Accurate bait casting control system and method for crab breeding

Abstract

The invention discloses an accurate bait casting control system and method for crab breeding. The accurate bait casting control system comprises a main control system, a workboat, a bait casting device, an image module and a positioning module; the main control system is connected with the workboat and the bait casting device, controls the operation and moving path of the workboat, and controls the bait casting operation and bait casting amount of the bait casting device; the image module comprises an image acquisition module and a recognition module, the image acquisition module is installed on the workboat and used for acquiring underwater images, and the recognition module dynamically recognizes the density of the blue crabs according to the images. Through the main control system, the workboat, the bait casting device, the image module and the positioning module, bait can be accurately controlled and saved, and the yield of the blue crabs is increased; the breeding scale is increased, the breeding benefits are improved, and the purpose of scientifically breeding the blue crabs is expected to be achieved.

Description

Technical Field

[0001] This invention belongs to the field of mud crab farming and relates to a precise feeding control system and method for crab farming. Background Technology

[0002] Traditional feeding methods, involving manual rowing by farmers, result in low feed utilization and high labor intensity, making them unsuitable for large-scale mud crab farming. Furthermore, mud crabs are highly territorial and often fight with their own kind for food. Insufficient feed hinders their growth and development, and in severe cases, can lead to fighting and even cannibalism; excessive feed not only increases costs but also causes eutrophication due to uneaten feed. Therefore, current mud crab farming management practices are outdated, lacking mechanization and automation, and their extensive operations and low feed utilization negatively impact farming efficiency. Summary of the Invention

[0003] In order to overcome at least one deficiency of the prior art, the present invention provides a precision feeding control system and method for crab farming.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a precision feeding control system for crab farming, comprising a main control system, a workboat, a feeding device, an image module, and a positioning module; The main control system is connected to the work vessel and the feeding device respectively, controlling the operation and movement path of the work vessel, and controlling the feeding operation and feeding amount of the feeding device. The image module includes an image acquisition module and a recognition module. The image acquisition module is installed on the work vessel to acquire underwater images, and the recognition module dynamically identifies the density of mud crabs based on the images. The positioning module is used to collect the location of underwater images. Combining the underwater images with their location makes it easier for the identification module to dynamically identify the density of blue crabs. The main control system connects to a remote server. The image module and positioning module upload data to the main control system, which in turn uploads data to the remote server. The remote server then connects to the terminal device.

[0005] A method for precise feeding control in crab farming includes the following steps: Step 1: Divide the pond into aquaculture areas and mark them; Step 2: Dynamically identify the density of blue crabs in different zones; Step 3: Using the crab pond water temperature, pH value, and dissolved oxygen content as inputs to the GA-BP neural network, environmental impact factors are obtained after training. Step 4: Establish a mud crab growth model. The mud crab growth model combines the breeding density and breeding area of ​​each area to obtain the total mass of mud crabs in each area, and obtains the empirical feeding amount for each area at different growth stages based on the feeding rate of mud crabs at different growth stages. Step 5: Based on environmental impact factors, empirical feeding amount and crab survival rate, determine the total feeding amount for each area, generate a feeding formula table, and send the feeding formula table to the work vessel. Step 6: According to the feeding formula, the main control system controls the work vessel and feeding device to complete the automatic feeding in each area.

[0006] Furthermore, the method for dynamically identifying the density of blue crabs in different zones in step 2 is as follows: Step 21: Acquire underwater images of each zone and preprocess the images based on image denoising algorithms and image color correction and enhancement algorithms; Step 22: Construct a live crab detection model; Step 23: The live crab detection model introduces a composite scaling factor to globally extend the efficiently fused feature network and category / boundary box prediction network to construct a live crab detector; Step 24: The live crab detector automatically identifies mud crabs and integrates location data to statistically analyze their distribution information, generating a mud crab feeding density prescription map.

[0007] Furthermore, the GA-BP neural network algorithm flow is as follows: Step 31: Set the initial BP weights and threshold values, and encode the initial values ​​using GA; Step 32: Input the crab pond water temperature, pH value, and dissolved oxygen content data, and preprocess the data; Step 33: The error obtained from training the BP neural network is used as the fitness. Step 34: After selection, crossover, and mutation, calculate the fitness and determine whether the fitness has reached the set value. If yes, obtain the optimal weight and threshold; otherwise, recalculate the fitness. Step 35: The error is obtained by training the BP neural network again, the weights and thresholds are updated, and it is determined whether the set conditions are met. If yes, the environmental factors are output; otherwise, the error is recalculated.

[0008] Furthermore, the steps for establishing the blue crab growth model are as follows: Step 41: Regularly harvest mud crabs, weigh them, and measure the water quality to obtain the individual weight of the mud crabs and the parameters of the breeding environment; Step 42: Extract, transform, and analyze the data, perform grey relational analysis on the data, and establish a growth model for mud crabs; Step 43: Evaluate the growth model of mud crabs.

[0009] Furthermore, the grey relational analysis involves identifying and representing explicit and implicit relationships from various types of information based on the collected data. The process is as follows: Step 421: Selecting reference and comparison sequences: Select the weight gain change of mud crabs as the reference sequence, and select the normalized aquaculture environment parameters as the comparison sequence; The normalization method is as follows:

[0010] In the formula, , These represent the values ​​before and after data normalization, respectively. To be the minimum value, It is the maximum value; Step 422: Calculate the grey relational coefficient ; ; In the formula, As a reference sequence, For comparing sequences, The resolution coefficient is adjustable. Step 423: Calculate the mean correlation degree ;

[0011] in, The mean of the correlation. The number of samples; Step 424: Sort and arrange the data according to the mean correlation degree to form a correlation sequence; use grey relational analysis to identify the environmental factors with the highest correlation to the growth of mud crabs. var ; Step 425: Establish the blue crab growth model G:

[0012] in, m This refers to the maximum weight limit parameter for mud crabs. k Instantaneous growth rate t This refers to the current number of days the mud crabs have been raised. p For calibration parameters, c The constant coefficient, var As environmental factors, the above model is iteratively fitted with nonlinear parameters to obtain... m , p , c , k The optimal solution.

[0013] Furthermore, the backbone network of the live crab detector is based on EfficientNet with linkage extension. This backbone network constructs a live crab detection model that limits the target computational load and storage space from three aspects: the manually designed MBConv basic computing module, the network structure searched by the neural network architecture, and the deep pruning and compression with accompanying computational constraints.

[0014] Furthermore, the aquaculture environment parameters include temperature, dissolved oxygen content, and pH value.

[0015] Furthermore, the formula for the total mass Z of the blue crab partition is: ; in, Z The total mass of mud crabs in the zone. D To determine the stocking density for different zones, S The area represents the zone.

[0016] Furthermore, the formula for calculating the feeding amount F in each zone is as follows:

[0017] in: F Feeding amount per zone The feeding rate is closely related to the various growth stages of mud crabs. Based on feeding experience and historical data, it is set as [missing value]. ; The formula for the total amount of feed H in each zone is:

[0018] in: H This represents the total amount of feed given to the mud crabs in each zone. To improve the survival rate of mud crabs, R These are environmental impact factors calculated using the GA-BP neural network.

[0019] In summary, the advantages of this invention are: This invention, through a main control system, a work vessel, a feeding device, an image module, and a positioning module, helps to precisely control the feed, save feed, increase the yield of mud crabs, expand the scale of aquaculture, and improve aquaculture efficiency, in order to achieve the goal of scientific mud crab farming. Attached Figure Description

[0020] Figure 1 This is a flowchart of the GA-BP neural network algorithm of the present invention.

[0021] Figure 2 This is a flowchart of the feeding decision-making process of the present invention.

[0022] Figure 3 This is a diagram illustrating the construction of the blue crab growth model of the present invention.

[0023] Figure 4 This is a flowchart of the grayscale correlation analysis of the present invention. Detailed Implementation

[0024] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that, unless otherwise specified, the following embodiments and features described therein can be combined with each other.

[0025] Example: A precision feeding control system for crab farming includes a main control system, a work vessel, a feeding device, an image module, and a positioning module; The main control system is connected to the work vessel and the feeding device respectively, controlling the operation and movement path of the work vessel, and controlling the feeding operation and feeding amount of the feeding device. Through the remote control of the main control system, unmanned fully automatic feeding is realized.

[0026] The image module includes an image acquisition module and a recognition module. The image acquisition module is installed on the work vessel to acquire underwater images, and the recognition module dynamically identifies the density of mud crabs based on the images.

[0027] The image acquisition module can be an underwater camera.

[0028] Before feeding the pond, the work vessel collects information on the location of the pond boundary and identifies the density of mud crabs. The main control system controls the work vessel and the feeding device to automatically and evenly feed the pond.

[0029] The positioning module is used to collect the location of underwater images. Combining the underwater images with their location makes it easier for the recognition module to dynamically identify the density of mud crabs.

[0030] The main control system can connect to a remote server. The image and positioning modules upload data to the main control system, which in turn uploads data to the remote server. The remote server then connects to the terminal device. This allows relevant parameters to be viewed in real time on the mobile terminal. Simultaneously, these parameters also serve as one of the bases for real-time early warning and input parameters for the blue crab growth model.

[0031] Using a handheld terminal, the operation parameters of the feeding boat can be adjusted in real time for different zones of the crab pond. Before feeding, the feeding boat collects information about the pond's boundaries. A remote server intelligently plans the path based on this boundary information, generates a map of the operation trajectory, and sends it to the main control system. The main control system then controls the feeding boat and the feeding device to automatically feed the crabs in the pond. Simultaneously, the handheld terminal allows for real-time monitoring of the feeding boat's course, speed, location, remaining feed, and equipment status.

[0032] like Figures 1-4As shown, a method for precise feeding control in crab farming includes the following steps: Step 1: Divide the pond into aquaculture areas and mark them; Step 2: Dynamically identify the density of blue crabs in different zones; Step 3: Using the crab pond water temperature, pH value, and dissolved oxygen content as inputs to the GA-BP neural network, environmental impact factors are obtained after training. Step 4: Establish a mud crab growth model. The mud crab growth model combines the breeding density and breeding area of ​​each area to obtain the total mass of mud crabs, and obtain the empirical feeding amount of each area at different growth stages based on the feeding rate of mud crabs at different growth stages. Step 5: Based on environmental impact factors, empirical feeding amount and crab survival rate, determine the total feeding amount for each area, generate a feeding formula table, and send the feeding formula table to the work vessel. Step 6: According to the feeding formula, the main control system controls the work vessel and feeding device to complete the automatic feeding in each area.

[0033] The method for dynamically identifying the density of blue crabs in different zones in step 2 is as follows: Step 21: Acquire underwater images of each zone and preprocess the images based on image denoising algorithms and image color correction and enhancement algorithms; Image denoising algorithms based on sparse representation and image color correction and enhancement algorithms based on color theory can completely abstract the image formation process without prior environmental knowledge, ensuring the reliability of image information and being simpler and more effective than image restoration techniques.

[0034] Step 22: Construct a live crab detection model; The backbone network of the live crab detector is EfficientNet based on linkage extension. This backbone network constructs a live crab detection model that can limit the target computational volume and storage space from three aspects: the manually designed MBConv basic computing module, the network structure of neural network architecture search, and the deep pruning and compression with accompanying computational constraints. Step 23: The live crab detection model introduces a composite scaling factor to globally extend the efficiently fused feature network and category / boundary box prediction network to construct a live crab detector.

[0035] Step 24: The live crab detector automatically identifies mud crabs and integrates location data to statistically analyze their distribution information, generating a mud crab feeding density prescription map.

[0036] The specific process of the GA-BP neural network algorithm is as follows: Step 31: Set the initial BP weights and threshold values, and encode the initial values ​​using GA; Step 32: Input the crab pond water temperature, pH value, and dissolved oxygen content data, and preprocess the data; Step 33: The error obtained from training the BP neural network is used as the fitness. Step 34: After selection, crossover, and mutation, calculate the fitness and determine whether the fitness has reached the set value. If yes, obtain the optimal weight and threshold; otherwise, recalculate the fitness. Step 35: The error is obtained by training the BP neural network again, the weights and thresholds are updated, and it is determined whether the set conditions are met. If yes, the environmental factors are output; otherwise, the error is recalculated.

[0037] The steps to establish a growth model for mud crabs are as follows: Step 41: Regularly harvest mud crabs, weigh them, and measure the water quality to obtain the individual weight of the mud crabs and the parameters of the breeding environment; Aquaculture environmental parameters include temperature, dissolved oxygen content, pH value, etc. Step 42: Extract, transform, and analyze the data, perform grey relational analysis on the data, and establish a growth model for mud crabs; Step 43: Evaluate the growth model of mud crabs.

[0038] Grey relational analysis is based on collected data to discover and represent explicit and implicit relationships from various types of information. The specific process is as follows: Step 421: Selecting reference and comparison sequences: Select the weight gain change of mud crabs as the reference sequence, and select the normalized aquaculture environment parameters as the comparison sequence; Because aquaculture environmental parameters have different physical meanings, their dimensions also differ. To reduce the discrepancies in absolute data values ​​and ensure the reliability of the results, the data is normalized and standardized to an approximate range.

[0039] The normalization method is as follows:

[0040] In the formula, , These represent the values ​​before and after data normalization, respectively. To be the minimum value, This is the maximum value.

[0041] Step 422: Calculate the grey relational coefficient ; ; In the formula, As a reference sequence, For comparing sequences, This is an adjustable resolution coefficient, with a value of (0, 1), typically set to 0.5. Step 423: Calculate the mean correlation degree ;

[0042] in, The mean of the correlation. The number of samples; Step 424: Sort and arrange according to the mean correlation degree to form a correlation sequence; the environmental factors with the greatest correlation to the growth of mud crabs can be obtained through grey relational analysis. var ; The correlation sequence can reveal the extent to which aquaculture environment parameters affect the reference sequence, preparing for further optimization of the growth model; Step 425: Establish the blue crab growth model G:

[0043] in, m This refers to the maximum weight limit parameter for mud crabs. k Instantaneous growth rate t This refers to the current number of days the mud crabs have been raised. p For calibration parameters, c The constant coefficient, var Environmental factors are considered. Through nonlinear fitting of the above model, iterative results are obtained. m , p , c , k The optimal solution.

[0044] The formula for the total mass Z of the blue crab partition is: ; in, Z The total mass of mud crabs in the zone. D To determine the stocking density for different zones, S The area represents the zone.

[0045] The formula for calculating the feeding amount F in each zone is: ; in: F Feeding amount per zone The feeding rate is given. The feeding rate is closely related to each stage of the mud crab's growth. Based on feeding experience and historical data, it is set as [value missing]. .

[0046] The formula for the total amount of feed H in each zone is:

[0047] in: H This represents the total amount of feed given to the mud crabs in each zone. To improve the survival rate of mud crabs, R These are environmental impact factors calculated using the GA-BP neural network.

[0048] During the feeding process, the work vessel also monitors the water environment in the pond. The relevant data serves as an environmental factor for the growth model of mud crabs, which is used to optimize related algorithms. On the other hand, users can use mobile terminal devices such as smartphones and IoT-related technologies to achieve comprehensive monitoring of water quality parameters in the aquaculture area.

[0049] Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort should fall within the scope of protection of the present invention.

Claims

1. A precision feeding control system for crab farming, characterized in that: Includes main control system, workboat, feeding device, image module, and positioning module; The main control system is connected to the work vessel and the feeding device respectively, controlling the operation and movement path of the work vessel, and controlling the feeding operation and feeding amount of the feeding device. The image module includes an image acquisition module and a recognition module. The image acquisition module is installed on the work vessel to acquire underwater images, and the recognition module dynamically identifies the density of mud crabs based on the images. The positioning module is used to collect the location of underwater images. Combining the underwater images with their location makes it easier for the identification module to dynamically identify the density of blue crabs. The main control system connects to a remote server. The image module and positioning module upload data to the main control system, which in turn uploads data to the remote server. The remote server then connects to the terminal device.

2. A method for precise feeding control in crab farming, characterized in that: Includes the following steps: Step 1: Divide the pond into aquaculture areas and mark them; Step 2: Dynamically identify the density of blue crabs in different zones; Step 3: Using the crab pond water temperature, pH value, and dissolved oxygen content as inputs to the GA-BP neural network, environmental impact factors are obtained after training. Step 4: Establish a mud crab growth model. The mud crab growth model combines the breeding density and breeding area of ​​each area to obtain the total mass of mud crabs in each area, and obtains the empirical feeding amount for each area at different growth stages based on the feeding rate of mud crabs at different growth stages. Step 5: Based on environmental impact factors, empirical feeding amount and crab survival rate, determine the total feeding amount for each area, generate a feeding formula table, and send the feeding formula table to the work vessel. Step 6: According to the feeding formula, the main control system controls the work vessel and feeding device to complete the automatic feeding in each area.

3. The method for precise feeding control in crab farming according to claim 2, characterized in that: The method for dynamically identifying the density of blue crabs in different zones in step 2 is as follows: Step 21: Acquire underwater images of each zone and preprocess the images based on image denoising algorithms and image color correction and enhancement algorithms; Step 22: Construct a live crab detection model; Step 23: The live crab detection model introduces a composite scaling factor to globally extend the efficiently fused feature network and category / boundary box prediction network to construct a live crab detector; Step 24: The live crab detector automatically identifies mud crabs and integrates location data to statistically analyze their distribution information, generating a mud crab feeding density prescription map.

4. The method for precise feeding control in crab farming according to claim 2, characterized in that: The GA-BP neural network algorithm flow is as follows: Step 31: Set the initial BP weights and threshold values, and encode the initial values ​​using GA; Step 32: Input the crab pond water temperature, pH value, and dissolved oxygen content data, and preprocess the data; Step 33: The error obtained from training the BP neural network is used as the fitness. Step 34: After selection, crossover, and mutation, calculate the fitness and determine whether the fitness has reached the set value. If yes, obtain the optimal weight and threshold; otherwise, recalculate the fitness. Step 35: The error is obtained by training the BP neural network again, the weights and thresholds are updated, and it is determined whether the set conditions are met. If yes, the environmental factors are output; otherwise, the error is recalculated.

5. The method for precise feeding control in crab farming according to claim 2, characterized in that: The steps for establishing the growth model of the mud crab are as follows: Step 41: Regularly harvest mud crabs, weigh them, and measure the water quality to obtain the individual weight of the mud crabs and the parameters of the breeding environment; Step 42: Extract, transform, and analyze the data, perform grey relational analysis on the data, and establish a growth model for mud crabs; Step 43: Evaluate the growth model of mud crabs.

6. The method for precise feeding control in crab farming according to claim 5, characterized in that: The grey relational analysis involves identifying and representing explicit and implicit relationships from various types of information based on collected data. The process is as follows: Step 421: Selecting reference and comparison sequences: Select the weight gain change of mud crabs as the reference sequence, and select the normalized aquaculture environment parameters as the comparison sequence; The normalization method is as follows: ； In the formula, , These represent the values ​​before and after data normalization, respectively. To be the minimum value, It is the maximum value; Step 422: Calculate the grey relational coefficient ; ； In the formula, As a reference sequence, For comparing sequences, The resolution coefficient is adjustable. Step 423: Calculate the mean correlation degree ; ； in, The mean of the correlation. The number of samples; Step 424: Sort and arrange the data according to the mean correlation degree to form a correlation sequence; use grey relational analysis to identify the environmental factors with the strongest correlation to the growth of mud crabs. var ; Step 425: Establish the blue crab growth model G: ； in, m This refers to the maximum weight limit parameter for mud crabs. k Instantaneous growth rate t This refers to the current number of days the mud crabs have been raised. p For calibration parameters, c The constant coefficient, var As environmental factors, the above model is iteratively fitted with nonlinear parameters to obtain... m , p , c , k The optimal solution.

7. The method for precise feeding control in crab farming according to claim 3, characterized in that: The backbone network of the live crab detector is based on EfficientNet with linkage extension. This backbone network constructs a live crab detection model that limits the target computational load and storage space from three aspects: the manually designed MBConv basic computing module, the network structure searched by the neural network architecture, and the deep pruning and compression with accompanying computational constraints.

8. The method for precise feeding control in crab farming according to claim 5, characterized in that: The aquaculture environment parameters include temperature, dissolved oxygen content, and pH value.

9. The method for precise feeding control in crab farming according to claim 2, characterized in that: The formula for the total mass Z of the blue crab partition is: ； in, Z The total mass of mud crabs in the zone. D To determine the stocking density for different zones, S The area represents the zone.

10. The method for precise feeding control in crab farming according to claim 2, characterized in that: The formula for calculating the feeding amount F in the designated area is: ； in: F Feeding amount per zone The feeding rate is closely related to the various growth stages of mud crabs. Based on feeding experience and historical data, it is set as [missing value]. ; The formula for the total amount of feed H in each zone is: ； in: H This represents the total amount of feed given to the mud crabs in each zone. To improve the survival rate of mud crabs, R These are the environmental impact factors calculated using the GA-BP neural network.

Images