Method, device, readable medium, electronic device and cleaning device for cleaning a mesh

Abstract

This application belongs to the field of marine aquaculture technology, disclosing a method, apparatus, readable medium, electronic equipment, and cleaning equipment for netting cleaning. This solution considers the type, degree of adhesion, and location of the attached organisms during the netting cleaning strategy, allowing for differentiated treatment of different types, degrees of adhesion, and locations of attachments on the netting surface during the cleaning process. This achieves rational allocation of cleaning resources and priority cleaning of key areas, significantly improving cleaning efficiency and reducing cleaning energy consumption and maintenance costs. Furthermore, by introducing an environmental factor, areas prone to rapid reattachment can be cleaned specifically, helping to extend the cleaning cycle, slow down the reattachment process, and further enhance the overall safety and economy of the aquaculture system.

Description

Technical Field

[0001] This application relates to the field of marine aquaculture technology, and in particular to a method, apparatus, readable medium, electronic device, and cleaning equipment for cleaning netting. Background Technology

[0002] With the rapid development of the marine economy, deep-sea aquaculture, due to its advantages such as excellent water quality and vast aquaculture space, has become an important direction for the transformation of aquaculture towards large-scale and intensive operations. Cage aquaculture, as one of the core models of deep-sea aquaculture, relies heavily on the performance of its key equipment—the netting—which directly affects aquaculture efficiency and safety. However, in the complex marine environment of the deep sea, the netting, constantly immersed in seawater, faces serious problems of biological and abiotic attachment: algae (such as seaweed, Sargassum, and Ulva), shellfish (such as oysters and mussels), and various other attachments continuously accumulate on the netting surface, forming a dense covering layer.

[0003] These deposits clog the mesh of the netting, significantly reducing water permeability and hindering water exchange. This leads to a decrease in dissolved oxygen transport efficiency within the net cage, consequently affecting the healthy growth of aquaculture organisms. Furthermore, the accumulation of deposits increases the weight of the netting, reducing the overall load-bearing capacity of the cage. In extreme cases, this can cause the cage to deform, tilt, or even disintegrate due to overloading, threatening aquaculture safety. Some deposits (such as acidic metabolic products of shellfish and corrosive substances in sludge) can also accelerate the aging and corrosion of the netting materials (such as polyethylene and nylon), shortening the netting's lifespan and increasing maintenance costs. In addition, uneven distribution of deposits can disrupt the balance and stability of the net cage, further reducing aquaculture efficiency.

[0004] Therefore, the netting needs to be cleaned. Related technologies mainly rely on manual cleaning or the use of high-pressure water guns / mechanical scrubbing devices. Manual cleaning requires divers or workers to approach the netting from a boat and manually remove the attached materials. This method is extremely inefficient, requiring a large amount of manpower and time per cleaning session, making it difficult to meet the high-frequency cleaning needs of deep-sea aquaculture. Furthermore, it is constrained by natural factors such as weather (e.g., wind, waves, heavy rain) and tides (e.g., low tide limiting the operating window), making it difficult to guarantee the continuity and reliability of cleaning operations, often resulting in substandard netting cleanliness. While high-pressure water guns and mechanical scrubbing devices partially replace manual cleaning, these devices mostly use a "non-discriminatory" physical rinsing mode. Due to a lack of targeting, the cleaning effect is limited, especially when dealing with unevenly distributed and different types of attached materials, making precise cleaning impossible. Summary of the Invention

[0005] To address the aforementioned problems, this application provides a method, apparatus, readable medium, electronic device, and cleaning equipment for cleaning mesh.

[0006] According to an embodiment of this application, a mesh cleaning method is disclosed. The mesh cleaning method includes: acquiring image data of the mesh; based on the image data, acquiring the adhesion degree, adhesion type, and location of the attachments in each target area on the mesh, where the adhesion degree indicates the difficulty of removing the attachments from the mesh, and the location indicates the position of the target area on the mesh, and the mesh contains multiple target areas; and based on the adhesion degree, attachment type, location, and environmental coefficient of the attachments in each target area on the mesh, determining a cleaning strategy for the mesh, where the cleaning strategy includes at least one of the cleaning priority, cleaning method, and cleaning intensity for each target area on the mesh, where the cleaning priority indicates the priority order of cleaning the target areas, the cleaning method indicates the cleaning means for the target areas, the cleaning intensity indicates the cleaning effort for the target areas, and the environmental coefficient reflects the degree of influence of environmental parameters on the risk of re-attachment of the attachments and the growth rate of the attachments.

[0007] In some embodiments, a cleaning strategy for the mesh is determined based on the degree of adhesion, type of adhesion, location of the area, and environmental coefficient of the adhesions in each target area of ​​the mesh. This includes: determining a type coefficient for each target area based on the type of adhesions in each target area of ​​the mesh, where the type coefficient is used to quantify the removal difficulty corresponding to the type of adhesion; determining a location coefficient for each target area based on the location of the area, where the location coefficient is used to quantify the importance of the target area relative to the entire mesh cleaning task; and determining the cleaning priority of each target area based on a weighted calculation of the degree of adhesion, type coefficient, location coefficient, and environmental coefficient of the adhesions in the target area.

[0008] In some embodiments, determining the position coefficient of each target area based on its location on the net includes: determining a boundary factor for each target area based on its distance from the edge and / or preset boundary of the net, whereby the boundary factor represents the impact of the area's proximity to the edge and / or preset boundary of the net; determining a depth factor for each target area based on the water depth characteristics at its location, whereby the depth factor represents the impact of the seawater depth at its location; determining an external factor for each target area based on at least one of the seawater current velocity, water temperature, and season of the seawater environment in which the target area is located, whereby the external factor represents the impact of the external environment on the location; and determining the position coefficient of each target area on the net based on its boundary factor, depth factor, and external factor.

[0009] In some embodiments, determining the type coefficient of each target area based on the type of attachments in each target area on the net includes: determining the type weight coefficient of each target area based on the type of attachments in each target area on the net, wherein the type weight coefficient is used to reflect the removal difficulty corresponding to the type of attachments and the degree of influence on the hydrodynamic performance of the net; calculating the total area of ​​each target area on the net and the area of ​​attachments in each target area; and determining the type coefficient corresponding to each target area on the net based on the type weight coefficient of the target area and the ratio of the area of ​​attachments to the total area.

[0010] In some embodiments, obtaining the attachment type of the attachments in each target region on the mesh based on image data includes: preprocessing the image data to obtain a target image; determining the attachment type label matching each pixel in the target image to obtain the classification result corresponding to each pixel in the target image; dividing the target image into multiple target regions based on the classification result, and obtaining the attachment type of the attachments in each target region, wherein the attachments in each target region have the same attachment type.

[0011] In some embodiments, obtaining the adhesion degree of the attachment in each target area on the mesh includes: calculating the total area of ​​each target area on the mesh and the area of ​​the attachment corresponding to the attachment type in each target area, and determining the attachment coverage rate of each target area based on the ratio of the attachment area of ​​each target area to the total area; determining the attachment thickness and attachment strength of the attachment in the normal direction on each target area, wherein the attachment strength is used to represent the bonding strength between the attachment and the mesh; and determining the adhesion degree of the attachment in each target area on the mesh based on the attachment coverage rate, attachment thickness and attachment strength of the target area.

[0012] In some embodiments, obtaining the adhesion degree of the attachment in each target area on the mesh fabric further includes: obtaining the attachment area and attachment boundary length in each target area on the mesh fabric; and for each target area on the mesh fabric, obtaining the shape factor of the attachment in the target area based on the attachment area and attachment boundary length. Determining the adhesion degree of the attachment in the target area based on the attachment coverage, attachment thickness, and attachment strength of the target area includes: determining the adhesion degree of the attachment in the target area based on the attachment coverage, attachment thickness, attachment strength, and shape factor of the target area.

[0013] In some embodiments, environmental parameters include the season, the water temperature of the seawater environment in which the net is located, the seawater current velocity, and the nutrient concentration.

[0014] In some embodiments, a cleaning strategy for the mesh fabric is determined based on the degree of adhesion, type of adhesion, location of the area, and environmental factors of the adhesions in each target area of ​​the mesh fabric. This includes: determining the cleaning priority of each target area of ​​the mesh fabric based on the degree of adhesion, type of adhesion, location of the area, and environmental factors; determining the cleaning intensity of each target area of ​​the mesh fabric based on the cleaning priority, degree of adhesion, and type of adhesion; and determining the cleaning method of each target area of ​​the mesh fabric based on the type of adhesions in each target area.

[0015] In some embodiments, after determining the cleaning priority of each target area on the mesh based on the degree of adhesion, type of adhesion, location of the area, and environmental coefficient of the attachments in each target area on the mesh, the method further includes: determining the cleaning path of the mesh based on the cleaning priority of each target area on the mesh and the ant colony algorithm; and cleaning the mesh based on the cleaning path and the cleaning method and cleaning intensity corresponding to each target area on the mesh.

[0016] According to an embodiment of this application, a mesh cleaning device is also disclosed. This mesh cleaning device includes a first acquisition module, a second acquisition module, and a strategy determination module. The first acquisition module is configured to acquire image data of the mesh. The second acquisition module is configured to acquire, based on the image data, the adhesion degree, attachment type, and location of attachments in each target area on the mesh. The adhesion degree indicates the difficulty of removing the attachments from the mesh, and the location indicates the position of the target area on the mesh. The mesh contains multiple target areas. The strategy determination module is configured to determine a cleaning strategy for the mesh based on the adhesion degree, attachment type, location, and environmental coefficient of the attachments in each target area on the mesh. The cleaning strategy includes at least one of the following: cleaning priority, cleaning method, and cleaning intensity for each target area on the mesh. The cleaning priority indicates the priority order of cleaning the target areas, the cleaning method indicates the cleaning means for the target areas, the cleaning intensity indicates the cleaning effort for the target areas, and the environmental coefficient reflects the influence of environmental parameters on the risk of re-attachment and the growth rate of the attachments.

[0017] According to an embodiment of this application, a computer-readable medium is also disclosed, on which a computer program is stored, which, when executed by a processor, implements the above-described mesh cleaning method.

[0018] According to an embodiment of this application, an electronic device is also disclosed, which includes one or more processors and a storage device for storing one or more programs. When the one or more programs are executed by one or more processors, the one or more processors implement the above-described mesh cleaning method.

[0019] According to an embodiment of this application, a mesh cleaning device is also disclosed. The mesh cleaning device includes a main body and the aforementioned electronic equipment. The main body is used to perform cleaning operations on the mesh, and the electronic equipment is installed in the main body.

[0020] The technical solutions provided by the embodiments of this application have at least the following beneficial effects: The scheme disclosed in this application considers the type, degree of adhesion, and location of the attached substances during the netting cleaning strategy. This allows for differentiated treatment of different types, degrees of adhesion, and locations of attached substances on the netting surface during the cleaning process, achieving rational allocation of cleaning resources and prioritizing the cleaning of key areas. This significantly improves cleaning efficiency and reduces cleaning energy consumption and maintenance costs. Furthermore, by introducing an environmental factor, areas prone to rapid re-attachment can be cleaned specifically, helping to extend the cleaning cycle, slow down the re-attachment process, and further enhance the overall safety and economy of the aquaculture system.

[0021] It should be understood that the above general description and the following detailed description are merely exemplary and do not limit this application. Attached Figure Description

[0022] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the principles of this application.

[0023] Figure 1 A flowchart of a mesh cleaning method according to an embodiment of this application is shown.

[0024] Figure 2 A flowchart of the type analysis sub-step of one embodiment of this application is shown.

[0025] Figure 3 A flowchart of the adhesion analysis sub-step of an embodiment of this application is shown.

[0026] Figure 4 A flowchart of the shape factor determination sub-step of an embodiment of this application is shown.

[0027] Figure 5 A flowchart of the location analysis sub-step of one embodiment of this application is shown.

[0028] Figure 6 A flowchart of the cleaning priority determination sub-step of an embodiment of this application is shown.

[0029] Figure 7 A block diagram of a mesh cleaning apparatus according to an embodiment of this application is shown.

[0030] Figure 8 A block diagram of an electronic device according to an embodiment of this application is shown.

[0031] Figure 9 A computer system architecture block diagram is shown for implementing some embodiments of this application.

[0032] The annotations in the attached figures are explained as follows: 700. Mesh cleaning device; 701. First acquisition module; 702. Second acquisition module; 703. Strategy determination module; 800. Computer equipment; 801. Processor; 802. Memory; 900. Computer system; 901. CPU; 902. ROM; 903. RAM; 904. Bus; 905. I / O interface; 906. Input section; 907. Output section; 908. Storage section; 909. Communication section; 910. Driver; 911. Removable medium. Detailed Implementation

[0033] To make the objectives, implementation methods, and advantages of this application clearer, exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided to make the description of this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. It should be noted that the brief descriptions of terminology in this application are merely for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0034] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined with "first" and "second" may explicitly or implicitly include one or more features.

[0035] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0036] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0037] In the past, manual cleaning of netting required divers or workers to approach the net cages by boat and manually remove the attached materials. This was extremely inefficient, with each cleaning requiring a large amount of manpower and taking a long time. It was difficult to meet the high-frequency cleaning needs of deep-sea aquaculture. Furthermore, the continuity and reliability of the cleaning operation were difficult to guarantee due to natural factors such as weather (e.g., wind, waves, heavy rain) and tides (e.g., low tide limiting the operating window), often resulting in the netting not meeting the cleanliness standards.

[0038] With the development of intelligent technology, some cleaning devices and robots have attempted to introduce automated methods such as mechanical brushing or high-pressure water jetting, which have improved cleaning efficiency to some extent. However, such equipment adopts a "non-discriminatory" physical rinsing mode (such as uniform water pressure and the same path), which makes it difficult to optimize cleaning parameters (such as water pressure, brushing force, and cleaning time) for different attachment scenarios (such as local high-density shellfish attachment and large-area algae coverage). This results in serious waste of resources (water, electricity, and time), and the treatment effect on complex attachment problems (such as the type and uneven distribution of attachments) is limited.

[0039] To this end, this application provides a mesh cleaning solution, which first acquires image data of the mesh, then, based on the acquired image data, obtains the degree of adhesion, type of adhesion, and location of the attachments in each target area of ​​the mesh, and finally, based on the degree of adhesion, type of adhesion, location of the attachments in each target area of ​​the mesh, and environmental factors, determines the cleaning strategy for the mesh.

[0040] By accurately identifying the deposits on the surface of the netting and formulating cleaning strategies based on the type, degree of adhesion, location, and environmental factors of the deposits, intelligent decision-making and regional priority scheduling of netting cleaning operations can be achieved, thereby improving netting cleaning efficiency, reducing waste of cleaning resources, extending the service life of net cages, and promoting the healthy growth of aquaculture organisms.

[0041] Figure 1 A flowchart illustrating a mesh cleaning method according to an embodiment of this application is shown. Figure 1 As shown, the mesh cleaning method includes at least the following steps S110-S130, which are described in detail below.

[0042] In step S110, image data of the mesh fabric is acquired.

[0043] Image data refers to visual information collected by a camera device that reflects the surface condition of the mesh, including but not limited to single-frame still photographs or multiple consecutive video frames. It records the characteristics of the attachments in various areas of the mesh and can be used for subsequent analysis and identification of the type, degree of attachment, and location of the attachments.

[0044] In some embodiments, an underwater high-definition camera or underwater image sensor can be used to acquire real-time images of the net surface to obtain image data of the net. The image acquisition device can have a high-definition resolution of not less than 1080p and a frame rate of not less than 30fps, thereby ensuring that the acquired image data has high clarity and detail reproduction capabilities, and can identify small attached objects. To further improve image quality, some embodiments employ an underwater multi-view image acquisition system to simultaneously or sequentially photograph the net surface from multiple directions, thereby mitigating image distortion and local occlusion problems caused by underwater light reflection, mud coverage, bubble interference, and other factors, and improving the reliability of subsequent analysis results.

[0045] In step S120, based on image data, the degree of adhesion, type of attachment, and location of the attachment in each target area on the mesh are obtained.

[0046] The degree of adhesion indicates the difficulty of removing the attached material from the mesh. The higher the degree of adhesion in the target area, the more difficult it is to remove the material from that area; conversely, the lower the degree of adhesion, the easier it is to remove the material from that area.

[0047] The "attachment type" indicates the category of attachments in the target area. These categories (i.e., attachment types) can include: algae (such as seaweed, sargassum, ulna, large brown algae, and green algae), shellfish (such as oysters, mussels, and scallop larvae), soft sludge (a mixture of mud, sand, organic debris, and microorganisms), biofilms (a slimy surface layer composed of bacteria, fungi, and their secreted extracellular polymers), barnacles (hard shell structures formed by crustaceans), and other possible marine attachment organisms or debris (such as microplastics and fishing net fragments). These categories can be distinguished based on morphology, composition, and attachment characteristics, providing a basis for selecting subsequent cleaning methods and intensities.

[0048] The "region location" indicates the exact location of the target area within the mesh. The mesh contains multiple target areas, each with a different location within the mesh. For example, a target area's location within the mesh could be: the upper left corner, upper right corner, lower left corner, lower right corner, the center area, the upper edge near the floating frame, the lower edge near the bottom anchor, the middle area along the longitudinal direction of the mesh, the area near the four edges, or based on a pre-defined grid number (e.g., grid R in row 3, column 5). 35 The specific location of the target area within the overall mesh structure reflects its spatial distribution and may influence water flow conditions, light exposure, and the distribution characteristics of attached materials, thus serving as an important reference in cleaning decisions.

[0049] In step S130, a cleaning strategy for the mesh is determined based on the degree of adhesion, type of adhesion, location of the area, and environmental factor of the attachments in each target area on the mesh.

[0050] The environmental coefficient reflects the influence of environmental parameters on the reattachment risk and growth rate of marine organisms. These environmental parameters include factors such as season, seawater temperature, ocean current velocity, and nutrient concentration. Season determines the peak reproductive period and activity patterns of marine organisms (such as algae, shellfish larvae, and biofilms); water temperature directly affects the metabolic activity and attachment rate; ocean current velocity affects the balance between nutrient transport and the erosion and shedding of marine organisms; and nutrient concentration drives the proliferation of algae and some microorganisms.

[0051] The environmental coefficient is calculated by integrating real-time or forecast data on parameters such as season, seawater temperature, current velocity, and nutrient concentration in the seawater environment where the net is located. It outputs a value characterizing the likelihood and rate of reattachment of marine organisms in the target area during the current or future period. During high-risk seasons (such as algal blooms) or under harsh environmental conditions (such as high temperature, high nutrient concentration, and low current velocity), the environmental coefficient is larger, indicating the need for earlier or more intensified cleaning to address the rapidly accumulating attachment risk. Conversely, under low-risk conditions, a smaller value can be used, appropriately delaying cleaning to conserve resources. Therefore, the environmental coefficient provides a dynamic environmental adaptation basis for net cleaning strategies, making cleaning plans more aligned with the actual changes in the complex deep-sea environment.

[0052] In some embodiments, via the relation: Calculate the environmental coefficient; where, This represents the environmental factor.

[0053] The cleaning strategy includes at least one of the following: cleaning priority, cleaning method, and cleaning intensity for each target area on the netting. Cleaning priority indicates the order in which target areas are cleaned, clarifying the sequence of cleaning each target area on the netting. The order is based on factors such as the degree of attachment of the attachment material, the type of attachment material, the location of the area, and environmental factors. Target areas with high cleaning priority indicate a high risk of re-attachment, a heavy degree of attachment, or a critical location (such as the center of the net cage or a high-flow-rate area). These areas require priority treatment during the cleaning operation to avoid a sharp decline in netting performance or an increase in aquaculture risks due to delayed cleaning.

[0054] The cleaning method is used to indicate the cleaning approach for the target area, specifying the removal method or technology employed for a particular target area. The selection is primarily based on the type and physical characteristics of the deposits. For example, for hard deposits (such as shellfish and barnacles) or heavy deposits, a mechanical impact method combining high-pressure water flow and rigid scrubbing can be used; for soft deposits (such as algae and loose sludge) or light deposits, a gentler method using low-pressure water flow or soft scrubbing can be used, thus ensuring cleaning effectiveness while minimizing damage to the mesh material.

[0055] Cleaning intensity is used to indicate the cleaning effort required for a target area, clearly defining the appropriate cleaning intensity or operational parameter level for that area. It can be reflected through quantifiable indicators such as cleaning time, number of cleaning cycles, spray pressure, and brushing speed. The cleaning intensity setting must balance removal requirements with mesh protection needs. For areas with heavy adhesion or rapid growth rates, the cleaning intensity can be increased (by extending cleaning time, increasing the number of cycles, or increasing pressure). For areas with light adhesion or fragile mesh, a lower intensity should be used to avoid excessive rinsing that could cause fiber aging or damage.

[0056] In some embodiments, in step S120, based on image data, the degree of attachment, type of attachment, and location of the attachment in each target area on the mesh are obtained. This includes an image preprocessing step that preprocesses the image data and an image analysis step that analyzes the target image obtained after preprocessing, so as to obtain the degree of attachment, type of attachment, and location of the attachment in each target area on the mesh.

[0057] Image preprocessing refers to a series of image optimization and data processing operations performed after the acquired image data is transmitted to a computer or edge computing device to improve the accuracy and efficiency of subsequent attachment identification and analysis. In some embodiments, the image preprocessing steps include a distortion correction sub-step, a denoising sub-step, and a contrast enhancement sub-step. In the distortion correction sub-step, the image data is distorted using the camera intrinsic parameter matrix to eliminate geometric distortions caused by underwater optical refraction, lens characteristics, etc., restoring the true spatial proportions of the image. In the denoising sub-step, median filtering or Gaussian filtering is used to denoise the image data, reducing or removing random noise interference caused by tiny particles, suspended matter, and bubbles in the water, improving image clarity, and ensuring that the outline and texture features of the attachments can be accurately extracted. In the contrast enhancement sub-step, histogram equalization is used to expand the grayscale dynamic range of the image, especially in low-light underwater environments, increasing the contrast between the attachments and the netting background, making previously difficult-to-distinguish attachment details more distinct, facilitating subsequent classification and identification.

[0058] In some embodiments, the image preprocessing step further includes a keyframe extraction sub-step and a change region focusing sub-step. In the keyframe extraction sub-step, frames are extracted from the continuously acquired video stream at set time intervals (e.g., 1 frame per second), and key frames representing the scene state are selected for analysis and recognition. This reduces data redundancy and improves processing efficiency while ensuring the continuity of temporal information. This keyframe extraction sub-step can be performed before the distortion correction, denoising, and contrast enhancement sub-steps. In the change region focusing sub-step, pixel differences between adjacent frames are compared using the frame difference method to extract significantly changing local regions in the image. This focuses the subsequent attachment identification processing on areas where attachment, detachment, or morphological changes may occur, thereby improving detection efficiency and response speed.

[0059] Image analysis refers to the process of analyzing the surface state of a mesh fabric after image preprocessing and obtaining a quality-optimized target image. This analysis utilizes a series of visual feature extraction and intelligent recognition algorithms to determine the adhesion degree, type, and location of the attachments in each target region. Specifically, the image analysis steps include adhesion degree analysis, type analysis, and location analysis sub-steps.

[0060] In the type analysis sub-step, based on the image data, the attachment type of the attachments in each target area on the mesh is obtained. In some embodiments, such as Figure 2 As shown, the type analysis sub-steps include the following steps S210-S250, which are described in detail below.

[0061] In step S210, the attachment type label matching each pixel in the target image is determined, and the classification result corresponding to each pixel in the target image is obtained.

[0062] Determining the attachment type label matching each pixel in a target image can be achieved using a semantic segmentation model to identify the attachment type for each pixel. In some embodiments, a convolutional neural network (CNN) is used to identify the attachment type label matching each pixel in the target image. This CNN is trained using a training set constructed from image samples containing various attachment types (such as algae, shellfish, soft sludge, biofilm, etc.) collected under different marine environments (different lighting, turbidity, and shooting angles). The CNN undergoes supervised learning using the training set, enabling it to learn the visual features (color, texture, shape, edges, etc.) of various attachments and their pixel-level correspondences, thereby outputting the attachment type label for each pixel, i.e., the classification result.

[0063] During the recognition process, the convolutional neural network receives the input target image and processes each pixel in the target image. Extract its feature vector and output the posterior probability that the pixel belongs to each attachment type k∈K. The attachment type label for a pixel is determined by maximizing the probability. ;in, For pixels Belongs to the type of attachment The probability, For pixels The feature vector, K={"algae","shellfish","soft sludge","biofilm","no attachment"}, is the set of attachment types. Based on this, a pixel-level semantic segmentation map of the entire target image can be obtained, intuitively presenting the spatial distribution areas of various attachments.

[0064] In step S220, based on the classification results, the target image is divided into multiple target regions, and the attachment type of the attachments in each target region is obtained. The attachments in each target region have the same attachment type.

[0065] Based on the classification results, the target image is divided into multiple target regions. That is, based on the classification results, adjacent pixels belonging to the same attachment type are aggregated into continuous regions, forming spatial segmentation blocks of the attachments, i.e., multiple target regions. In some embodiments, after forming the spatial segmentation blocks of the attachments, the non-maximum suppression (NMS) algorithm is used to remove overlapping or scattered redundant detection boxes / regions, retaining the most representative and highest-confidence region boundaries, thereby dividing the target image into multiple target regions. Each target area The attachments within the area are of the same type, which improves the accuracy of target area delineation and ensures the accuracy of subsequent analysis results (such as attachment degree assessment and cleaning strategy formulation).

[0066] In the above embodiments, the determination of attachment type labels is based on pixel-level examples. In some embodiments, the area belonging to the mesh surface in the target image can be divided into several grid-level attachment areas according to preset rules. Within each grid unit, the attachment type of the area is labeled based on the overall features or statistical information of the pixels contained therein, rather than classifying individual pixels one by one. Specifically, the mesh surface can first be mapped into several regular or irregular grids in the image space (such as rectangular grids evenly divided by rows and columns). Then, feature aggregation (such as calculating the main color tone, texture pattern, attachment coverage, and other statistics) is performed on the pixel set within each grid. Combined with a trained model (which can be a CNN, a traditional machine learning model, or a rule base), the main attachment type of the grid (such as "algae", "shellfish", "soft sludge", "biofilm", etc.) is determined. Compared with pixel-level recognition, grid-level recognition focuses more on the overall category classification of the area, which can reduce the amount of data processing and computational complexity while ensuring recognition accuracy.

[0067] In some embodiments, after obtaining the attachment type of the attachment in each target area, a type coefficient for each target area is further determined based on the attachment type of each target area on the mesh. This type coefficient is used to quantify the removal difficulty corresponding to the attachment type.

[0068] In some embodiments, the step of determining the type coefficient of each target area based on the type of attachment in each target area on the mesh fabric further includes: determining the type weight coefficient of each target area based on the type of attachment in each target area on the mesh fabric; calculating the total area of ​​each target area on the mesh fabric and the area of ​​attachment in each target area; and determining the type coefficient corresponding to each target area on the mesh fabric based on the type weight coefficient of the target area and the ratio of the attachment area to the total area. That is, the type analysis sub-step further includes... Figure 2 Steps S230-S250 are shown.

[0069] In step S230, the type weight coefficient of each target area is determined based on the type of attachment on each target area of ​​the mesh.

[0070] Among them, the type weight coefficient is used to reflect the difficulty of removing the attachment type and its effect on the mesh. Hydrodynamics The degree of impact on learning performance.

[0071] In some embodiments, different type weighting coefficients are pre-assigned to various types of attachments (such as algae, shellfish, soft sludge, biofilm, etc.) based on their removal difficulty (e.g., hard shellfish are more difficult to remove than soft algae) and their impact on the hydrodynamic performance of the net (e.g., dense shellfish attachments significantly hinder water exchange). Generally, attachment types that are difficult to remove and have a significant impact on the hydrodynamic performance of the net are assigned higher type weighting coefficients. For example, hard attachments such as shellfish are difficult to clean and have higher type weighting coefficients; while soft attachments such as algae and sludge have lower type weighting coefficients.

[0072] In step S240, the total area of ​​each target area on the mesh and the area of ​​the attached material in each target area are calculated.

[0073] The area of ​​the attached organism in the target region, that is, the pixel or actual area occupied by the attached organism in that region. For example, if the type of attached organism in the target region is algae, the area of ​​the attached organism is the pixel or actual area occupied by the algae in the target region.

[0074] In some embodiments, calculating the total area of ​​each target region on the mesh can be done by counting the total number of pixels in each target region on the mesh. The area of ​​the attachment in the target region can be the number of pixels corresponding to the attachment in the target region.

[0075] In step S250, for each target area on the mesh, the type coefficient corresponding to the target area is determined based on the type weight coefficient of the target area and the ratio of the area of ​​the attached material to the total area.

[0076] In some embodiments, based on relational expressions: Determine the type coefficient corresponding to the target region; where, For type coefficients, For type weight coefficients, For target area The middle type is The area of ​​the attached material, The total area of ​​the target region. The type coefficient comprehensively reflects the average difficulty of removing all attachments within the target region and their impact on the hydrodynamic performance of the netting. The larger the type coefficient, the higher the cleaning difficulty and risk caused by the type of attachments in the target region.

[0077] In the adhesion analysis sub-step, the adhesion degree of the adhesive in each target area on the mesh is obtained based on image data. In some embodiments, such as Figure 3 As shown, the adhesion degree analysis sub-step includes the following steps S310-S330, which are described in detail below.

[0078] In step S310, the total area of ​​each target area on the netting and the area of ​​the attachment corresponding to the attachment type in each target area are calculated, and the attachment coverage rate of each target area is determined based on the ratio of the attachment area of ​​each target area to the total area.

[0079] In some embodiments, based on the formula: Calculate the attachment coverage rate for each target area; where, For target area The coverage of attachments provides information about the target area. The distribution density of deposits in the medium is an important basis for subsequent cleaning decisions. For target area The area of ​​the pixels identified as attachments, i.e., the area of ​​the attachments. For target area The total area.

[0080] In step S320, the adhesion thickness and adhesion strength of the adhering material in the normal direction on each target area are determined.

[0081] The adhesion thickness of the deposit in the normal direction can be measured or estimated using methods such as laser rangefinders, structured light, or ultrasonic ranging. In some embodiments, the adhesion thickness in the normal direction is calculated using the reflection time of the laser / ultrasonic echo signal, as shown in the formula: ,in, For adhesion thickness, Speed ​​of light or speed of sound This represents the time difference of the echo signal.

[0082] Adhesion strength indicates the bond strength between the substrate and the mesh. The stronger the bond strength, the more difficult it is to remove the substrate; conversely, the weaker the bond strength, the easier it is to remove the substrate.

[0083] Adhesion strength can be determined based on the type of substrate, or it can be obtained through force sensors, empirical models, or fitting of historical data. Adhesion strength corresponds to the shear force or water flow impact force required per unit area during the cleaning process.

[0084] In step S330, for each target area on the mesh, the degree of adhesion of the attachment in the target area is determined based on the coverage, thickness and strength of the attachment in the target area.

[0085] In some embodiments, the degree of adhesion of each target area is quantitatively characterized based on the coverage of the attachment, the thickness of the attachment, and the strength of the attachment, and a comprehensive scoring function for the degree of adhesion is constructed. In step S330, for each target area on the mesh, a weighted calculation is performed based on the coverage of the attachment, the thickness of the attachment, and the strength of the attachment in the target area to obtain the degree of adhesion of the attachment in the target area.

[0086] Specifically, the comprehensive scoring function for adhesion degree is: ,in, For the degree of adhesion, The weighting coefficients for indicators such as adhesion coverage, adhesion thickness, and adhesion strength can be determined experimentally or optimized based on historical cleaning results. For adhesion thickness, For the coverage of attachments, This refers to the adhesion strength.

[0087] In some embodiments, the adhesion score is compared with a first preset threshold. Second preset threshold Based on the relationship, the degree of adhesion is classified into mild, moderate, and severe. For example: when At that time, the adhesion degree was determined to be light adhesion. At that time, the adhesion level was determined to be moderate. When the adhesion level is determined to be severe, the first preset threshold is used. Second preset threshold It can be set according to the actual situation.

[0088] In some embodiments, the determination of the degree of adhesion of the adhering material in the target area also considers the geometric characteristics of the adhering material; that is, the adhesion degree analysis sub-step also includes a shape factor determination sub-step. The shape factor helps measure the complexity of the adhering material, thus affecting the cleaning difficulty. For example... Figure 4 As shown, the shape factor determination sub-step includes the following steps S410-S430, which are described in detail below.

[0089] In step S410, the area of ​​the attachment in each target area on the mesh is obtained.

[0090] In step S420, the length of the attachment boundary in each target area on the mesh is obtained.

[0091] Among them, the attachment boundary length is used to represent the total length of the outline boundary line between the attachment area with attachment in the target area and the surrounding medium, such as the non-attached mesh background. It reflects the extent of the extension and morphological complexity of the attachment in spatial distribution.

[0092] In step S430, for each target area on the mesh, the shape factor of the attachment in the target area is obtained based on the area of ​​the attachment and the length of the attachment boundary in the target area.

[0093] In some embodiments, the shape factor of the attachment in the target area is obtained based on the area of ​​the attachment and the length of the attachment boundary in the target area, i.e., by the formula: Calculate the shape factor of the attachments in the target region; where, For shape factor, The length of the attachment boundary. The area of ​​the attached material.

[0094] The degree of adhesion of the attachment in the target area is determined based on the coverage, thickness, and strength of the attachment in the target area. This includes determining the degree of adhesion of the attachment in the target area based on the coverage, thickness, strength, and shape factor of the attachment in the target area. Accordingly, in step S330, the degree of adhesion of the attachment in the target area is determined based on the coverage, thickness, strength, and shape factor of the attachment in the target area. Specifically, this can be achieved by performing a weighted calculation based on the coverage, thickness, strength, and shape factor of the attachment in the target area to obtain the degree of adhesion of the attachment in the target area.

[0095] In some embodiments, the determination of the degree of adhesion of the attachments in the target area can also consider characteristics such as the number and average size of the attachment's connected regions to enrich the description of the spatial distribution of the attachments. In particular, by calculating the connected regions of each attachment area, it can be determined whether the shape of the attachments affects cleaning. That is, the determination of the shape factor also considers the number and average size of the attachment's connected regions.

[0096] In the location analysis sub-step, based on image data, the location of the attachment in each target area on the mesh is obtained. This location represents the position of the target area within the mesh; different locations have different location coefficients. In some embodiments, such as... Figure 5 As shown, the location analysis sub-step includes the following steps S510-S550, which are described in detail below.

[0097] In step S510, based on the image data, the location of the attachments in each target area on the mesh is obtained.

[0098] Based on image data, the location of the attachments in each target area on the mesh can be obtained. This can be achieved by using the known geometric parameters of the mesh, calibration points, or pre-set grid markers. First, a correspondence between the image coordinate system and the physical coordinate system of the mesh is established. Then, the mesh surface areas identified in the target image are mapped back to the actual mesh structure to ensure that each target area can uniquely correspond to a physical location on the mesh, i.e., the area location. The boundaries of each area are marked in the target image, so that each target area has a clear spatial range on the mesh. Finally, the location of each target area on the mesh is determined according to the coordinate system of the mesh or the preset position description rules.

[0099] In step S520, the boundary factor of each target area is determined based on the distance between the location of each target area on the mesh and the edge and / or preset boundary of the mesh.

[0100] Boundary factors reflect whether the target area is close to the edge of the netting, the boundary of the floating frame, or the connection point with other structures (such as anchor points). These locations are usually more prone to water flow disturbance, accumulation of deposits, or greater sensitivity to netting stability, and therefore require close attention. In other words, boundary factors are used to represent the impact of a region's location being close to the edge of the netting and / or a pre-defined boundary.

[0101] In step S530, the depth factor of each target area is determined based on the water depth characteristics at the location of each target area on the net.

[0102] Water depth characteristics can be the difference between deep and shallow water areas. Deep water areas may face higher water pressure, more complex water flow environment, or lower maintenance accessibility, making cleaning more difficult and necessary.

[0103] The depth factor is used to represent the impact of seawater depth at a given location, such as complex water flow environments and greater cleaning difficulties.

[0104] In step S540, the external factors of each target area are determined based on at least one of the seawater current velocity, water temperature and season of the seawater environment in which each target area is located on the net.

[0105] In some embodiments, external factors for each target area are determined based on the seawater flow velocity, water temperature, and season of the marine environment in which each target area is located on the netting. That is, various external uncertainties faced by the target area are considered, including seasonal attachment risks (such as algal bloom season), seawater flow velocity (attachments are easily washed away in high-speed flow areas but may also be aggravated by wear), and water temperature (which affects biological attachment activity).

[0106] External factors are used to represent the influence of the external environment on the location of a region. For example, the larger the external factors, the greater the urgency of cleaning.

[0107] In step S550, for each target area on the mesh, the position coefficient of the target area is determined based on the boundary factor, depth factor and external factor of the target area.

[0108] In some embodiments, the location coefficient of the target region is obtained by weighting the boundary factor, depth factor, and external factors of the target region. Specifically, the formula is: Quantify the location coefficients of the target area; where... For position coefficients, Boundary factor For depth factor, As external factors, These are weighting coefficients used to adjust the proportion of various factors. For example, if seasonal risk dominates in a certain sea area, the weighting coefficient can be increased. To amplify the influence of the external environment.

[0109] exist Figure 5 In the illustrated embodiment, the position coefficient is determined by weighted fusion of multiple factors such as boundary factor, depth factor and external factor, which helps to prioritize according to the actual situation of each target area, thereby achieving efficient and accurate cleaning scheduling.

[0110] In some embodiments, the step of determining the cleaning strategy for the mesh based on the degree of adhesion, type of adhesion, location of the area, and environmental factor of the attachments in each target area on the mesh includes a cleaning priority determination sub-step, a cleaning intensity determination sub-step, and a cleaning method determination sub-step.

[0111] In the cleaning priority determination sub-step, the cleaning priority of each target area on the mesh is determined based on the adhesion degree, type coefficient, location coefficient, and environmental coefficient of the attachments in each target area of ​​the mesh. In some embodiments, such as Figure 6 As shown, the cleaning priority determination sub-step includes the following steps S610-S630, which are described in detail below.

[0112] In step S610, based on the type of attachment in each target area on the mesh, a type coefficient for each target area is determined. This type coefficient is used to quantify the removal difficulty corresponding to the type of attachment.

[0113] Specifically, to determine the type coefficient of each target area based on the type of attachment on each target area of ​​the mesh, one can refer to the above... Figure 2 The description in the illustrated embodiment.

[0114] In step S620, based on the location of each target area on the mesh, a position coefficient for each target area is determined. This position coefficient is used to quantify the importance of the target area relative to the entire mesh cleaning task.

[0115] For details on how to determine the location coefficient based on regional location, please refer to the above. Figure 5 The description in the illustrated embodiment.

[0116] In step S630, for each target area on the mesh, a weighted calculation is performed based on the adhesion degree, type coefficient, location coefficient, and environmental coefficient of the attachments in the target area to determine the cleaning priority of the target area.

[0117] In some embodiments, based on relational expressions: Determine the cleaning priority of the target area; among which, As a cleaning priority, To adjust the weighting parameters of the degree of influence of each item, For type coefficients, For the degree of adhesion, For position coefficients, This is the environmental factor.

[0118] exist Figure 6 In the illustrated embodiment, the cleaning priority is calculated by weighting the adhesion degree, type coefficient, location coefficient, and environmental coefficient of the attachments in the target area. This cleaning priority accurately reflects the urgency of cleaning each target area. For example, when the adhesion degree is similar, target areas with high environmental coefficients and high seasonal growth rates can be prioritized for cleaning, thereby reducing the risk of rapid re-attachment.

[0119] After determining the cleaning priority for each target area on the mesh, the cleaning can be carried out according to the cleaning priority. The size of each target area is used to sort all target areas, thus determining the cleaning priority order for each target area of ​​the mesh. During cleaning, the cleaning order of each target area can be determined directly based on its cleaning priority, or it can be determined by combining other factors.

[0120] In the cleaning method determination sub-step, based on the type of attachment, the most suitable cleaning method for each target area on the mesh is matched through rule mapping or decision function. In some embodiments, the physical characteristics of the attachment are first determined based on the attachment type, and then the cleaning method is determined based on the physical characteristics of the attachment; finally, the cleaning method corresponding to the target area is output.

[0121] Specifically, firstly, based on the type of attachment obtained from the image analysis step, such as... ={"Algae", "Shellfish", "Soft Sludge", "Biofilm", "No Attachment"}, distinguishing their physical characteristics, such as soft or hard. For soft attachments, such as algae, low-pressure water flow or gentle brushing can be used to avoid damaging the mesh. For hard attachments, such as shellfish and hard-shelled organisms, high-pressure water flow combined with mechanical brushing or bubble cleaning can be used to break down the attachment structure through strong impact. For medium-weight attachments, such as soft sludge and biofilm, a combination of medium-pressure water flow and brushing can be used to balance removal intensity with mesh protection. Finally, through a mapping function... Automatic matching of deposit type to cleaning method, wherein, For attachment type, The cleaning method is specified. For example, if the type of attached organism is "shellfish", the output will be "high-pressure water flow combined with mechanical scrubbing or bubble cleaning"; if the type of attached organism is "algae", the output will be "low-pressure water flow or gentle scrubbing"; and if the type of attached organism is "soft sludge and biofilm", the output will be "suitable for medium-intensity water flow or scrubbing, often using a combination of medium-pressure water flow and scrubbing".

[0122] In the step of determining the cleaning intensity, the cleaning priority, the degree of adhesion, and the type of adhesion are taken into consideration. That is, the cleaning intensity of each target area on the mesh is determined based on the cleaning priority, the degree of adhesion, and the type of adhesion of each target area.

[0123] Specifically, the cleaning intensity of the target area is adjusted based on the degree of adhesion in the target area. For example, areas with high adhesion require more intensive cleaning, while areas with low adhesion can require less intensive cleaning.

[0124] Adjust the cleaning intensity of the target area based on the type of deposits in the target area. For example, harder attachments (such as shellfish and hard-shelled organisms) usually require a higher cleaning intensity, while softer attachments (such as algae and sludge) can be cleaned with a lower intensity.

[0125] Adjust the cleaning intensity of the target area based on the cleaning priority of the target area. The final cleaning intensity can be adjusted by using cleaning priorities. For example, for target areas where the degree and type of adhesion have the same overall impact on cleaning intensity, if the cleaning priority is high, the final cleaning intensity will be increased due to the amplified cleaning priority, ensuring that critical areas such as high-risk environmental areas and high-adhesion areas are prioritized and cleaned vigorously; if the cleaning priority is low, the final cleaning intensity will be reduced due to the decreased cleaning priority, avoiding over-cleaning that would waste resources or damage the mesh.

[0126] In some embodiments, the final cleaning intensity of the target area It can be calculated using the following formula: ,in, For cleaning intensity, For the degree of adhesion, For attachment type, As a cleaning priority, and These are the weighting coefficients. This indicates the degree of adhesion affecting cleaning intensity; the greater the adhesion, the greater the cleaning intensity. This indicates the degree to which the type of attachment affects the cleaning intensity; for example, the cleaning intensity for shellfish and hard-shell attachments is higher than that for algae. Cleaning Intensity By combining the type and degree of adhesion of the deposits with the cleaning priority, it ensures that strong cleaning is carried out in high-priority areas, while the cleaning intensity is appropriately reduced in low-priority areas.

[0127] In other embodiments, the cleaning intensity of the target area can be determined solely based on the degree of adhesion and the type of deposits in the target area. Specifically, based on Determine the cleaning intensity of the target area. For cleaning intensity, For the degree of adhesion, For attachment type, and These are the weighting coefficients.

[0128] In some embodiments, cleaning intensity may include not only cleaning force but also cleaning frequency and cleaning time. For high-value target areas, increasing the cleaning frequency and extending the time of each cleaning session will improve cleaning intensity. Lower target areas reduce cleaning frequency and shorten single cleaning time.

[0129] In some embodiments, the mesh cleaning method further includes a path planning step. In the path planning step, a cleaning path for the mesh is determined based on the cleaning priority of each target area on the mesh and an ant colony algorithm.

[0130] In some embodiments, during the path planning step, all target areas are first sorted from highest to lowest score according to the calculated cleaning priority, forming a set arranged by cleaning priority, such as... This dataset categorizes the mesh surface into multiple cleaning priority levels, with higher-scoring target areas ranked higher, indicating they should be prioritized in the cleaning task. Next, these target areas are divided into several groups, with higher-priority groups entering the cleaning execution phase first, while lower-priority groups are processed in subsequent batches to optimize overall operational efficiency and resource utilization. For example, target areas can be divided into groups based on the type and degree of adhesion of the attached material and their cleaning priority. Finally, based on the grouping, and considering the motion constraints of the actual cleaning execution mechanism, such as underwater cleaning robots, robotic arms, or mesh lifting and cleaning devices (e.g., maximum working radius, turning radius, obstacle avoidance requirements, underwater travel speed), path planning algorithms, such as ant colony optimization and Algorithm, are employed. Algorithms such as Dijkstra's algorithm or sampling-based RRT are used to generate cleaning paths for the mesh. Finally, the mesh is cleaned based on the cleaning paths and the corresponding cleaning methods and intensities for each target area on the mesh.

[0131] In some embodiments, to simplify the path optimization problem under complex constraints, an ant colony algorithm is adopted. By simulating the cooperative search of an ant colony, the algorithm gradually approaches the globally optimal path. Specifically, it includes an initialization sub-step, a pheromone update sub-step, a path selection sub-step, and a path optimization sub-step.

[0132] In the initialization sub-step, the number of ants N participating in the path search is set, and an initial position is assigned to each ant; initial pheromone values ​​are assigned to all possible path segments (i,j), that is, feasible paths from node i to node j. Reflecting the initial path attractiveness, It is usually a constant, such as τ0.

[0133] In the pheromone update sub-step, the ants dynamically update the pheromone based on the path quality during the path search process. Specifically, after each iteration, the pheromone on all paths decays according to the evaporation coefficient ρ (0 < ρ < 1), simulating natural evaporation and preventing old pheromones from excessively dominating the search. The formula is as follows: ,in, The evaporation coefficient of pheromones. Path segment The increase in pheromones indicates the quality of a path. For example, a shorter path, lower energy consumption, and better fulfillment of cleaning priorities result in a larger increase in pheromones, attracting more ants to choose that path. for Pheromones of the moment for The pheromone at a given moment, that is, the updated pheromone.

[0134] In the path selection sub-step, each ant chooses its next path based on a selection probability that is directly proportional to the pheromone concentration. ,in, For the probability of selection, Let ηij be the heuristic function for the path, such as ηij = 1 / dij, where diij is the path length, reflecting the immediate attractiveness of the path. The shorter the distance and the higher the priority, the larger the heuristic value. Path segment The concentration of pheromones reflects the quality of historical pathways. and To adjust the importance of pheromones and heuristic functions, Allowed is the set of nodes currently accessible to the ant, which satisfies the cleaning order constraint, such as prioritizing the inclusion of high-cleaning-priority regions.

[0135] In the path optimization sub-step, the path is continuously updated based on the ant's path search results until the maximum number of iterations is reached or the path quality converges, ultimately yielding the shortest cleaning path. Once path planning is complete, the order of the cleaning tasks and the specific execution path for each task are output.

[0136] In some embodiments, the mesh cleaning method further includes a monitoring and feedback step. In the monitoring and feedback step, the cleaning effect is monitored in real time using cameras and sensors. If the cleaning effect is found to be unsatisfactory, the cleaning strategy is adjusted promptly to ensure that the attached substances are completely removed. After cleaning, the cleaning effect is reviewed. If residual attached substances are found, the cleaning priority and cleaning method are readjusted, and supplementary cleaning is performed.

[0137] In some embodiments, after the cleaning operation is completed, image acquisition and attachment identification are repeated, and the degree of attachment is recalculated for each target area. and the degree of adhesion before cleaning Compare the results to evaluate the cleaning effectiveness. If certain areas... If the value still exceeds the set threshold, the subsequent cleaning priority will be increased accordingly, or the cleaning method and intensity will be adjusted. Simultaneously, the aforementioned weighting parameters will be adjusted based on historical cleaning results. Adaptive optimization is performed to improve the accuracy and robustness of subsequent cleaning strategies.

[0138] In addition, the degree of adhesion can be recalculated based on feedback data from different seasons. For environmental coefficient The calculation model was revised to better reflect the actual adhesion changes.

[0139] In some embodiments, the model may further include a model training and parameter calibration step. In the model training and parameter calibration step, calibration samples are collected under different types and degrees of adhesion to establish a correlation between the degree of adhesion and the actual cleaning difficulty. Simultaneously, adhesion evolution data under different seasons and water temperatures are collected to establish a mapping relationship between environmental coefficients and re-adhesion risk, thereby optimizing weight parameters and threshold settings to ensure a high correlation between the degree of adhesion, cleaning priority, and indicators such as actual cleaning time, energy consumption, or residual rate.

[0140] In summary, this application considers the type, degree of adhesion, and location of the attached substances during the netting cleaning strategy. This allows for differentiated treatment of different types, degrees of adhesion, and locations of substances on the netting surface during the cleaning process, achieving rational allocation of cleaning resources and prioritizing the cleaning of key areas. This significantly improves cleaning efficiency and reduces cleaning energy consumption and maintenance costs. Furthermore, by introducing an environmental factor, areas prone to rapid re-attachment can be cleaned specifically, helping to extend the cleaning cycle, slow down the re-attachment process, and further enhance the overall safety and economy of the aquaculture system.

[0141] See next. Figure 7 This embodiment provides a mesh cleaning device 700, which includes a first acquisition module 701, a second acquisition module 702, and a strategy determination module 703. The first acquisition module 701 is configured to acquire image data of the mesh. The second acquisition module 702 is configured to acquire, based on the image data, the adhesion degree, attachment type, and location of the attachments in each target area on the mesh. The adhesion degree indicates the difficulty of removing the attachments from the mesh, and the location indicates the position of the target area on the mesh. The mesh contains multiple target areas. The strategy determination module 703 is configured to determine a cleaning strategy for the mesh based on the adhesion degree, attachment type, location, and environmental coefficient of the attachments in each target area on the mesh. The cleaning strategy includes at least one of the following: cleaning priority, cleaning method, and cleaning intensity for each target area on the mesh. The cleaning priority indicates the priority order of cleaning the target areas, the cleaning method indicates the cleaning means for the target areas, and the cleaning intensity indicates the cleaning effort for the target areas. The environmental coefficient reflects the influence of environmental parameters on the risk of re-attachment and the growth rate of the attachments.

[0142] For detailed information on the first acquisition module 701, the second acquisition module 702, and the strategy determination module 703, please refer to the description in the above method steps, which will not be repeated here.

[0143] Understandably, the mesh cleaning device 700 may also include other modules, such as a path planning module, a monitoring and feedback module, a model training and parameter calibration module, etc.

[0144] See next. Figure 8 This embodiment provides a computer device 800, which includes one or more processors 801 and a memory 802. The memory 802 is used to store one or more programs. When one or more programs are executed by one or more processors 801, the computer device 800 implements the mesh cleaning method of this application.

[0145] Figure 9 The diagram shows a computer system architecture block diagram for implementing some embodiments of this application. It should be noted that... Figure 9 The computer system shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0146] like Figure 9 As shown, the computer system 900 includes a CPU (Central Processing Unit) 901, which can perform various appropriate actions and processes according to programs stored in ROM (Read-Only Memory) 902 or programs loaded from storage portion 908 into RAM (Random Access Memory) 903, such as performing the mesh cleaning method in the above embodiment. The RAM 903 also stores various programs and data required for system operation. The CPU 901, ROM 902, and RAM 903 are interconnected via a bus 904. An I / O (Input / Output) interface 905 is also connected to the bus 904.

[0147] The following components are connected to I / O interface 905: an input section 906 including a keyboard, mouse, etc.; an output section 907 including CRT (Cathode Ray Tube), LCD (Liquid Crystal Display), etc., and speakers, etc.; a storage section 908 including a hard disk, etc.; and a communication section 909 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 909 performs communication processing via a network such as the Internet. A drive 910 is also connected to I / O interface 905 as needed. Removable media 911, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 910 as needed so that computer programs read from them can be installed into storage section 908 as needed.

[0148] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing all or part of the steps shown in the flowcharts of the mesh cleaning method. In such embodiments, the computer program can be downloaded and installed from a network via communication section 909, and / or installed from removable medium 911. When the computer program is executed by central processing unit (CPU) 901, it performs various functions defined in the system of this application.

[0149] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0150] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0151] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0152] In another aspect, this application also provides a computer-readable medium, which may be included in the computer device described in the above embodiments; or it may exist independently and not assembled into the computer device. The computer-readable medium carries one or more programs that, when executed by the computer device, cause the computer device to implement the methods described in the above embodiments.

[0153] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0154] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this application.

[0155] In addition, this application also provides a net cleaning device, which includes a main body for performing cleaning operations on nets and electronic equipment installed on the main body. The main body can be an underwater cleaning robot or the like.

[0156] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this application are indicated by the appended claims.

Claims

1. A method for cleaning mesh, characterized in that, include: Acquire image data of the mesh garment; Based on the image data, the adhesion degree, attachment type, and location of the attachment in each target area on the mesh are obtained. The adhesion degree is used to indicate the difficulty of removing the attachment from the mesh, and the location of the area is used to indicate the position of the target area on the mesh. The mesh contains multiple target areas. Based on the degree of adhesion, type of adhesion, location of area, and environmental factor of the adhesions in each target area of ​​the mesh, a cleaning strategy for the mesh is determined. The cleaning strategy includes at least one of the following: cleaning priority, cleaning method, and cleaning intensity for each target area of ​​the mesh. The cleaning priority is used to indicate the order of priority for cleaning the target areas. The cleaning method is used to indicate the means of cleaning the target areas. The cleaning intensity is used to indicate the cleaning effort for the target areas. The environmental factor is used to reflect the degree of influence of environmental parameters on the risk of adhesion re-adhesion and the growth rate of adhesions.

2. The mesh cleaning method according to claim 1, characterized in that, The method for determining the cleaning strategy for the mesh fabric based on the degree of adhesion, type of adhesion, location of the area, and environmental factors in each target area of ​​the mesh fabric includes: Based on the type of attachments in each target area on the mesh, a type coefficient for each target area is determined. The type coefficient is used to quantify the removal difficulty corresponding to the type of attachment. Based on the location of each target area on the mesh, a position coefficient for each target area is determined. The position coefficient is used to quantify the importance of the target area relative to the entire mesh cleaning task. For each target area on the mesh, a weighted calculation is performed based on the adhesion degree of the attachment material in the target area, the type coefficient, the position coefficient, and the environmental coefficient to determine the cleaning priority of the target area.

3. The mesh cleaning method according to claim 2, characterized in that, The determination of the position coefficient of each target area based on the regional position of each target area on the mesh includes: Based on the distance between the location of each target area on the mesh and the edge and / or preset boundary of the mesh, a boundary factor for each target area is determined. The boundary factor is used to represent the impact of the location of the area being close to the edge and / or preset boundary of the mesh. Based on the water depth characteristics at the location of each target area on the net, a depth factor for each target area is determined, and the depth factor is used to represent the influence of the seawater depth at the location of the area. Based on at least one of the seawater current velocity, water temperature and season of the seawater environment in which each target area is located on the netting, the external factors of each target area are determined, and the external factors are used to represent the influence of the external environment on the location of the area. For each target region on the mesh, the position coefficient of the target region is determined based on the boundary factor, depth factor and external factor of the target region.

4. The mesh cleaning method according to claim 2, characterized in that, The determination of the type coefficient for each target area based on the type of attachment on each target area of ​​the mesh includes: Based on the type of attachments in each target area of ​​the netting, a type weighting coefficient for each target area is determined. The type weighting coefficient is used to reflect the removal difficulty and the degree of influence on the hydrodynamic performance of the netting corresponding to the type of attachment. Calculate the total area of ​​each target region on the mesh and the area of ​​the attached material in each target region; For each target area on the mesh, the type coefficient corresponding to the target area is determined based on the type weight coefficient of the target area and the ratio of the area of ​​the attached material to the total area.

5. The mesh cleaning method according to claim 1, characterized in that, Based on the image data, the attachment types of the attachments in each target area on the mesh are obtained, including: The image data is preprocessed to obtain the target image; Determine the attachment type label matching each pixel in the target image to obtain the classification result corresponding to each pixel in the target image; Based on the classification results, the target image is divided into multiple target regions, and the attachment type of the attachments in each target region is obtained. The attachments in each target region have the same attachment type.

6. The mesh cleaning method according to claim 5, characterized in that, The step of obtaining the adhesion degree of the attachment material in each target area on the mesh includes: Calculate the total area of ​​each target area on the mesh and the area of ​​the attachment corresponding to the attachment type in each target area, and determine the attachment coverage rate of each target area based on the ratio of the attachment area of ​​each target area to the total area; Determine the adhesion thickness and adhesion strength of the attachment in the normal direction on each target area, wherein the adhesion strength is used to represent the bonding strength between the attachment and the mesh. For each target area on the mesh, the degree of adhesion of the attachment in the target area is determined based on the coverage, thickness and strength of the attachment in the target area.

7. The mesh cleaning method according to claim 6, characterized in that, The step of obtaining the adhesion degree of the attachment in each target area on the mesh also includes: Obtain the area and boundary length of the attachment in each target region on the mesh. For each target area on the mesh, the shape factor of the attachment in the target area is obtained based on the area of ​​the attachment and the length of the attachment boundary in the target area; The determination of the adhesion degree of the attachment in the target area based on the attachment coverage, adhesion thickness, and adhesion strength of the attachment in the target area includes: The degree of adhesion of the attachments in the target area is determined based on the coverage, thickness, strength and shape factor of the attachments in the target area.

8. The mesh cleaning method according to claim 1, characterized in that, The environmental parameters include the season, the water temperature of the seawater environment in which the netting is located, the seawater flow rate, and the nutrient concentration.

9. The mesh washing method according to any one of claims 1 to 8, characterized in that, The method for determining the cleaning strategy for the mesh fabric based on the degree of adhesion, type of adhesion, location of the area, and environmental factors in each target area of ​​the mesh fabric includes: Based on the degree of adhesion, type of adhesion, location of area and environmental factor of the attachments in each target area of ​​the mesh, the cleaning priority of each target area of ​​the mesh is determined. The cleaning intensity of each target area on the mesh is determined based on the cleaning priority of each target area, the degree of adhesion of the adhering material, and the type of adhering material. Based on the type of attachment in each target area, the cleaning method for each target area on the mesh is determined.

10. The mesh cleaning method according to claim 9, characterized in that, After determining the cleaning priority of each target area on the mesh based on the adhesion degree, type of adhesion, location of the area, and environmental factors, the process further includes: Based on the cleaning priority of each target area on the mesh and the ant colony algorithm, the cleaning path of the mesh is determined; The mesh is cleaned based on the cleaning path and the cleaning method and intensity corresponding to each target area on the mesh.

11. A mesh washing device, characterized in that, include: The first acquisition module is configured to acquire image data of the mesh fabric. The second acquisition module is configured to acquire, based on the image data, the adhesion degree, attachment type, and location of the attachment in each target area on the mesh, wherein the adhesion degree is used to indicate the difficulty of removing the attachment from the mesh, and the location of the area is used to indicate the position of the target area on the mesh, and the mesh contains multiple target areas; The strategy determination module is configured to determine a cleaning strategy for the mesh based on the adhesion degree, type, location, and environmental coefficient of the attachments in each target area on the mesh. The cleaning strategy includes at least one of the cleaning priority, cleaning method, and cleaning intensity for each target area on the mesh. The cleaning priority indicates the order of priority for cleaning the target areas, the cleaning method indicates the means of cleaning the target areas, the cleaning intensity indicates the cleaning effort of the target areas, and the environmental coefficient reflects the influence of environmental parameters on the risk of re-attachment and the growth rate of the attachments.

12. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the mesh cleaning method as described in any one of claims 1 to 10.

13. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the mesh cleaning method as described in any one of claims 1 to 10.

14. A mesh cleaning device, characterized in that, include: The main body of the equipment is used to perform cleaning operations on the mesh. An electronic device installed in the main body of the device, as described in claim 13.

Images