A feeding path evaluation method and system based on intelligent perception

Abstract

The present invention discloses a feeding path evaluation method and system based on intelligent perception, including: obtaining water surface monitoring information, performing water surface environment perception according to the water surface monitoring information, and obtaining water surface environment perception information; obtaining underwater monitoring information, performing underwater environment perception according to the underwater monitoring information, and obtaining underwater environment perception information; obtaining feeding area information, combining underwater environment perception information and water surface environment perception information to formulate feeding paths, and obtaining candidate feeding path information; evaluating each candidate feeding path according to the candidate feeding path information, and selecting the final feeding path according to the evaluation result. The method helps feeding ships to perform feeding tasks intelligently, thereby improving breeding efficiency and reducing breeding costs.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned feeding ship path assessment, and in particular to a feeding path assessment method and system based on intelligent perception. Background Art

[0002] At present, the traditional aquaculture feeding method has the problems of feeding waste and unevenness, so the introduction of unmanned feeding fleets has become an important way to improve aquaculture efficiency and reduce costs. With the expansion of the scale of aquaculture, the limitations of traditional feeding methods have gradually emerged. Traditional methods are difficult to adapt to the changes in different water environments, resulting in increased feeding waste and reduced aquaculture benefits. To solve this problem, unmanned feeding fleets came into being, which have more flexible feeding capabilities. However, the existing unmanned feeding fleets have deficiencies in path planning and evaluation, and lack of intelligent perception, resulting in the failure to fully utilize the feeding efficiency.

[0003] In traditional aquaculture, feeding is usually done manually or with simple mechanical equipment, which results in uncontrollable feeding amount and large waste. In recent years, with the continuous advancement of technology, unmanned feeding fleets have begun to be used in aquaculture. These fleets can autonomously navigate in the waters and perform fixed-point and quantitative feeding operations, bringing new development opportunities to the aquaculture industry.

[0004] Therefore, it is an important issue to improve the environmental perception and path assessment capabilities of large-surface unmanned feeding fleets, thereby improving the overall efficiency of the feeding system and addressing the shortcomings of current feeding methods in the aquaculture industry. Summary of the invention

[0005] The present invention overcomes the defects of the prior art and provides a feeding path evaluation method and system based on intelligent perception, the important purpose of which is to help feeding ships to perform feeding tasks intelligently, thereby improving aquaculture efficiency and reducing aquaculture costs.

[0006] To achieve the above-mentioned purpose, the first aspect of the present invention provides a feeding path evaluation method based on intelligent perception, comprising:

[0007] Acquire water surface monitoring information, perform water surface environment perception according to the water surface monitoring information, and obtain water surface environment perception information;

[0008] Acquiring underwater monitoring information, and performing underwater environment perception according to the underwater monitoring information to obtain underwater environment perception information;

[0009] Obtain feeding area information, formulate feeding paths based on underwater environment perception information and surface environment perception information, and obtain candidate feeding path information;

[0010] Each candidate feeding path is evaluated according to the candidate feeding path information, and a final feeding path is selected according to the evaluation result.

[0011] In this solution, the water surface monitoring information is obtained and the water surface environment perception is performed according to the water surface monitoring information, specifically:

[0012] Acquire water surface monitoring information, pre-process the water surface monitoring information, acquire characteristic information of various water surface obstacles and ships and characteristic information of the objects to be fed based on big data retrieval, and form a comparative data set;

[0013] Constructing a water surface environment recognition model, and constructing a training data set according to the comparison data set to perform deep learning and training on the water surface environment recognition model, wherein the water surface environment recognition model includes a feature extraction layer, a target detection layer, and an environment recognition layer;

[0014] Performing feature extraction on the water surface monitoring information according to the feature extraction layer, extracting color features, contour features and texture features, and constructing a feature map according to the scale of each feature to obtain water surface monitoring feature information;

[0015] The water surface monitoring feature information is input into the target detection layer for target detection, an attention mechanism is introduced, and the attention scores of the feature maps of each scale are calculated by combining the water surface monitoring feature information with the attention mechanism, and an attention feature map is constructed to obtain attention feature map information;

[0016] Based on YOLOv3, an FPN feature pyramid is constructed, the water surface monitoring feature information and the attention feature map information are fused, the target position is predicted using Yolo Head, a target detection frame is generated, and target detection information is obtained;

[0017] Performing image segmentation on the detection target according to the target detection information, and performing feature extraction on the target object image after image segmentation to obtain feature information of the target object;

[0018] The target object feature information is input into the environment recognition layer for environment recognition, obstacles in the navigation direction of the feeding ship are identified, and water surface environment perception information is obtained.

[0019] In this solution, the underwater monitoring information is obtained and the underwater environment perception is performed according to the underwater monitoring information, specifically:

[0020] Obtain underwater monitoring information, obtain the appearance feature information of the object to be fed based on big data retrieval, and form an appearance comparison data set;

[0021] Extract features from underwater monitoring information, extract contour features and texture features of the target object, calculate similarity with the appearance comparison data set, and make a judgment with a preset threshold to analyze whether the detected object is an object to be fed, and obtain pre-identification information;

[0022] Marking the objects to be fed in combination with the pre-identification information and the underwater monitoring information, and performing quantity statistics, comparing the quantity statistics result with a preset threshold, and classifying the objects to be fed according to the quantity to obtain classification information;

[0023] Extracting the position information and underwater depth information of each category of the objects to be fed according to the underwater monitoring information, constructing a movement change trend graph according to the position change and depth change per unit time, analyzing the movement direction of the objects to be fed, and obtaining movement direction prediction information;

[0024] The underwater environment perception information is formed by combining the pre-identification information, category classification information and movement direction prediction information.

[0025] In this solution, the feeding area information is obtained, and the feeding path is formulated in combination with the underwater environment perception information and the surface environment perception information to obtain the candidate feeding path information, specifically:

[0026] Acquire feeding area information, divide the area to be fed into a plurality of sub-areas according to the feeding area information to obtain area division information, and set the feeding area of ​​the feeding boat according to the area division information to obtain the feeding area information;

[0027] Acquire water surface environment perception information, combine it with feeding area information to screen the navigable area, and obtain navigable area information;

[0028] Acquire underwater environment perception information, mark the feeding areas in combination with the feeding area information, determine the feeding level according to the number of objects to be fed in each area, and obtain feeding level analysis information;

[0029] Extracting the position change information of the object to be fed according to the underwater environment perception information, setting the initial point and the monitoring point, calculating the moving speed of the object to be fed, and obtaining the moving speed information;

[0030] Constructing a feeding path formulation model, inputting the surface environment perception information, underwater environment perception information, navigable area information and feeding level analysis information into the feeding path formulation model to formulate a path, and obtaining initial feeding path information;

[0031] According to the moving speed information and the initial feeding path information, the moving position of the object to be fed arriving at the feeding area on each initial path is predicted to obtain moving position prediction information;

[0032] Each initial feeding path is evaluated according to the mobile position prediction information to determine whether the path is invalid due to the change of the position of the object to be fed, and the path is screened to obtain candidate feeding path information.

[0033] In this solution, the input is input into the feeding path formulation model to formulate the path to obtain the initial feeding path information, and further includes:

[0034] According to the water surface environment perception information and the navigable area information, the obstruction area and the navigable area are marked to obtain the feeding area marking information;

[0035] The number of objects to be fed in the obstruction area is determined according to the feeding area marking information, the underwater environment perception information and the feeding level judgment information, and whether feeding is required is analyzed to obtain feeding judgment information;

[0036] A particle swarm optimization algorithm and an ant colony algorithm are introduced to formulate a feeding path, initial parameters of the particle swarm algorithm are set according to the feeding judgment information, the feeding level analysis information and the navigable area information, and the particle position and speed are randomly initialized to obtain an initial particle swarm;

[0037] Calculate the fitness value of each individual particle in each initial particle group, compare it with the preset threshold, select the optimal particle and population based on the judgment result, update the optimal solution of the individual and population, and perform initial global planning and iterative optimization;

[0038] Iterative control rules are preset, and the inertia weight and learning factor are dynamically adjusted according to the iterative control adjustment rules until the optimization result meets the termination conditions and the initial global planning information is obtained;

[0039] The initial pheromone of the ant colony is set according to the feeding judgment information, the feeding level analysis information and the navigable area information, the initial global planning information is used as the enhancement variable of the ant colony pheromone, the enhancement variable is allocated to the initial pheromone of the ant colony, and the initial optimization parameters are set;

[0040] Calculate the state transition probability of each path node, optimize the path through the state transition probability, and update the pheromone of each optimized path until the optimization result meets the iteration termination condition to obtain the initial feeding path information.

[0041] In this solution, each candidate feeding path is evaluated according to the candidate feeding path information, and the final feeding path is selected according to the evaluation result, specifically:

[0042] Acquire candidate feeding path information, perform feature extraction on the candidate feeding path information, extract path length features, path planning features, and feeding sequence features of each candidate feeding path, and obtain candidate feeding path feature information;

[0043] Calculate the feeding time and energy consumption of each candidate feeding path according to the candidate feeding path characteristic information, use it as a feeding efficiency judgment index, and compare it with a preset threshold to obtain feeding efficiency evaluation information;

[0044] Evaluate the rationality of the path planning of each candidate feeding path according to the candidate feeding path feature information, analyze the repetition rate of the driving area through the path planning features and the feeding sequence features, and obtain feasibility evaluation information;

[0045] Acquire underwater environment perception information, calculate feeding coverage according to the candidate feeding path information, and evaluate feeding effect in combination with the underwater environment perception information to obtain feeding effect evaluation information;

[0046] A weighted calculation is performed on each candidate feeding path according to the feeding efficiency evaluation information, the feasibility evaluation information and the feeding effect evaluation information, and a final feeding path is selected according to the weighted calculation result to obtain final feeding path information.

[0047] A second aspect of the present invention provides a feeding path evaluation system based on intelligent perception, the system comprising: a memory, a processor, the memory containing a feeding path evaluation method program based on intelligent perception, the feeding path evaluation method program based on intelligent perception when executed by the processor implements the following steps:

[0048] Acquire water surface monitoring information, perform water surface environment perception according to the water surface monitoring information, and obtain water surface environment perception information;

[0049] Acquiring underwater monitoring information, and performing underwater environment perception according to the underwater monitoring information to obtain underwater environment perception information;

[0050] Obtain feeding area information, formulate feeding paths based on underwater environment perception information and surface environment perception information, and obtain candidate feeding path information;

[0051] Each candidate feeding path is evaluated according to the candidate feeding path information, and a final feeding path is selected according to the evaluation result.

[0052] The present invention discloses a feeding path evaluation method and system based on intelligent perception, including: obtaining water surface monitoring information, performing water surface environment perception according to the water surface monitoring information, and obtaining water surface environment perception information; obtaining underwater monitoring information, performing underwater environment perception according to the underwater monitoring information, and obtaining underwater environment perception information; obtaining feeding area information, combining underwater environment perception information and water surface environment perception information to formulate feeding paths, and obtaining candidate feeding path information; evaluating each candidate feeding path according to the candidate feeding path information, and selecting the final feeding path according to the evaluation result. The method helps feeding ships to perform feeding tasks intelligently, thereby improving breeding efficiency and reducing breeding costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments or exemplary embodiments of the present invention, the drawings required for use in the embodiments or exemplary descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained according to the drawings without paying creative work.

[0054] Figure 1 A flow chart of a feeding path evaluation method based on intelligent perception provided by one embodiment of the present invention;

[0055] Figure 2 A flow chart of feeding path planning and adjustment provided in one embodiment of the present invention;

[0056] Figure 3 A block diagram of a feeding path evaluation system based on intelligent perception provided by an embodiment of the present invention;

[0057] 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

[0058] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0059] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific embodiments disclosed below.

[0060] Figure 1 A flow chart of a feeding path evaluation method based on intelligent perception provided by one embodiment of the present invention;

[0061] like Figure 1 As shown, the present invention provides a flow chart of a feeding path evaluation method based on intelligent perception, comprising:

[0062] S102, acquiring water surface monitoring information, and performing water surface environment perception according to the water surface monitoring information to obtain water surface environment perception information;

[0063] Acquire water surface monitoring information, pre-process the water surface monitoring information, acquire characteristic information of various water surface obstacles and ships and characteristic information of the objects to be fed based on big data retrieval, and form a comparative data set;

[0064] Constructing a water surface environment recognition model, and constructing a training data set according to the comparison data set to perform deep learning and training on the water surface environment recognition model, wherein the water surface environment recognition model includes a feature extraction layer, a target detection layer, and an environment recognition layer;

[0065] Performing feature extraction on the water surface monitoring information according to the feature extraction layer, extracting color features, contour features and texture features, and constructing a feature map according to the scale of each feature to obtain water surface monitoring feature information;

[0066] The water surface monitoring feature information is input into the target detection layer for target detection, an attention mechanism is introduced, and the attention scores of the feature maps of each scale are calculated by combining the water surface monitoring feature information with the attention mechanism, and an attention feature map is constructed to obtain attention feature map information;

[0067] Based on YOLOv3, an FPN feature pyramid is constructed, the water surface monitoring feature information and the attention feature map information are fused, the target position is predicted using Yolo Head, a target detection frame is generated, and target detection information is obtained;

[0068] Performing image segmentation on the detection target according to the target detection information, and performing feature extraction on the target object image after image segmentation to obtain feature information of the target object;

[0069] The target object feature information is input into the environment recognition layer for environment recognition, obstacles in the navigation direction of the feeding ship are identified, and water surface environment perception information is obtained.

[0070] It should be noted that, first, the water surface monitoring information is obtained and preprocessed. This information is obtained through various water surface monitoring equipment, including obstacles, ships, and objects to be fed on the water surface. Next, a water surface environment recognition model is constructed, which includes a feature extraction layer, a target detection layer, and an environment recognition layer. By comparing the data set to build a training data set, the model is deeply learned and trained to obtain a model that meets the expectations. In the feature extraction stage, the water surface monitoring information is extracted through the feature extraction layer, and the color, contour and texture features are extracted, and a feature map is constructed to obtain the water surface monitoring feature information. The obstacles, ships, and objects to be fed on the monitored water surface are detected through the water surface monitoring feature information, and their locations are locked. Subsequently, the target detection layer is used for target detection, and the attention mechanism is introduced. Through the attention mechanism, the attention scores of the feature maps of each scale are calculated, the attention feature map is constructed, the target detection information is obtained, and the accuracy and precision of the target detection are improved. Next, an FPN feature pyramid is constructed based on YOLOv3 to fuse the water surface monitoring feature information and the attention feature map information. Yolo Head is used to predict the target position, generate the target detection frame, and obtain the target detection information. Then, the detected target is segmented according to the target detection information to obtain the image of the target object. The features of these images are extracted to obtain the detailed feature information of the target object. Finally, the feature information of these target objects is input into the environment recognition layer for environment recognition, and obstacles in the navigation direction of the feeding boat are identified, so as to perceive the direction, position and category of the obstacles, providing highly accurate and detailed information for the subsequent feeding path evaluation.

[0071] S104, acquiring underwater monitoring information, and performing underwater environment perception according to the underwater monitoring information to obtain underwater environment perception information;

[0072] Obtain underwater monitoring information, obtain the appearance feature information of the object to be fed based on big data retrieval, and form an appearance comparison data set;

[0073] Extract features from underwater monitoring information, extract contour features and texture features of the target object, calculate similarity with the appearance comparison data set, and make a judgment with a preset threshold to analyze whether the detected object is an object to be fed, and obtain pre-identification information;

[0074] Marking the objects to be fed in combination with the pre-identification information and the underwater monitoring information, and performing quantity statistics, comparing the quantity statistics result with a preset threshold, and classifying the objects to be fed according to the quantity to obtain classification information;

[0075] Extracting the position information and underwater depth information of each category of the objects to be fed according to the underwater monitoring information, constructing a movement change trend graph according to the position change and depth change per unit time, analyzing the movement direction of the objects to be fed, and obtaining movement direction prediction information;

[0076] The underwater environment perception information is formed by combining the pre-identification information, category classification information and movement direction prediction information.

[0077] It should be noted that, first, underwater monitoring information is obtained through underwater detection equipment such as sonar or radar, and the appearance feature information of the object to be fed is obtained based on big data retrieval technology to form an appearance comparison data set. Next, feature extraction is performed on the underwater monitoring information. The contour features and texture features of the target detection object are extracted to map the appearance features of the detection object. These features are similarity calculated with the appearance comparison data set, and the detection object is judged whether it is an object to be fed by comparison with the preset threshold, so as to obtain pre-identification information, which reflects whether there is an object to be fed in the monitoring area, so as to distinguish whether the monitoring area needs to be fed. Subsequently, the objects to be fed are marked and counted in quantity in combination with the pre-identification information and the underwater monitoring information. The quantity statistics result is judged with the preset threshold, and the objects to be fed are classified according to the quantity, and the classification information is obtained. By counting the number of objects to be fed monitored in the monitoring area, the quantity is used to judge whether the area meets the feeding standard, so as to avoid blind feeding and waste of resources. Further, the location information and underwater depth information of each category of objects to be fed are extracted according to the underwater monitoring information. By constructing a movement change trend graph based on the position change and depth change per unit time, the moving direction of the object to be fed is analyzed, thereby obtaining the movement direction prediction information, which reflects whether the monitored object to be fed is moving, providing a basis for subsequent path planning. Finally, the underwater environment perception information is formed by combining the pre-identification information, category classification information and movement direction prediction information. This comprehensive information provides comprehensive information about the objects to be fed in the underwater environment for subsequent feeding path evaluation, including their type, quantity and movement trend, and provides key data for the system's intelligent perception and path planning.

[0078] S106, obtaining feeding area information, formulating a feeding path in combination with the underwater environment perception information and the surface environment perception information, and obtaining candidate feeding path information;

[0079] Acquire feeding area information, divide the area to be fed into a plurality of sub-areas according to the feeding area information to obtain area division information, and set the feeding area of ​​the feeding boat according to the area division information to obtain the feeding area information;

[0080] Acquire water surface environment perception information, combine it with feeding area information to screen the navigable area, and obtain navigable area information;

[0081] Acquire underwater environment perception information, mark the feeding areas in combination with the feeding area information, determine the feeding level according to the number of objects to be fed in each area, and obtain feeding level analysis information;

[0082] Extracting the position change information of the object to be fed according to the underwater environment perception information, setting the initial point and the monitoring point, calculating the moving speed of the object to be fed, and obtaining the moving speed information;

[0083] Constructing a feeding path formulation model, inputting the surface environment perception information, underwater environment perception information, navigable area information and feeding level analysis information into the feeding path formulation model to formulate a path, and obtaining initial feeding path information;

[0084] According to the moving speed information and the initial feeding path information, the moving position of the object to be fed arriving at the feeding area on each initial path is predicted to obtain moving position prediction information;

[0085] Evaluate each initial feeding path according to the mobile position prediction information, determine whether the path fails due to a change in the position of the object to be fed, perform path screening, and obtain candidate feeding path information;

[0086] The input is input into the feeding path formulation model to formulate the path to obtain the initial feeding path information, and further includes:

[0087] According to the water surface environment perception information and the navigable area information, the obstruction area and the navigable area are marked to obtain the feeding area marking information;

[0088] The number of objects to be fed in the obstruction area is determined according to the feeding area marking information, the underwater environment perception information and the feeding level judgment information, and whether feeding is required is analyzed to obtain feeding judgment information;

[0089] A particle swarm optimization algorithm and an ant colony algorithm are introduced to formulate a feeding path, initial parameters of the particle swarm algorithm are set according to the feeding judgment information, the feeding level analysis information and the navigable area information, and the particle position and speed are randomly initialized to obtain an initial particle swarm;

[0090] Calculate the fitness value of each individual particle in each initial particle group, compare it with the preset threshold, select the optimal particle and population based on the judgment result, update the optimal solution of the individual and population, and perform initial global planning and iterative optimization;

[0091] Iterative control rules are preset, and the inertia weight and learning factor are dynamically adjusted according to the iterative control adjustment rules until the optimization result meets the termination conditions and the initial global planning information is obtained;

[0092] The initial pheromone of the ant colony is set according to the feeding judgment information, the feeding level analysis information and the navigable area information, the initial global planning information is used as the enhancement variable of the ant colony pheromone, the enhancement variable is allocated to the initial pheromone of the ant colony, and the initial optimization parameters are set;

[0093] Calculate the state transition probability of each path node, optimize the path through the state transition probability, and update the pheromone of each optimized path until the optimization result meets the iteration termination condition to obtain the initial feeding path information.

[0094] It should be noted that, first, the feeding area information is obtained, the area to be fed is divided into multiple sub-areas, and the specific feeding area of ​​the feeding ship is set to form a complete feeding area information. Next, the obtained feeding area information is used to screen the navigable area in combination with the surface environment perception information. The area without obstacles is defined as the navigable area, which effectively limits the driving range of the feeding ship and improves the safety of driving. At the same time, by obtaining the underwater environment perception information and combining the feeding area information, the feeding area is marked. The number of objects to be fed in each area is analyzed, the feeding level is judged, and the feeding level analysis information is finally formed, thereby reflecting the feeding degree of each area. Subsequently, based on the underwater environment perception information, the position change information of the object to be fed is extracted. By setting the initial point and the monitoring point, the moving rate of the object to be fed is calculated to form the moving rate information. Further, a feeding path formulation model is constructed, and the surface environment perception information, the underwater environment perception information, the navigable area information and the feeding level analysis information are input into the model. The path is formulated through the model to obtain the initial feeding path information. According to the moving speed information, combined with the initial feeding path information, the moving position of the object to be fed in each initial path to the feeding area is predicted to form the moving position prediction information. Based on the moving position prediction information, each initial feeding path is evaluated to determine whether the path is invalid due to the change in the position of the object to be fed. Through path screening, the candidate feeding path information is obtained.

[0095] It should be noted that in the analysis process of the feeding path formulation model, first of all, the feeding area marking information, underwater environment perception information and feeding level judgment information are used to judge the number of objects to be fed in the obstruction area, analyze whether feeding is needed, and obtain feeding judgment information. Then, the particle swarm optimization algorithm and ant colony algorithm are introduced to formulate the feeding path. In the first step, the particle swarm optimization algorithm is used to perform initial global planning of the feeding path, plan a preliminary feeding path, and thus search for a preliminary feasible feeding path; when performing preliminary planning through the particle swarm algorithm, the inertia weight and learning factor are adjusted by preset iterative control rules, and dynamic regulation is performed, because the values ​​of the inertia weight and learning factor affect the accuracy of the algorithm search. By adjusting them, the premature convergence of the algorithm can be avoided, and the global optimization ability of the particle swarm optimization algorithm can be improved to obtain more accurate output results. Then, the ant colony algorithm is used to globally optimize the feeding path. The initial global planning information is used as the enhanced variable of the ant colony pheromone to allocate the enhanced variables of the pheromone, enhance the ant colony algorithm's recognition of the path, improve the diversity of ants' path selection, and thus obtain an accurate feeding path.

[0096] S108, evaluating each candidate feeding path according to the candidate feeding path information, and selecting a final feeding path according to the evaluation result;

[0097] Acquire candidate feeding path information, perform feature extraction on the candidate feeding path information, extract path length features, path planning features, and feeding sequence features of each candidate feeding path, and obtain candidate feeding path feature information;

[0098] Calculate the feeding time and energy consumption of each candidate feeding path according to the candidate feeding path characteristic information, use it as a feeding efficiency judgment index, and compare it with a preset threshold to obtain feeding efficiency evaluation information;

[0099] Evaluate the rationality of the path planning of each candidate feeding path according to the candidate feeding path feature information, analyze the repetition rate of the driving area through the path planning features and the feeding sequence features, and obtain feasibility evaluation information;

[0100] Acquire underwater environment perception information, calculate feeding coverage according to the candidate feeding path information, and evaluate feeding effect in combination with the underwater environment perception information to obtain feeding effect evaluation information;

[0101] A weighted calculation is performed on each candidate feeding path according to the feeding efficiency evaluation information, the feasibility evaluation information and the feeding effect evaluation information, and a final feeding path is selected according to the weighted calculation result to obtain final feeding path information.

[0102] It should be noted that, first, the candidate feeding path information is obtained, and the feature extraction of these candidate paths is performed to obtain the candidate feeding path feature information. Including path length features, path planning features and feeding sequence features. Next, the feeding time and energy consumption of each path are calculated using the candidate feeding path feature information, which is used as a judgment indicator of feeding efficiency. Compared with the preset threshold, the evaluation information of the feeding efficiency is obtained, which is used to measure the feeding efficiency of each path. At the same time, the rationality of the path planning is evaluated according to the characteristic information of the candidate feeding path. By analyzing the path planning features and the feeding sequence features, the repetition rate of the driving area is investigated, and the feasibility evaluation information is obtained to evaluate the rationality of the path planning of each path, so as to avoid repeated passing through multiple areas during the feeding process, which leads to unreasonable routes and reduced feeding efficiency. Subsequently, the underwater environment perception information is obtained, and the underwater environment perception information can be used to know the number of objects to be fed in each sub-area in the feeding area. By calculating the area to be fed covered by each feeding path, the feeding effect of the path is judged, thereby judging whether the path can be adopted. Finally, weighted calculation is performed on each candidate feeding path according to the feeding efficiency evaluation information, feasibility evaluation information and feeding effect evaluation information. According to the result of weighted calculation, the final feeding path is selected to obtain the final feeding path information. The optimal feeding path is selected based on various factors to improve feeding efficiency and effect.

[0103] Figure 2 A flow chart of feeding path planning and adjustment provided in one embodiment of the present invention;

[0104] like Figure 2 As shown, the present invention provides a feeding path planning and adjustment flow chart, including:

[0105] S202, sensing the surface environment and underwater environment of the feeding area, sensing surface obstacles and underwater objects to be fed;

[0106] S204, formulating feeding paths, evaluating the feeding efficiency, rationality and feeding effect of each feeding path, and selecting the optimal feeding path;

[0107] S206, monitoring the arrival of the objects to be fed in the target feeding area, determining whether the objects to be fed have moved, and analyzing whether a route change is required;

[0108] S208, predicting the activity area of ​​the object to be fed according to the movement of the object to be fed;

[0109] S210, judging whether the area is a fed area according to the predicted activity area of ​​the object to be fed;

[0110] S212, if it is a fed area, extract the amount of food to be fed and the amount of food already fed in the area, evaluate whether repeated feeding is needed, and adjust the route;

[0111] S214: If the area is not fed, a route change plan is performed to drive to the area for feeding.

[0112] Furthermore, water environment monitoring information of the area where the feeding boat is traveling is obtained, and the water environment monitoring information is sorted and classified according to the collected different areas to obtain water environment monitoring information of each area; a number of water environment level judgment thresholds are preset, and the water environment monitoring information of each water area is judged with the water environment level judgment threshold, and the water environment of each area is graded to obtain water environment level information; a water environment assessment standard is preset, and the water environment of each area is assessed in combination with the water environment monitoring information of each area and the water environment level information to obtain water environment assessment information; according to the water environment assessment standard, the water environment of each area is assessed. The method comprises the following steps: analyzing the water pollution situation in the feeding area based on the water environment assessment information to obtain the water environment pollution analysis information; obtaining the suitable environment information of the objects to be fed, and conducting the living environment assessment on the objects to be fed in each area based on the water environment pollution analysis information to obtain the living environment assessment information; judging whether water quality regulation is needed based on the living environment assessment information, and analyzing the regulation priority of each area based on the water environment pollution analysis information to obtain the regulation priority information; if water quality regulation is needed, formulating the water environment regulation strategy based on the water environment monitoring information and the regulation priority information of each area to improve the living environment quality of the objects to be fed, thereby improving the economic benefits.

[0113] Figure 3 A feeding path evaluation system framework based on intelligent perception is provided in one embodiment of the present invention Figure 3 The system includes: a memory 31 and a processor 32. The memory 31 includes a feeding path evaluation method program based on intelligent perception. When the feeding path evaluation method program based on intelligent perception is executed by the processor 32, the following steps are implemented:

[0114] Acquire water surface monitoring information, perform water surface environment perception according to the water surface monitoring information, and obtain water surface environment perception information;

[0115] Acquiring underwater monitoring information, and performing underwater environment perception according to the underwater monitoring information to obtain underwater environment perception information;

[0116] Obtain feeding area information, formulate feeding paths based on underwater environment perception information and surface environment perception information, and obtain candidate feeding path information;

[0117] Each candidate feeding path is evaluated according to the candidate feeding path information, and a final feeding path is selected according to the evaluation result.

[0118] It should be noted that the present invention provides a feeding path evaluation method and system based on intelligent perception, which performs surface environment perception and underwater environment perception of the feeding area to understand the obstacles on the surface environment and the location of the objects to be fed in the underwater environment, thereby providing a planning basis for the feeding path. Through the underwater environment perception information and the surface environment perception information, the areas that do not need to be fed and the areas that need to be fed are determined, and a safe feeding path is formulated based on the identified surface obstacles. The formulated feeding path is evaluated from three aspects: feeding efficiency, rationality, and feeding effect, to ensure the effectiveness of the feeding path, while reducing resource consumption, feeding in a targeted manner, and improving feeding efficiency, avoiding time consumption due to repeated feeding.

[0119] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0120] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units; they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0121] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0122] A person of ordinary skill in the art can understand that: all or part of the steps of implementing the above method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0123] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention can be essentially or partly reflected in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0124] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A feeding path evaluation method based on intelligent perception, characterized in that: include: Acquiring water surface monitoring information, and performing water surface environment perception according to the water surface monitoring information to obtain water surface environment perception information; Acquiring underwater monitoring information, and performing underwater environment perception according to the underwater monitoring information to obtain underwater environment perception information; Obtain feeding area information, formulate feeding paths based on underwater environment perception information and surface environment perception information, and obtain candidate feeding path information; Evaluate each candidate feeding path according to the candidate feeding path information, and select a final feeding path according to the evaluation result; The acquiring of water surface monitoring information and performing water surface environment perception according to the water surface monitoring information specifically includes: Acquire water surface monitoring information, pre-process the water surface monitoring information, acquire characteristic information of various water surface obstacles and ships and characteristic information of the objects to be fed based on big data retrieval, and form a comparative data set; Constructing a water surface environment recognition model, and constructing a training data set according to the comparison data set to perform deep learning and training on the water surface environment recognition model, wherein the water surface environment recognition model includes a feature extraction layer, a target detection layer, and an environment recognition layer; Performing feature extraction on the water surface monitoring information according to the feature extraction layer, extracting color features, contour features and texture features, and constructing a feature map according to the scale of each feature to obtain water surface monitoring feature information; The water surface monitoring feature information is input into the target detection layer for target detection, an attention mechanism is introduced, and the attention scores of the feature maps of each scale are calculated by combining the water surface monitoring feature information with the attention mechanism, and an attention feature map is constructed to obtain attention feature map information; Based on YOLOv3, an FPN feature pyramid is constructed, the water surface monitoring feature information and the attention feature map information are fused, the target position is predicted using Yolo Head, a target detection frame is generated, and target detection information is obtained; Performing image segmentation on the detection target according to the target detection information, and performing feature extraction on the target object image after image segmentation to obtain feature information of the target object; Inputting the target object feature information into the environment recognition layer for environment recognition, identifying obstacles in the navigation direction of the feeding ship, and obtaining water surface environment perception information; The acquiring of underwater monitoring information and performing underwater environment perception according to the underwater monitoring information specifically includes: Obtain underwater monitoring information, obtain the appearance feature information of the object to be fed based on big data retrieval, and form an appearance comparison data set; Extract features from underwater monitoring information, extract contour features and texture features of the target object, calculate similarity with the appearance comparison data set, and make a judgment with a preset threshold to analyze whether the detected object is an object to be fed, and obtain pre-identification information; Marking the objects to be fed in combination with the pre-identification information and the underwater monitoring information, and performing quantity statistics, comparing the quantity statistics result with a preset threshold, and classifying the objects to be fed according to the quantity to obtain classification information; Extracting the position information and underwater depth information of each category of the objects to be fed according to the underwater monitoring information, constructing a movement change trend graph according to the position change and depth change per unit time, analyzing the movement direction of the objects to be fed, and obtaining movement direction prediction information; Combine pre-identification information, classification information and movement direction prediction information to form underwater environment perception information; The step of obtaining the feeding area information and formulating the feeding path in combination with the underwater environment perception information and the surface environment perception information to obtain the candidate feeding path information specifically includes: Acquire feeding area information, divide the area to be fed into a plurality of sub-areas according to the feeding area information to obtain area division information, and set the feeding area of ​​the feeding boat according to the area division information to obtain the feeding area information; Acquire water surface environment perception information, combine it with feeding area information to screen the navigable area, and obtain navigable area information; Acquire underwater environment perception information, mark the feeding areas in combination with the feeding area information, determine the feeding level according to the number of objects to be fed in each area, and obtain feeding level analysis information; Extracting the position change information of the object to be fed according to the underwater environment perception information, setting the initial point and the monitoring point, calculating the moving speed of the object to be fed, and obtaining the moving speed information; Constructing a feeding path formulation model, inputting the surface environment perception information, underwater environment perception information, navigable area information and feeding level analysis information into the feeding path formulation model to formulate a path, and obtaining initial feeding path information; According to the moving speed information and the initial feeding path information, the moving position of the object to be fed arriving at the feeding area on each initial path is predicted to obtain moving position prediction information; Each initial feeding path is evaluated according to the mobile position prediction information to determine whether the path is invalid due to the change of the position of the object to be fed, and the path is screened to obtain candidate feeding path information.

2. A feeding path evaluation method based on intelligent perception according to claim 1, characterized in that: The input is input into the feeding path formulation model to formulate the path to obtain the initial feeding path information, and further includes: According to the water surface environment perception information and the navigable area information, the obstruction area and the navigable area are marked to obtain the feeding area marking information; The number of objects to be fed in the obstruction area is determined according to the feeding area marking information, the underwater environment perception information and the feeding level judgment information, and whether feeding is required is analyzed to obtain feeding judgment information; A particle swarm optimization algorithm and an ant colony algorithm are introduced to formulate a feeding path, initial parameters of the particle swarm algorithm are set according to the feeding judgment information, the feeding level analysis information and the navigable area information, and the particle position and speed are randomly initialized to obtain an initial particle swarm; Calculate the fitness value of each individual particle in each initial particle group, compare it with the preset threshold, select the optimal particle and population based on the judgment result, update the optimal solution of the individual and population, and perform initial global planning and iterative optimization; Iterative control rules are preset, and the inertia weight and learning factor are dynamically adjusted according to the iterative control adjustment rules until the optimization result meets the termination conditions and the initial global planning information is obtained; The initial pheromone of the ant colony is set according to the feeding judgment information, the feeding level analysis information and the navigable area information, the initial global planning information is used as the enhancement variable of the ant colony pheromone, the enhancement variable is allocated to the initial pheromone of the ant colony, and the initial optimization parameters are set; Calculate the state transition probability of each path node, optimize the path through the state transition probability, and update the pheromone of each optimized path until the optimization result meets the iteration termination condition to obtain the initial feeding path information.

3. A feeding path evaluation method based on intelligent perception according to claim 1, characterized in that: The step of evaluating each candidate feeding path according to the candidate feeding path information and selecting a final feeding path according to the evaluation result specifically includes: Acquire candidate feeding path information, perform feature extraction on the candidate feeding path information, extract path length features, path planning features, and feeding sequence features of each candidate feeding path, and obtain candidate feeding path feature information; Calculate the feeding time and energy consumption of each candidate feeding path according to the candidate feeding path characteristic information, use it as a feeding efficiency judgment index, and compare it with a preset threshold to obtain feeding efficiency evaluation information; Evaluate the rationality of the path planning of each candidate feeding path according to the candidate feeding path feature information, analyze the repetition rate of the driving area through the path planning features and the feeding sequence features, and obtain feasibility evaluation information; Acquire underwater environment perception information, calculate feeding coverage according to the candidate feeding path information, and evaluate feeding effect in combination with the underwater environment perception information to obtain feeding effect evaluation information; A weighted calculation is performed on each candidate feeding path according to the feeding efficiency evaluation information, the feasibility evaluation information and the feeding effect evaluation information, and a final feeding path is selected according to the weighted calculation result to obtain final feeding path information.

4. A feeding path evaluation system based on intelligent perception, characterized in that: The system comprises: a memory and a processor, wherein the memory comprises a feeding path evaluation method program based on intelligent perception, and when the feeding path evaluation method program based on intelligent perception is executed by the processor, the feeding path evaluation method steps based on intelligent perception as described in any one of claims 1 to 3 are implemented.

Images