood through visual shooting images, and can accurately identify the types and states of various seafood. According to the matching identification data, detailed processing control instructions are generated, including sorting, cutting, and cleaning, to ensure that the robot can operate according to the specific requirements of each seafood. The automatically generated instructions reduce manual intervention and improve the efficiency and consistency of the overall processing process. The energy consumption of sorting is calculated and fed back, which can effectively optimize the energy consumption of the robot and reduce energy waste. The robot arm motion control is carried out through the feedback of the sorting energy consumption data to ensure the accuracy of the sorting action and reduce the possibility of misoperation. The fixed cutting instruction feedback improves the stability of cutting, reduces the error and unstable factors in the cutting process, and ensures the consistency of the cutting effect. Dynamic adjustment based on the cleaning instruction can respond to different cleaning requirements in real time, and improve the flexibility and adaptability of the cleaning process. The cleaning operation is optimized through the robot arm control and instruction feedback, which ensures the thorough cleaning of the seafood, reduces the residue and pollution, and the generated cleaning instruction feedback data helps to adjust the cleaning strategy in time, which improves the efficiency and effect of the cleaning process. Acquiring and analyzing environmental perception data can more accurately understand the position of the robot and the feeding target, and improve the accuracy of the operation. The regional obstacle avoidance path planning data ensures the safe driving of the robot in the working area, reducing the impact of collisions and obstacles. The availability of the target position is evaluated to ensure that the target position of the loading operation is suitable for the actual operation, thereby improving the success rate and efficiency of the operation. By building a multi-robot collaborative network, coordination and cooperation between robots are achieved, and processing efficiency and operation quality are improved. The real-time control monitoring and early warning system can detect and handle abnormal situations in robot control in a timely manner, reducing the occurrence rate of failures and downtime. The stability and reliability of the entire system are enhanced, the efficiency and consistency of robot operation are ensured, and the work efficiency of the overall system is improved. Therefore, the present invention improves the robot control accuracy and adaptability through accurate seafood identification, dynamic energy consumption management, real-time cleaning adjustment, environmental perception and obstacle avoidance, and the construction of a multi-robot collaborative network.

[0080] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of a robot control method based on data model driving of the present invention. In this example, the robot control method based on data model driving includes the following steps:

[0081] Step S1: obtaining a visual photograph of seafood; matching the seafood visual image with seafood types to generate seafood matching identification data; marking the seafood segmentation image with seafood status according to the seafood matching identification data to generate seafood status marking data; generating processing control instructions for a processing and loading robot for aquaculture vessels based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions, and cleaning instructions;

[0082] In an embodiment of the present invention, a visual image of seafood is captured in real time by using a camera installed on a breeding vessel. Ensure that the camera can capture images with sufficient resolution for subsequent processing. Obtain continuous visual images of seafood from the camera. The image should contain an overall view of the seafood and should have good lighting conditions for subsequent processing. Apply a denoising algorithm (such as median filtering, Gaussian filtering) to process the original image to reduce the noise in the image. Image filtering technology is used to enhance the edges and details of the image and improve the image quality. Adjust the brightness and contrast of the image to ensure that the seafood features in the image are more obvious. Use a super-resolution algorithm to improve the resolution of the image and the clarity of the image details. The image after preprocessing will be used as a standard seafood visual image for subsequent analysis. Use an image segmentation algorithm (such as Canny edge detection, U-Net, etc.) to perform edge contour segmentation on the standard seafood visual image to generate a seafood segmentation map. Based on a preset seafood database (including feature information of various seafood), match the seafood in the seafood segmentation map with the type in the database. Use an image classification algorithm (such as a convolutional neural network) for recognition to generate seafood matching recognition data. Based on the preset seafood database (including the characteristic information of various seafood), the seafood in the seafood segmentation map is matched with the type in the database. Use image classification algorithms (such as convolutional neural networks) for identification and generate seafood matching identification data. Generate appropriate processing control instructions based on seafood status marking data. Generate corresponding robot processing control instructions according to the type and status of seafood, including: Sorting instructions: determine how to sort different types of seafood to designated areas. Cutting instructions: determine the method, location, depth, etc. of cutting seafood. Cleaning instructions: determine the time, cleaning medium, cleaning intensity, etc. for cleaning seafood. Transmit these instructions to the processing and loading robots on the aquaculture ship to guide them to perform specific operations.

[0083] Step S2: When the aquaculture vessel processing and loading robot receives a sorting instruction, it calculates the sorting energy consumption of the seafood matching and identification data to obtain the sorting energy consumption data; the sorting action control feedback of the robot arm is performed on the aquaculture vessel processing and loading robot through the sorting energy consumption data, thereby generating the robot sorting instruction feedback data; based on the cutting instruction, the fixed cutting instruction feedback is performed on the aquaculture vessel processing and loading robot, thereby generating the robot cutting instruction feedback data;

[0084] In an embodiment of the present invention, the sorting instructions generated by step S1 are received by the processing and loading robot of the aquaculture vessel. These instructions include the types of seafood to be sorted and their corresponding sorting requirements. The type, size and weight of each seafood are analyzed using the information in the seafood matching identification data. These data will be used to calculate the energy consumption during the sorting process. Energy consumption models (such as energy consumption calculation formulas based on power and time) are applied to estimate the energy consumption of seafood of different types and weights. Taking into account factors such as the movement of the robotic arm, the complexity of the sorting action, and the force required to grab and move seafood, the sorting energy consumption data is calculated. For example: energy consumption = power × time; wherein power is the power consumed by the robotic arm during the sorting process, and time is the time required to complete the sorting. The energy efficiency of the robotic arm is evaluated based on the sorting energy consumption data. The speed, strength, etc. of the sorting action are adjusted to optimize energy consumption. These adjustment instructions are converted into robot sorting instruction feedback data. The robot sorting instruction feedback data is transmitted to the robot to guide it to make necessary action adjustments during the actual sorting process to ensure maximum energy efficiency and accuracy of operation. The processing and loading robot of the aquaculture vessel receives the cutting instruction generated by step S1. The instruction includes parameters such as cutting position, cutting depth and cutting shape. Based on the cutting instruction, the robot control system is used to perform position and retention analysis on the current seafood. Ensure the stability of the seafood during the cutting process. Convert the retention cutting instruction feedback into robot cutting instruction feedback data to guide the robot to make necessary adjustments and corrections during the cutting process. Transmit the robot cutting instruction feedback data to the robot to guide it to perform the cutting action and ensure that the cutting operation meets the instruction requirements.

[0085] Step S3: When the aquaculture vessel processing and loading robot receives a cleaning instruction, the aquaculture vessel processing and loading robot is dynamically adjusted based on the cleaning instruction to generate cleaning dynamic adjustment data; the aquaculture vessel processing and loading robot is controlled to perform a cleaning action of a mechanical arm according to the cleaning dynamic adjustment data, thereby generating robot cleaning mechanical arm control data; and the cleaning instruction is fed back to the robot cleaning mechanical arm control data, thereby generating robot cleaning instruction feedback data;

[0086] In the embodiment of the present invention, the cleaning instruction generated by step S1 is received by the aquaculture vessel processing and loading robot. The instruction includes parameters such as cleaning time, cleaning medium, and cleaning intensity. The cleaning instruction is parsed to obtain data such as cleaning time, cleaning medium, and cleaning intensity. Cleaning instruction parsing data is generated. The marine product is imaged in real time by using a sensor to generate a marine product cleaning image set. The marine product cleaning image set is segmented to generate a non-stain area image set and a stain area image set. The stain area change analysis is performed on the stain area image set to generate stain area change data. According to the stain area change data, the stain residual rate data on the surface of the stain area is calculated. The surface finish analysis is performed on the non-stain area image set to generate non-stain area surface finish data. The cleaning completion analysis is performed using the stain residual rate data on the surface of the stain area and the non-stain area surface finish data to generate cleaning completion data. If the cleaning completion data is greater than or equal to the preset cleaning completion threshold, low completion cleaning adjustment data is generated, and the robot is cleaned with reduced water pressure. If the cleaning completion data is less than the preset cleaning completion threshold, generate high completion cleaning adjustment data and extend the cleaning time of the robot. Integrate the low completion and high completion adjustment data to generate dynamic cleaning adjustment data. Compare the cleaning completion data with the preset cleaning completion threshold. Based on the dynamic cleaning adjustment data, control the cleaning action of the robot arm of the aquaculture vessel processing and feeding robot. Adjust the motion trajectory, strength, and use of cleaning tools of the robot arm to achieve the best cleaning effect. Transmit the robot cleaning robot arm control data to the robot to guide it to perform the cleaning action. Feedback the cleaning command of the robot cleaning robot arm control data. Record the actual execution of the cleaning process, including cleaning effect, cleaning time, etc. Feedback the robot cleaning command feedback data to the system for subsequent adjustment and optimization. Ensure that the cleaning process meets the expected requirements.

[0087] Step S4: Acquire robot environmental perception data; extract position information from the robot environmental perception data to obtain robot location data and loading target location data; perform regional obstacle avoidance path planning based on the robot location data and loading target location data to generate loading area obstacle avoidance path planning data; perform availability evaluation on the loading target location data to generate loading target location availability evaluation data; perform loading transportation control feedback on the loading area obstacle avoidance path planning data based on the loading target location availability evaluation data to generate robot loading transportation control feedback data;

[0088] In an embodiment of the present invention, sensor systems (such as laser radar, camera, ultrasonic sensor, etc.) on the robot are used to obtain perception data of the robot's surrounding environment. These data include obstacle information, spatial layout, target position, etc. in the environment. Position information is extracted from the robot's environmental perception data to obtain robot position data and feeding target position data. The robot position data includes the robot's current position and posture information. The feeding target position data includes the position and related spatial information of the target feeding area. Based on the robot position data and the feeding target position data, obstacles in the path area are identified. These obstacles include fixed obstacles (such as walls, equipment) and dynamic obstacles (such as moving objects). Regional obstacle avoidance path planning is performed using a path planning algorithm (such as Dijkstra algorithm, RRT algorithm, etc.). Feeding area obstacle avoidance path planning data is generated to describe the optimal path for the robot to reach the target position from the current position, avoiding obstacles in the path. Further data collection is performed on the feeding target position through sensors to generate feeding area monitoring data. These data include the available space, temperature status, equipment operation status, etc. in the feeding area. According to the monitoring data of the feeding area, the feeding position availability status analysis is performed to generate the feeding target position availability evaluation data. This includes whether the area is available and whether it meets the feeding requirements. The area availability status data indicates whether the feeding area is suitable for operation. The area unavailable status data indicates that the feeding area is not currently suitable for operation. If the feeding position is unavailable, the robot is controlled to wait in place until the target position becomes available. If the feeding position is available, the next step of control feedback is continued. Based on the feeding target position availability evaluation data and the feeding area obstacle avoidance path planning data, the robot feeding transportation control feedback data is generated. This data is used to guide the robot to perform corresponding actions during the feeding process. The robot feeding transportation control feedback data is transmitted to the robot control system. Adjust the robot's motion trajectory, speed, posture, etc. to ensure that the robot can smoothly transport materials from the current position to the target position.

[0089] Step S5: construct a multi-robot collaborative network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0090] In an embodiment of the present invention, the following control feedback data are collected and integrated: robot sorting instruction feedback data: about the execution of sorting tasks. Robot cutting instruction feedback data: about the execution of cutting tasks. Robot cleaning instruction feedback data: about the execution of cleaning tasks. Robot loading and transportation control feedback data: about the execution of loading tasks. These feedback data are integrated into robot comprehensive control feedback data through data integration technology (such as data fusion algorithm, data merging technology, etc.). The robot comprehensive control feedback data is analyzed using data model technology (such as model-based control, machine learning model, etc.), the specific role and responsibility of each robot in the task are determined, and a multi-robot collaborative control matrix is ​​generated, which describes the task allocation and collaborative relationship of each robot in the multi-robot system. Determine the collaborative strategy between robots, including task allocation, time scheduling, priority setting, etc., to ensure that the robots can efficiently complete their respective tasks and coordinate and cooperate. Construct a robot control network based on the multi-robot collaborative control matrix. The network includes communication links, collaborative relationships and control processes between each robot. Use network topology and control protocols (such as CAN bus, Ethernet, Robot Operating System (ROS), etc.) to achieve communication and collaboration between robots. Implement a real-time monitoring system to track the status, task progress, and potential problems of each robot in the robot control network. Configure early warning mechanisms (such as anomaly detection algorithms, condition monitoring systems, etc.) to detect and respond to potential failures or abnormal situations. For example, the system can monitor whether the robot exceeds the predetermined operating range or whether there are abnormalities in task execution. Generate real-time robot control monitoring early warning data based on the monitoring data, and adjust the robot control strategy and execution operation based on the early warning data to ensure the stability of the system and the smooth completion of the task.

[0091] Preferably, step S1 comprises the following steps:

[0092] Step S11: using the camera on the processing and loading robot of the aquaculture vessel to obtain a visual picture of the seafood;

[0093] Step S12: performing image preprocessing on the seafood visual image to generate a standard seafood visual shooting image, wherein the image preprocessing includes image denoising, image filtering, image brightness enhancement and image resolution improvement;

[0094] Step S13: segmenting the edge contour of the standard seafood visual photograph to generate a seafood segmentation map; matching the seafood segmentation map with the type of seafood based on a preset seafood database to generate seafood matching recognition data;

[0095] Step S14: Mark the seafood status of the seafood segmentation map according to the seafood matching and identification data to generate seafood status marking data; generate processing control instructions for the aquaculture ship processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions.

[0096] In an embodiment of the present invention, a high-resolution industrial camera is installed on a processing and feeding robot of a breeding ship to obtain a visual image of seafood. The camera should have good imaging capabilities and adapt to the light and humidity conditions in the marine environment. When the seafood passes through the processing line, the camera captures the visual image of the seafood in real time. The captured image should include the overall appearance and main features of the seafood to ensure the accuracy of subsequent processing. A denoising algorithm (such as Gaussian filtering, mean filtering or median filtering) is used to remove noise and interference in the visual image of the seafood to improve the clarity of the image. Image filtering technology is applied to enhance the edge and detail information in the image. Common filtering methods include Sobel filtering, Canny edge detection, etc. Brightness adjustment is performed according to the brightness histogram of the image to enhance the contrast of the image and make the outline of the seafood clearer. The super-resolution reconstruction algorithm is used to improve the resolution of the image to improve the detail performance of the visual image of the seafood. Finally, a standard seafood visual shot is generated. The edge detection algorithm (such as Canny edge detection) is used to segment the edge contour of the standard seafood visual shot to extract the contour information of the seafood. A seafood segmentation map is generated to identify the edge of each seafood. Use deep learning-based image recognition algorithms (such as convolutional neural networks (CNNs)) to identify the contours in the seafood segmentation map. Perform type matching based on a preset seafood database (including images and features of common seafood) to generate seafood matching recognition data. The recognition process should include feature extraction, pattern matching, and type classification. Mark the status of seafood based on the seafood matching recognition data. The status mark may include whether it is fresh, whether it is damaged, whether it needs to be processed, etc. Generate seafood status mark data and attach the status information to the corresponding seafood. Based on the seafood status mark data, automatically generate processing control instructions for the processing and loading robot of the aquaculture ship. The instructions include the following: Sorting instructions: According to the type and state of the seafood, decide whether to sort it to a specific processing area or classification container. Cutting instructions: For seafood that needs to be processed (such as shelling and de-spine), generate corresponding cutting instructions to control the robot to perform precise cutting. Cleaning instructions: Generate cleaning instructions for seafood that needs to be cleaned, and control the robot to send it into the cleaning process. The processing and loading robot performs actual operations according to the generated control instructions to complete the automated processing and sorting of seafood.

[0097] As an example of the present invention, refer to Figure 2As shown, in this example, step S2 includes:

[0098] Step S21: When the processing and loading robot of the aquaculture vessel receives the sorting instruction, it performs a seafood characteristic analysis on the seafood matching identification data to generate seafood characteristic attribute data, wherein the seafood characteristic analysis includes type analysis, size analysis and weight analysis;

[0099] Step S22: Calculate the sorting energy consumption of the aquaculture vessel processing and loading robot according to the seafood characteristic attribute data to obtain the sorting energy consumption data; sort the seafood characteristic attribute data according to the sorting energy consumption data to generate the seafood sorting priority data;

[0100] Step S23: performing sorting path planning based on the seafood sorting priority ranking data to generate seafood sorting planning path data; controlling the robot arm sorting action of the aquaculture vessel processing and loading robot through the seafood sorting planning path data, thereby generating robot sorting robot arm control data;

[0101] Step S24: Control feedback is performed on the control data of the robot sorting mechanical arm, and robot sorting instruction feedback data is generated; based on the cutting instruction, the fixed cutting instruction feedback is performed on the aquaculture vessel processing and loading robot, thereby generating robot cutting instruction feedback data.

[0102] In an embodiment of the present invention, by using the seafood matching identification data generated in step S13, the specific type of each seafood is analyzed according to the information in the seafood database. The type information is extracted and recorded as part of the seafood characteristic attribute data. By analyzing the edge contour segmentation map of the seafood, the length, width, height and other size data of each seafood are calculated. Commonly used techniques include edge detection, morphological operations and measurement algorithms. According to the size data and type information of the seafood, combined with known density data or actual measurement using sensors, the weight of each seafood is calculated. Weight data is also an important part of the seafood characteristic attribute data. The type, size and weight information obtained by the above analysis are integrated to generate complete seafood characteristic attribute data. This data will serve as the basis for subsequent sorting energy consumption calculation and sorting priority sorting. According to the seafood characteristic attribute data (such as size and weight), combined with the energy consumption model of the processing and loading robot of the aquaculture ship, the sorting energy consumption of each seafood is calculated. Energy consumption calculation can involve factors such as the moving distance, required torque, speed and load of the robot arm. Energy consumption calculation formula example: E = f (d, w, v, F); where E represents the sorting energy consumption, d represents the moving distance of the robot arm, w represents the weight of the seafood, v represents the moving speed, and F represents the required torque. Combine the characteristic attribute data of seafood with its corresponding sorting energy consumption data to perform sorting priority sorting. The sorting criteria can generate seafood sorting priority sorting data based on factors such as the lowest energy consumption priority, seafood freshness, size, etc., to guide sorting path planning. Based on the seafood sorting priority sorting data, plan the optimal sorting path. Path planning should consider factors such as energy consumption, time efficiency, and seafood processing sequence to reduce the total sorting time and energy consumption. Use path optimization algorithms (such as Dijkstra algorithm, etc.) to generate seafood sorting planning path data. According to the sorting planning path data, control the robot arm of the aquaculture ship processing and loading robot to perform sorting operations. The actions of the robot arm include positioning, grasping, moving, and placing. Generate robot sorting robot arm control data to ensure that the robot arm accurately performs the sorting task according to the planned path. Monitor the movements of the robot sorting arm in real time and perform feedback control. The purpose of feedback control is to ensure the accuracy and stability of the sorting process and avoid misoperation and robot arm conflicts. Feedback mechanisms include position correction, torque adjustment, speed control, etc. Generate robot sorting instruction feedback data so that the sorting operation can be adjusted when necessary. After the sorting is completed, if the seafood needs to be cut, the sorted seafood will be passed to the cutting process. Based on the cutting instruction, generate the fixed cutting instruction feedback of the aquaculture ship processing and loading robot. The feedback data is used to control the robot arm to accurately position the seafood and perform the cutting operation, and generate the robot cutting instruction feedback data to ensure the accuracy and quality of the cutting process.

[0103] Preferably, the feedback of the fixed cutting instruction to the aquaculture vessel processing and feeding robot based on the cutting instruction includes:

[0104] When the processing and loading robot of the aquaculture vessel receives a cutting instruction, it updates the standard seafood visual image through the robot sorting instruction feedback data to generate a seafood sorting update image; it parses the cutting instruction to confirm the cutting position, cutting depth and cutting shape of the seafood;

[0105] Perform seafood location positioning on seafood sorting update photos to generate seafood location positioning data; perform seafood location fixation stability analysis based on seafood location positioning data to generate seafood retention stability analysis data;

[0106] Based on the cutting position, cutting depth and cutting shape of seafood, the cutting tool of the aquaculture vessel processing and loading robot is selected to generate cutting tool selection data; the cutting path is planned for the cutting tool selection data using the seafood retention stability analysis data to generate seafood cutting planning path data;

[0107] The seafood cutting planning path data is used to control the cutting action of the robot arm of the aquaculture vessel processing and loading robot, thereby generating robot cutting robot arm control data; the robot cutting robot arm control data is fed back with fixed cutting instructions, thereby generating robot cutting instruction feedback data.

[0108] In the embodiment of the present invention, when the processing and loading robot of the aquaculture vessel receives a cutting instruction, the standard seafood visual shooting image is first updated through the robot sorting instruction feedback data generated in the previous step to reflect the latest seafood position and status. The updated image will record any position change or posture change during the sorting and conveying process, thereby generating a seafood sorting update shooting image. The received cutting instruction is analyzed in detail to confirm the cutting position, cutting depth and cutting shape of the seafood. This process involves analyzing specific parameters in the instruction, such as the starting and ending points of the cutting line, the predetermined cutting depth, and the geometric shape required for cutting (such as straight line cutting, curve cutting or other complex shapes). The seafood sorting update shooting image is used to locate the position of the seafood. Through image processing technology (such as template matching or edge detection), the exact position of the seafood in the current view is determined, and the seafood position positioning data is generated to ensure the accuracy of subsequent cutting operations. According to the seafood position positioning data, the stability of the seafood in the current position is evaluated. The shape, weight distribution and external force of the seafood are considered during the analysis process to ensure that the seafood will not shift or shake during the cutting process. Generate seafood retention stability analysis data, which is used to verify whether the seafood is firmly fixed and provide guarantee for the cutting process. Select the appropriate cutting tool based on the previously analyzed cutting position, cutting depth and cutting shape. The selection of cutting tools is based on the cutting material, the required accuracy and the applicability of the tool. Generate cutting tool selection data, which will guide the robot to select and prepare appropriate cutting tools (such as laser cutters, knives, etc.). Combine seafood retention stability analysis data with cutting tool selection data to carry out detailed planning of cutting paths. Cutting path planning will include the starting point, end point, path trajectory and moving speed of the cutting tool. Generate seafood cutting planning path data to ensure that the cutting tool performs cutting operations according to the optimal path and correct parameters. According to the seafood cutting planning path data, the robotic arm of the aquaculture ship processing and loading robot is precisely controlled. The robotic arm will perform cutting actions according to the set path and parameters to ensure the stability and accuracy of the cutting process. Generate robot cutting robotic arm control data, which includes the movement instructions of the robotic arm, the start instructions of the cutting tool and the real-time adjustment instructions. During the cutting operation, the robot arm's movements are continuously monitored and adjusted to ensure the quality and consistency of the cutting. Deviations and errors generated during the cutting process are recorded during the monitoring process, and these deviations are adjusted in real time. Feedback data on the robot's cutting instructions is generated to optimize and improve subsequent cutting operations or similar tasks. The feedback data will be used as a reference for the next operation to ensure continuous improvement of the cutting process.

[0109] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0110] Step S31: when the aquaculture vessel processing and loading robot receives a cleaning instruction, it performs cleaning requirement instruction analysis on the cleaning instruction to generate cleaning instruction analysis data, wherein the cleaning requirement instruction analysis includes cleaning time analysis, cleaning medium analysis and cleaning intensity analysis;

[0111] Step S32: updating the view of the seafood sorting update shot image based on the robot cutting instruction feedback data to generate the seafood cutting update shot image; performing surface stain recognition on the seafood cutting update shot image to generate surface stain recognition data;

[0112] Step S33: selecting a cleaning tool for the aquaculture vessel processing and loading robot based on the surface stain recognition data to generate cleaning tool selection data; dynamically adjusting the cleaning of the aquaculture vessel processing and loading robot based on the cleaning tool selection data to generate cleaning dynamic adjustment data;

[0113] Step S34: Controlling the cleaning action of the robot arm of the aquaculture vessel processing and loading robot according to the dynamic cleaning adjustment data, thereby generating robot cleaning robot arm control data; providing cleaning instruction feedback to the robot cleaning robot arm control data, thereby generating robot cleaning instruction feedback data.

[0114] In the embodiment of the present invention, when the processing and loading robot of the aquaculture vessel receives the cleaning instruction, the instruction is first analyzed in detail to confirm the specific requirements of the cleaning. According to the type, surface state and processing requirements of the seafood, the required cleaning time is determined, and the cleaning time analysis data is generated. The required cleaning medium, such as fresh water, sea water or other specific cleaning liquids, is analyzed to generate the cleaning medium analysis data. According to the degree of stains, the tolerance of seafood and the requirements of the cleaning effect, the intensity of cleaning is determined, and the cleaning intensity analysis data is generated. Finally, the cleaning instruction analysis data is generated by combining the data of the above three aspects to guide the subsequent cleaning operation. Based on the robot cutting instruction feedback data, the view of the seafood sorting update shooting image is updated to obtain the latest state of the seafood after cutting. The updated image will show the new stains on the surface of the seafood or the residues generated during the cutting process, thereby generating the seafood cutting update shooting image. The updated image is subjected to surface stain recognition, and the stains on the surface of the seafood are identified and classified using image processing technology. Surface stain recognition data is generated, which includes information such as the location, area, type, etc. of the stains, which is used for subsequent cleaning tool selection and dynamic adjustment. Based on the surface stain recognition data, a suitable cleaning tool is selected. The selection of cleaning tools is based on the type of stains, seafood material and cleaning requirements. Cleaning tool selection data is generated to guide the robot to select suitable cleaning tools (such as spray devices, brushes, ultrasonic cleaners, etc.). Using the cleaning tool selection data, the cleaning process is dynamically adjusted according to the actual situation. The adjustment content includes the use angle of the cleaning tool, the moving speed, the cleaning pressure and the flow rate of the medium. The dynamic adjustment data of cleaning is generated to ensure that the cleaning process is efficient and meets the cleaning requirements. According to the dynamic adjustment data of cleaning, the robotic arm of the processing and feeding robot of the aquaculture ship is precisely controlled. The robotic arm performs the cleaning action according to the set path and parameters to ensure the coverage and effect of the cleaning. The robot cleaning robotic arm control data is generated, which includes the movement trajectory of the robotic arm, the operating parameters of the cleaning tool and the real-time adjustment instructions. During the cleaning operation, the cleaning effect and the execution of the robotic arm are monitored and recorded in real time. Through the feedback mechanism, the cleaning operation is adjusted and optimized to ensure that the final cleaning quality meets expectations. The robot cleaning instruction feedback data is generated to provide reference and optimization basis for the execution of subsequent similar tasks.

[0115] Preferably, using the cleaning tool selection data to dynamically adjust the cleaning of the aquaculture vessel processing and feeding robot includes:

[0116] The cleaning image of the processing and loading robot of the aquaculture vessel is taken in real time using the cleaning tool selection data to generate a seafood cleaning image set; the seafood cleaning image set is segmented to generate a non-stained area image set and a stained area image set;

[0117] Perform stain area change analysis on the stain area image set to generate stain area change data; perform surface stain residual rate calculation on the stain area image set based on the stain area change data to obtain surface stain residual rate data on the stain area; perform surface finish analysis on the non-stain area image set to generate non-stain area surface finish data;

[0118] The cleaning completion degree is analyzed by the surface stain residual rate data of the stained area and the surface finish data of the non-stained area to generate the cleaning completion degree data; the cleaning completion degree data is compared with the preset cleaning completion degree threshold value. When the cleaning completion degree data is greater than or equal to the preset cleaning completion degree threshold value, the cleaning water pressure of the aquaculture vessel processing and feeding robot is reduced to generate low completion degree cleaning adjustment data;

[0119] When the cleaning completion data is less than the preset cleaning completion threshold, the cleaning time of the aquaculture ship processing and loading robot is extended to generate high-completion cleaning adjustment data; the low-completion cleaning adjustment data and the high-completion cleaning adjustment data are integrated to generate dynamic cleaning adjustment data.

[0120] In an embodiment of the present invention, during the cleaning process, the processing and loading robot of the aquaculture vessel is dynamically adjusted based on the cleaning tool selection data, and the robot camera is used to capture the images during the cleaning process in real time to generate a seafood cleaning image set, which includes the surface images of seafood at different time points during the entire cleaning process. The generated seafood cleaning image set is processed and analyzed, and different areas in the image are separated using image segmentation technology. A non-stain area image set and a stain area image set are generated, corresponding to the area not covered by the stain during the cleaning process and the area where the stain still exists, respectively. The stain area in the stain area image set is dynamically monitored, the area change of the stain during the cleaning process is analyzed, and the area change data of the stain area is generated, which reflects the reduction of the stain during the cleaning process. According to the area change data of the stain area, the stain area image set is further analyzed, the residual rate of the stain after cleaning is calculated, and the surface stain residual rate data of the stain area is generated, which is used to evaluate the thoroughness of the cleaning effect. The surface finish of the non-stain area image set is analyzed to evaluate the smoothness and cleanliness of the surface of the seafood after cleaning, and the surface finish data of the non-stain area is generated, which is used to comprehensively evaluate the cleaning effect together with the residual situation of the stain area. The surface stain residual rate data of the stain area and the surface finish data of the non-stain area are used to analyze the cleaning completion and generate the cleaning completion data, which reflects the overall effect of the cleaning process. The cleaning completion data is compared with the preset cleaning completion threshold to determine whether the cleaning effect meets the standard. When the cleaning completion data is greater than or equal to the preset cleaning completion threshold, it means that the cleaning effect has met the standard, and the cleaning water pressure of the aquaculture processing and loading robot is reduced to reduce energy consumption, and low completion cleaning adjustment data is generated. When the cleaning completion data is less than the preset cleaning completion threshold, it indicates that the cleaning effect does not meet the requirements. The cleaning time of the aquaculture processing and loading robot is extended to enhance the cleaning effect, and high completion cleaning adjustment data is generated.

[0121] As an example of the present invention, refer to Figure 4 As shown, in this example, step S4 includes:

[0122] Step S41: Acquire robot environment perception data;

[0123] Step S42: extracting position information from the robot's environmental perception data to obtain the robot's position data and the loading target position data;

[0124] Step S43: performing path area obstacle identification on the robot location data and the loading target location data to generate path area obstacle identification data; performing regional obstacle avoidance path planning on the robot location data and the loading target location data using the path area obstacle identification data to generate loading area obstacle avoidance path planning data;

[0125] Step S44: perform availability evaluation on the loading target position data to generate loading target position availability evaluation data; perform loading transport control feedback on the loading area obstacle avoidance path planning data through the loading target position availability evaluation data to generate robot loading transport control feedback data.

[0126] In the embodiment of the present invention, by utilizing a variety of sensors (such as laser radar, ultrasonic sensor, camera, etc.) equipped with the feeding robot for aquaculture workboat, the perception data of the surrounding environment is obtained in real time. These data include the perception information of obstacles, terrain, lighting conditions and other dynamic environmental changes around the robot, and the robot environment perception data is generated to provide basic information for subsequent path planning and obstacle avoidance. The robot environment perception data is analyzed, the precise position of the robot in the current environment is located and extracted, and the robot location data is generated, which includes the current coordinates of the robot and its posture information in three-dimensional space. According to the task requirements, the feeding target position that the robot needs to reach is determined, and the precise coordinates of the target position are extracted from the robot environment perception data to generate the feeding target position data for subsequent path planning. Based on the robot location data and the feeding target position data, the path area between the two is subjected to obstacle recognition. By using technologies such as image processing, point cloud analysis or radar data, obstacles in the path, such as protruding ground, obstacles or dynamic interference objects, are identified, and path area obstacle recognition data is generated, which identifies the obstacle area to be avoided in the path. Obstacle avoidance path planning is performed by combining the obstacle identification data of the path area with the robot's location data and the loading target location data. The planned path needs to avoid all identified obstacles while ensuring the shortest distance or optimal energy consumption of the path, and generate loading area obstacle avoidance path planning data, which guides the robot to perform loading tasks on a safe and efficient path. Before the robot approaches or reaches the loading target location, the availability of the location is evaluated. The evaluation content includes whether there is enough space for loading operations at the location, whether there are new obstacles or environmental changes, and whether the target location is consistent with the preset plan. The loading target location availability evaluation data is generated to determine whether the target location is suitable for the current task execution. Based on the loading target location availability evaluation data, the previously generated loading area obstacle avoidance path planning data is dynamically adjusted. If the target location evaluation result shows that there are unsuitable factors, the path is replanned or the loading strategy is adjusted, and the robot loading transportation control feedback data is generated to guide the robot to make real-time adjustments to the path and tasks in actual operation to ensure the smooth completion of the loading process.

[0127] Preferably, step S44 includes the following steps:

[0128] Step S441: using sensors to collect data of the feeding area based on the feeding target position data, and obtaining monitoring data of the feeding area, wherein the monitoring data of the feeding area includes available space data of the feeding area, temperature status data of the feeding area, and equipment operation status data of the feeding area;

[0129] Step S442: performing a loading position availability status analysis based on the loading area available space data, the loading area temperature status data, and the loading area equipment operation status data, and generating loading target position availability evaluation data, wherein the loading target position availability evaluation data includes regional available status data and regional unavailable status data; when the loading position availability status is regional unavailable status data, the aquaculture vessel processing loading robot is controlled to wait in situ until the loading position availability status is regional available status data;

[0130] Step S443: when the availability status of the loading position is regional available status data, a spatial stereoscopic area analysis is performed on the available space data of the loading area to generate available space stereoscopic data, wherein the available space stereoscopic data includes two-dimensional plane area data and three-dimensional height stereoscopic data;

[0131] Step S444: Compare the two-dimensional plane area data and the three-dimensional height stereo data. When the two-dimensional plane area data is larger than the three-dimensional height stereo data, the aquaculture vessel processing loading robot is placed and loaded according to the loading area obstacle avoidance path planning data, and the robot loading and transportation control feedback data is generated; when the two-dimensional plane area data is smaller than the three-dimensional height stereo data, the aquaculture vessel processing loading robot is stacked and loaded according to the loading area obstacle avoidance path planning data, and the robot loading and transportation control feedback data is still generated.

[0132] In an embodiment of the present invention, by using the sensor equipped with the feeding robot for processing aquaculture vessels, detailed data collection is performed on the feeding target position. The collected data includes available space data of the feeding area (such as the free area of ​​the space, the available height), temperature status data of the feeding area (such as the current temperature, the temperature change trend), and equipment operation status data of the feeding area (such as the operating status of the equipment, whether there is a fault). These data are summarized to form the feeding area monitoring data, which provides a basis for the subsequent feeding position availability analysis. According to the available space data of the feeding area, the temperature status data of the feeding area, and the equipment operation status data of the feeding area in the feeding area monitoring data, a comprehensive analysis is performed to evaluate the availability of the feeding position. The feeding target position availability evaluation data generated by the analysis includes regional available status data and regional unavailable status data. If the evaluation result is regional unavailable status data, a waiting control instruction is issued to the feeding robot for processing aquaculture vessels to suspend operation and wait for the feeding position to become available. When the feeding position is evaluated as regional available status data, further spatial stereoscopic analysis is performed on the available space data of the feeding area. The analysis generates available spatial stereoscopic data, which includes two-dimensional plane area data and three-dimensional height stereoscopic data. The two-dimensional plane area data represents the size of the plane space of the feeding area, while the three-dimensional height stereo data represents the available height range of the area. Compare the two-dimensional plane area data and the three-dimensional height stereo data to determine the appropriate feeding method. When the two-dimensional plane area data is larger than the three-dimensional height stereo data, this indicates that the area is suitable for conventional placement and feeding operations. Based on the obstacle avoidance path planning data of the feeding area, the aquaculture ship processing feeding robot is placed and fed, and the robot feeding and transportation control feedback data is generated. When the two-dimensional plane area data is smaller than the three-dimensional height stereo data, this means that the height of the area is relatively sufficient, but the plane space is limited. At this time, the robot should perform stacking feeding operations to make full use of the vertical space, and the robot feeding and transportation control feedback data is also generated.

[0133] Preferably, step S5 comprises the following steps:

[0134] Step S51: integrating the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to generate robot comprehensive control feedback data;

[0135] Step S52: performing multi-robot coordinated collaborative task allocation on the robot integrated control feedback data through data model technology to generate a multi-robot collaborative control matrix;

[0136] Step S53: constructing a robot collaborative network based on the multi-robot collaborative control matrix to obtain a robot control network; performing real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0137] In an embodiment of the present invention, feedback data from each step is collected, including robot sorting instruction feedback data, robot cutting instruction feedback data, robot cleaning instruction feedback data, and robot loading and transportation control feedback data. The feedback data from different sources are integrated to generate robot comprehensive control feedback data. This data integrates various operation feedback information of the robot and provides basic data support for subsequent multi-robot collaboration. Based on the integrated robot comprehensive control feedback data, data model technology is used for analysis and processing to determine the task allocation strategy between multiple robots. The allocation process generates a multi-robot collaborative control matrix, which clarifies the specific tasks, collaborative relationships, and priority order of each robot. Through this allocation, it is ensured that the multi-robot system can coordinate in a complex working environment, avoid resource conflicts, and improve overall operating efficiency. Based on the generated multi-robot collaborative control matrix, a robot control network is constructed to connect multiple robots into a coordinated operation system. The network can dynamically adjust the interaction relationship between robots to ensure that each robot can coordinate in time during the work process and make corresponding adjustments according to environmental changes. The constructed robot control network is monitored in real time to ensure that the operation of each robot meets the predetermined task allocation and operation process. If an abnormality or potential operational risk is detected during the monitoring process, the system will issue an early warning and take necessary adjustment measures to ensure the stable operation of the entire system and the smooth completion of the task.

[0138] In this specification, a robot control system based on data model driving is provided, which is used to execute the above-mentioned robot control method based on data model driving. The robot control system based on data model driving includes:

[0139] The instruction division module is used to obtain visual images of seafood; match the seafood visual images with seafood types to generate seafood matching identification data; mark the seafood status of the seafood segmentation images according to the seafood matching identification data to generate seafood status marking data; generate processing control instructions for the aquaculture ship processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions;

[0140] The sorting and cutting control module is used to calculate the sorting energy consumption of the seafood matching identification data when the aquaculture ship processing and loading robot receives the sorting instruction, and obtain the sorting energy consumption data; to control the sorting action of the robot arm of the aquaculture ship processing and loading robot through the sorting energy consumption data, so as to generate the robot sorting instruction feedback data; to feedback the fixed cutting instruction to the aquaculture ship processing and loading robot based on the cutting instruction, so as to generate the robot cutting instruction feedback data;

[0141] A cleaning control module is used for dynamically adjusting the cleaning of the aquaculture vessel processing and loading robot based on the cleaning instruction when the aquaculture vessel processing and loading robot receives a cleaning instruction, thereby generating cleaning dynamic adjustment data; controlling the cleaning action of the robot arm of the aquaculture vessel processing and loading robot according to the cleaning dynamic adjustment data, thereby generating robot cleaning robot arm control data; and providing cleaning instruction feedback to the robot cleaning robot arm control data, thereby generating robot cleaning instruction feedback data;

[0142] The feeding path planning module is used to obtain the robot's environmental perception data; extract the position information of the robot's environmental perception data to obtain the robot's position data and the feeding target position data; perform regional obstacle avoidance path planning based on the robot's position data and the feeding target position data to generate the feeding area obstacle avoidance path planning data; perform availability evaluation on the feeding target position data to generate the feeding target position availability evaluation data; perform feeding transportation control feedback on the feeding area obstacle avoidance path planning data based on the feeding target position availability evaluation data to generate the robot feeding transportation control feedback data;

[0143] The robot collaboration monitoring module is used to construct a multi-robot collaboration network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; and to perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

[0144] The beneficial effect of the present invention is that by acquiring a visual image of seafood and performing type matching and state marking, seafood of different types and states can be accurately identified, and corresponding processing control instructions (including sorting, cutting, and cleaning instructions) can be generated, thereby ensuring that the robot can accurately process the specific seafood according to its characteristics, thereby improving the accuracy and efficiency of processing. By calculating and feedback controlling the sorting energy consumption, the robot's sorting action is optimized, energy consumption is reduced, and sorting efficiency is improved. In addition, the feedback of the fixed cutting instruction ensures the stability and accuracy of the cutting process, thereby improving the cutting quality. By dynamically adjusting the cleaning process and controlling the robotic arm, flexible optimization of the cleaning operation is achieved, the cleaning effect and efficiency are improved, and the accuracy of the cleaning process is ensured by the feedback mechanism, avoiding the problem of excessive or insufficient cleaning. By acquiring and analyzing environmental perception data, position extraction, obstacle avoidance path planning, and target position availability evaluation can be effectively performed, ensuring that the robot can safely and effectively perform loading operations in complex environments, reducing collision risks and transportation errors, and improving operational safety and efficiency. By building a multi-robot collaborative network and a real-time monitoring and early warning system, the coordination and collaboration capabilities between multiple robots are enhanced, ensuring the overall efficiency and stability of the system, and being able to discover and solve potential problems in a timely manner, thereby improving the reliability and work efficiency of the entire system. Therefore, the present invention improves the robot control accuracy and adaptability through accurate seafood identification, dynamic energy consumption management, real-time cleaning adjustment, environmental perception and obstacle avoidance, and the construction of a multi-robot collaborative network.

[0145] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0146] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A robot control method based on data model driving, characterized in that: The following steps are involved: Step S1: obtaining a visual photograph of seafood; matching the seafood visual photograph with the type of seafood to generate seafood matching recognition data; Marking the seafood status of the seafood segmentation map according to the seafood matching and identification data to generate seafood status marking data; generating processing control instructions for the aquaculture vessel processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions; Step S2: When the aquaculture vessel processing and loading robot receives a sorting instruction, it calculates the sorting energy consumption of the seafood matching identification data to obtain the sorting energy consumption data; the sorting action control feedback of the robot arm is performed on the aquaculture vessel processing and loading robot through the sorting energy consumption data, thereby generating the robot sorting instruction feedback data; based on the cutting instruction, the fixed cutting instruction feedback is performed on the aquaculture vessel processing and loading robot, thereby generating the robot cutting instruction feedback data; Step S2 includes the following steps: Step S21: When the processing and loading robot of the aquaculture vessel receives the sorting instruction, it performs a seafood characteristic analysis on the seafood matching identification data to generate seafood characteristic attribute data, wherein the seafood characteristic analysis includes type analysis, size analysis and weight analysis; Step S22: Calculate the sorting energy consumption of the aquaculture vessel processing and loading robot according to the seafood characteristic attribute data to obtain the sorting energy consumption data; sort the seafood characteristic attribute data according to the sorting energy consumption data to generate the seafood sorting priority data; Step S23: performing sorting path planning based on the seafood sorting priority ranking data to generate seafood sorting planning path data; controlling the robot arm sorting action of the aquaculture vessel processing and loading robot through the seafood sorting planning path data, thereby generating robot sorting robot arm control data; Step S24: providing control feedback to the robot sorting mechanical arm control data, and generating robot sorting instruction feedback data; providing fixed cutting instruction feedback to the aquaculture vessel processing and feeding robot based on the cutting instruction, thereby generating robot cutting instruction feedback data; Step S3: When the aquaculture vessel processing and loading robot receives a cleaning instruction, the aquaculture vessel processing and loading robot is dynamically adjusted based on the cleaning instruction to generate cleaning dynamic adjustment data; the aquaculture vessel processing and loading robot is controlled to perform a cleaning action of a mechanical arm according to the cleaning dynamic adjustment data, thereby generating robot cleaning mechanical arm control data; and the cleaning instruction is fed back to the robot cleaning mechanical arm control data, thereby generating robot cleaning instruction feedback data; Step S4: Acquire robot environmental perception data; extract position information from the robot environmental perception data to obtain robot location data and loading target location data; perform regional obstacle avoidance path planning based on the robot location data and loading target location data to generate loading area obstacle avoidance path planning data; perform availability evaluation on the loading target location data to generate loading target location availability evaluation data; perform loading transportation control feedback on the loading area obstacle avoidance path planning data based on the loading target location availability evaluation data to generate robot loading transportation control feedback data; Step S5: construct a multi-robot collaborative network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

2. The robot control method based on data model driving according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: using the camera on the processing and loading robot of the aquaculture vessel to obtain a visual picture of the seafood; Step S12: performing image preprocessing on the seafood visual image to generate a standard seafood visual shooting image, wherein the image preprocessing includes image denoising, image filtering, image brightness enhancement and image resolution improvement; Step S13: segmenting the edge contour of the standard seafood visual photograph to generate a seafood segmentation map; matching the seafood segmentation map with the type of seafood based on a preset seafood database to generate seafood matching recognition data; Step S14: Mark the seafood status of the seafood segmentation map according to the seafood matching and identification data to generate seafood status marking data; generate processing control instructions for the aquaculture ship processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions.

3. The robot control method based on data model driving according to claim 1 is characterized in that: Feedback of fixed cutting instructions for the aquaculture vessel processing and feeding robot based on cutting instructions includes: When the processing and loading robot of the aquaculture vessel receives a cutting instruction, it updates the standard seafood visual image through the robot sorting instruction feedback data to generate a seafood sorting update image; it parses the cutting instruction to confirm the cutting position, cutting depth and cutting shape of the seafood; Perform seafood location positioning on seafood sorting update photos to generate seafood location positioning data; perform seafood location fixation stability analysis based on seafood location positioning data to generate seafood retention stability analysis data; Based on the cutting position, cutting depth and cutting shape of seafood, the cutting tool of the aquaculture vessel processing and loading robot is selected to generate cutting tool selection data; the cutting path is planned for the cutting tool selection data using the seafood retention stability analysis data to generate seafood cutting planning path data; The seafood cutting planning path data is used to control the cutting action of the robot arm of the aquaculture vessel processing and loading robot, thereby generating robot cutting robot arm control data; the robot cutting robot arm control data is fed back with fixed cutting instructions, thereby generating robot cutting instruction feedback data.

4. The robot control method based on data model driving according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: when the aquaculture vessel processing and loading robot receives a cleaning instruction, it performs cleaning requirement instruction analysis on the cleaning instruction to generate cleaning instruction analysis data, wherein the cleaning requirement instruction analysis includes cleaning time analysis, cleaning medium analysis and cleaning intensity analysis; Step S32: updating the view of the seafood sorting update shot image based on the robot cutting instruction feedback data to generate the seafood cutting update shot image; performing surface stain recognition on the seafood cutting update shot image to generate surface stain recognition data; Step S33: selecting a cleaning tool for the aquaculture vessel processing and loading robot based on the surface stain recognition data to generate cleaning tool selection data; dynamically adjusting the cleaning of the aquaculture vessel processing and loading robot based on the cleaning tool selection data to generate cleaning dynamic adjustment data; Step S34: Controlling the cleaning action of the robot arm of the aquaculture vessel processing and loading robot according to the dynamic cleaning adjustment data, thereby generating robot cleaning robot arm control data; providing cleaning instruction feedback to the robot cleaning robot arm control data, thereby generating robot cleaning instruction feedback data.

5. The robot control method based on data model driving according to claim 4 is characterized in that: Using cleaning tool selection data to dynamically adjust the cleaning of the aquaculture vessel processing and feeding robot includes: The cleaning image of the processing and loading robot of the aquaculture vessel is taken in real time using the cleaning tool selection data to generate a seafood cleaning image set; the seafood cleaning image set is segmented to generate a non-stained area image set and a stained area image set; Perform stain area change analysis on the stain area image set to generate stain area change data; perform surface stain residual rate calculation on the stain area image set based on the stain area change data to obtain surface stain residual rate data on the stain area; perform surface finish analysis on the non-stain area image set to generate non-stain area surface finish data; The cleaning completion degree is analyzed by the surface stain residual rate data of the stained area and the surface finish data of the non-stained area to generate the cleaning completion degree data; the cleaning completion degree data is compared with the preset cleaning completion degree threshold value. When the cleaning completion degree data is greater than or equal to the preset cleaning completion degree threshold value, the cleaning water pressure of the aquaculture vessel processing and feeding robot is reduced to generate low completion degree cleaning adjustment data; When the cleaning completion data is less than the preset cleaning completion threshold, the cleaning time of the aquaculture ship processing and loading robot is extended to generate high-completion cleaning adjustment data; the low-completion cleaning adjustment data and the high-completion cleaning adjustment data are integrated to generate dynamic cleaning adjustment data.

6. The robot control method based on data model driving according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Acquire robot environment perception data; Step S42: extracting position information from the robot's environmental perception data to obtain the robot's position data and the loading target position data; Step S43: performing path area obstacle identification on the robot location data and the loading target location data to generate path area obstacle identification data; performing regional obstacle avoidance path planning on the robot location data and the loading target location data using the path area obstacle identification data to generate loading area obstacle avoidance path planning data; Step S44: perform availability evaluation on the loading target position data to generate loading target position availability evaluation data; perform loading transport control feedback on the loading area obstacle avoidance path planning data through the loading target position availability evaluation data to generate robot loading transport control feedback data.

7. The robot control method based on data model driving according to claim 6 is characterized in that: Step S44 includes the following steps: Step S441: using sensors to collect data of the feeding area based on the feeding target position data, and obtaining monitoring data of the feeding area, wherein the monitoring data of the feeding area includes available space data of the feeding area, temperature status data of the feeding area, and equipment operation status data of the feeding area; Step S442: performing a loading position availability status analysis based on the loading area available space data, the loading area temperature status data, and the loading area equipment operation status data, and generating loading target position availability evaluation data, wherein the loading target position availability evaluation data includes regional available status data and regional unavailable status data; when the loading position availability status is regional unavailable status data, the aquaculture vessel processing loading robot is controlled to wait in situ until the loading position availability status is regional available status data; Step S443: when the availability status of the loading position is regional available status data, a spatial stereoscopic area analysis is performed on the available space data of the loading area to generate available space stereoscopic data, wherein the available space stereoscopic data includes two-dimensional plane area data and three-dimensional height stereoscopic data; Step S444: Compare the two-dimensional plane area data and the three-dimensional height stereo data. When the two-dimensional plane area data is larger than the three-dimensional height stereo data, the aquaculture vessel processing loading robot is placed and loaded according to the loading area obstacle avoidance path planning data, and the robot loading and transportation control feedback data is generated; when the two-dimensional plane area data is smaller than the three-dimensional height stereo data, the aquaculture vessel processing loading robot is stacked and loaded according to the loading area obstacle avoidance path planning data, and the robot loading and transportation control feedback data is still generated.

8. The robot control method based on data model driving according to claim 1 is characterized in that: Step S5 includes the following steps: Step S51: integrating the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to generate robot comprehensive control feedback data; Step S52: performing multi-robot coordinated collaborative task allocation on the robot integrated control feedback data through data model technology to generate a multi-robot collaborative control matrix; Step S53: constructing a robot collaborative network based on the multi-robot collaborative control matrix to obtain a robot control network; performing real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

9. A robot control system driven by a data model, characterized in that: The robot control method based on data model driving according to claim 1 is used to execute the robot control method based on data model driving according to claim 1, and the robot control system based on data model driving comprises: The instruction division module is used to obtain visual images of seafood; match the seafood visual images with seafood types to generate seafood matching identification data; mark the seafood status of the seafood segmentation images according to the seafood matching identification data to generate seafood status marking data; generate processing control instructions for the aquaculture ship processing and loading robot based on the seafood status marking data to obtain robot processing control instructions, wherein the robot processing control instructions include sorting instructions, cutting instructions and cleaning instructions; The sorting and cutting control module is used to calculate the sorting energy consumption of the seafood matching identification data when the aquaculture ship processing and loading robot receives the sorting instruction, and obtain the sorting energy consumption data; to control the sorting action of the robot arm of the aquaculture ship processing and loading robot through the sorting energy consumption data, so as to generate the robot sorting instruction feedback data; to feedback the fixed cutting instruction to the aquaculture ship processing and loading robot based on the cutting instruction, so as to generate the robot cutting instruction feedback data; A cleaning control module is used for dynamically adjusting the cleaning of the aquaculture vessel processing and loading robot based on the cleaning instruction when the aquaculture vessel processing and loading robot receives a cleaning instruction, thereby generating cleaning dynamic adjustment data; controlling the cleaning action of the robot arm of the aquaculture vessel processing and loading robot according to the cleaning dynamic adjustment data, thereby generating robot cleaning robot arm control data; and providing cleaning instruction feedback to the robot cleaning robot arm control data, thereby generating robot cleaning instruction feedback data; The feeding path planning module is used to obtain the robot's environmental perception data; extract the position information of the robot's environmental perception data to obtain the robot's position data and the feeding target position data; perform regional obstacle avoidance path planning based on the robot's position data and the feeding target position data to generate the feeding area obstacle avoidance path planning data; perform availability evaluation on the feeding target position data to generate the feeding target position availability evaluation data; perform feeding transportation control feedback on the feeding area obstacle avoidance path planning data based on the feeding target position availability evaluation data to generate the robot feeding transportation control feedback data; The robot collaboration monitoring module is used to construct a multi-robot collaboration network based on the robot sorting instruction feedback data, the robot cutting instruction feedback data, the robot cleaning instruction feedback data and the robot loading and transportation control feedback data to obtain a robot control network; and to perform real-time robot control monitoring and early warning on the robot control network to execute robot control operations.

Images