Digital breeding method and system for shrimp culture
By building a digital breeding model and visual processing system, the problem of expanding the scale of breeding groups and genetic progress in shrimp breeding was solved, and the refined management and quality improvement of shrimp breeding was achieved.
Patent Information
- Application Number
- CN202510539334.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-27
- Publication Date
- 2025-08-01
AI Technical Summary
In shrimp farming, it is difficult for the existing technology to effectively expand the scale of breeding groups and improve genetic progress, and there are risks and challenges brought about by the introduction of exogenous germplasm.
The shrimp breeding method based on digital breeding model is adopted, and the shrimp population data analysis and trait observation are carried out through the construction of a neural network model, combined with a visual processing system, and the breeding data is optimized to achieve refined management and quality improvement.
It has achieved refined management and quality improvement of shrimp farming process, and can monitor and adjust breeding strategies in real time to improve shrimp quality.
Smart Images
Figure CN120409941A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of shrimp farming, and particularly to a digital breeding selection method and system for shrimp farming. Background Art
[0002] In recent years, for important farmed shrimps in China such as whiteleg shrimp, Chinese shrimp, and giant river prawn, after years of artificial breeding and farming, the degradation of germplasm resources has become increasingly prominent. The selection breeding technology system based on large-scale families has become the main means to improve the important economic traits of most shrimps. The main characteristics of this technology system are as follows: a certain scale of families is established in each generation. After physical marking of a certain number of individuals in each family, they are mixed and cultured for trait testing. Pedigree and trait test data are used to evaluate and screen excellent families and individuals, and then optimized breeding is carried out to produce the next-generation families. In the large-scale family selection breeding system, the genetic progress mainly depends on the heritability and selection intensity of the target traits. Since most of the core breeding populations have been closed and selected for multiple generations and have excellent performance in important economic traits, directly introducing wild populations without genetic improvement to increase the genetic variation of the breeding population will reduce the performance of the breeding target traits of the core breeding population. In addition, introducing exogenous improved germplasm populations also has uncontrollable risks such as introducing unknown pathogenic agents. Therefore, it is difficult to carry out the work of introducing exogenous germplasm in actual breeding. To sum up, when the construction of the basic population has been completed and parameters such as heritability have been fixed, expanding the scale of the selected breeding population and increasing the selection intensity are one of the main ways to further improve the genetic progress.
[0003] To solve these problems, there is an urgent need for a digital breeding selection method and system for shrimp farming. Summary of the Invention
[0004] To solve the above problems, the present application proposes a shrimp farming selection method based on a digital breeding selection model, including the following steps:
[0005] S1. Obtain past breeding selection information, analyze the past breeding selection information to obtain past breeding selection data and past breeding selection results;
[0006] S2. Construct a digital breeding selection model, use the past breeding selection data as input and the past breeding selection results as output, and train the digital breeding selection model;
[0007] S3. Obtain current shrimp population data and target breeding selection results, and bring the current shrimp population data and target breeding selection results into the digital breeding selection model for training to obtain target breeding selection data;
[0008] S4. Carry out farming treatment on the current shrimp population according to the target breeding selection data, and use a vision processing system to conduct stage trait observation on the current shrimp population to obtain stage trait data;
[0009] S5. Input the stage trait data into the digital breeding model to adjust the target breeding data and obtain the optimized breeding data;
[0010] S6. Breed the shrimp population according to the optimized breeding data to obtain the target shrimp body.
[0011] Preferably, the digital breeding model in S2 is based on a neural network model and includes:
[0012] Head training layer: The head training layer is used to receive the initial information, and complete the overall training in combination with the process training layer and the tail training layer to generate the target breeding data;
[0013] Process training layer: The process training includes completing the overall training in combination with the head training layer and the tail training layer or receiving stage information, and performing partial training in combination with the tail training layer to obtain the optimized breeding data;
[0014] Tail training layer; The tail training layer is used to cooperate with the head training layer and the process training layer to complete the overall training and partial training and output the past breeding results.
[0015] Preferably, the specific content of receiving the stage information and performing partial training in combination with the tail training layer to obtain the optimized breeding data is:
[0016] The process training layer includes several layers, arranged in sequence;
[0017] Different process training layers correspond to different shrimp growth stages. Input the stage information into the corresponding process training layer, and the corresponding process training layer and the process training layers arranged after it combine with the tail training layer to perform different partial trainings.
[0018] Preferably, during the training process of the digital breeding model, the overall training is the main body, and the overall training is carried out first and then the partial training;
[0019] The overall training and the partial training are respectively set with a main iteration number and a partial iteration number, and the main iteration number is less than the partial iteration number of the tail training layer.
[0020] Preferably, the specific content of S3, obtaining the current shrimp population data and the target breeding result, and bringing the current shrimp population data and the target breeding result into the digital breeding model for training to obtain the target breeding data is;
[0021] The current shrimp population data includes the shrimp population environment, the shrimp population gene pedigree, and the shrimp population trait data;
[0022] The target breeding result is the constraint condition of the digital breeding model;
[0023] If the breeding result does not meet the constraint conditions when the number of iterations reaches the main iteration number, continue the iteration until the breeding result meets the constraint conditions and output the target breeding data;
[0024] If the breeding result meets the constraint conditions when the number of iterations does not reach the main iteration number, directly output the target breeding data.
[0025] Preferably, in S4, the current shrimp population is cultured according to the target breeding data, and the specific content of obtaining the stage trait data by using the vision processing system to conduct stage-by-stage observation on the current shrimp population is as follows:
[0026] Set the installation positions of the aquaculture environment equipment according to the activities of the shrimp population, and install the image acquisition equipment at the installation positions to obtain the acquired images;
[0027] Based on the deep learning algorithm, analyze and process the acquired images to obtain the stage trait data;
[0028] The stage trait data includes growth rate, disease resistance, and individual size;
[0029] The stage-by-stage observation is the observation of the growth stage of shrimp.
[0030] Preferably, in S5, input the stage trait data into the digital breeding model to adjust the target breeding data to obtain the optimized breeding data:
[0031] Take the stage trait data as the stage information and input it into the corresponding process training layer, and conduct different parts of training to obtain the optimized breeding data.
[0032] A digital breeding system for shrimp farming includes:
[0033] Data acquisition unit: Obtain the past breeding information, analyze the past breeding information to obtain the past breeding data and past breeding results;
[0034] Model construction unit: Construct a digital breeding model, use the past breeding data as the input and the past breeding results as the output, and train the digital breeding model;
[0035] Breeding decision unit: The breeding decision unit includes an overall decision module and a partial decision module;
[0036] The overall decision module includes obtaining the current shrimp population data and the target breeding result, and bringing the current shrimp population data and the target breeding result into the digital breeding model for training to obtain the target breeding data;
[0037] The partial decision-making module includes performing aquaculture processing on the current shrimp population according to the target breeding data, using a vision processing system to conduct phased observation of the traits of the current shrimp population to obtain phased trait data, and inputting the phased trait data into a digital breeding model to adjust the target breeding data to obtain optimized breeding data.
[0038] An electronic device includes a memory and a processor. A computer program is stored in the memory. When the processor calls the computer program in the memory, the content of the digital breeding method for shrimp aquaculture is implemented.
[0039] A storage medium stores computer-executable instructions. When the computer-executable instructions are loaded and executed by a processor, the content of the digital breeding method for shrimp aquaculture is implemented.
[0040] In summary, for the digital breeding method and system for shrimp aquaculture of the present invention, compared with traditional technologies, the present invention constructs a digital breeding model, observes the shrimp aquaculture at different growth stages and assists in adjusting the aquaculture strategy, can form refined management in the process of shrimp aquaculture, and can realize real-time monitoring of the quality of shrimp, improving the quality of shrimp aquaculture.
[0041] Next, through the accompanying drawings and embodiments, the technical method of the present invention will be further described in detail. Description of the Drawings
[0042] Figure 1 It is a flowchart of the steps of a digital breeding method for shrimp aquaculture of the present invention;
[0043] Figure 2 It is a module diagram of a digital breeding system for shrimp aquaculture of the present invention. Detailed Embodiments
[0044] The technical method of the present invention will be further described below through the accompanying drawings and embodiments. It should be noted that: unless otherwise specifically stated, the relative arrangements, numerical expressions and values of the components and steps described in these embodiments do not limit the scope of the present application.
[0045] The description of at least one exemplary embodiment is actually only illustrative and in no way limits the present application and its application or use.
[0046] For technologies, systems and devices known to those of ordinary skill in the relevant art, detailed discussions may not be made, but where appropriate, the technologies, systems and devices should be regarded as part of the specification.
[0047] In all the examples shown and discussed here, any specific value should be construed as merely exemplary and not as a limitation. Therefore, other examples of the exemplary embodiments may have different values.
[0048] Unless otherwise defined, the technical terms or scientific terms used in the present invention shall have the ordinary meanings understood by those of ordinary skill in the field to which the present invention pertains.
[0049] The present invention provides a shrimp breeding and selection method based on a digital breeding and selection model, comprising the following steps:
[0050] S1. Obtain past breeding and selection information, analyze the past breeding and selection information to obtain past breeding and selection data and past breeding and selection results;
[0051] S2. Construct a digital breeding and selection model, use the past breeding and selection data as input and the past breeding and selection results as output to train the digital breeding and selection model;
[0052] Further, the digital breeding and selection model in S2 is based on a neural network model and includes:
[0053] Head training layer: The head training layer is used to receive initial information and complete overall training in combination with the process training layer and the tail training layer to generate target breeding and selection data;
[0054] Process training layer: The process training includes completing overall training in combination with the head training layer and the tail training layer or receiving stage information and performing partial training in combination with the tail training layer to obtain optimized breeding and selection data;
[0055] Tail training layer; The tail training layer is used to cooperate with the head training layer and the process training layer to complete overall training and partial training and output past breeding and selection results.
[0056] Further, the specific content of receiving stage information and performing partial training in combination with the tail training layer to obtain optimized breeding and selection data is:
[0057] The process training layer includes several layers arranged in sequence;
[0058] Different process training layers correspond to different shrimp growth stages. Input the stage information into the corresponding process training layer, and the corresponding process training layer and the process training layers arranged after it combine with the tail training layer to perform different partial trainings.
[0059] Further, during the training process of the digital breeding and selection model, overall training is the main body, and overall training is carried out first and then partial training;
[0060] The overall training and partial training are respectively set with a main iteration number and a partial iteration number, and the main iteration number is less than the partial iteration number of the tail training layer.
[0061] S3. Obtain the current shrimp population data and the target breeding results, and input the current shrimp population data and the target breeding results into the digital breeding model for training to obtain the target breeding data;
[0062] Further, the specific content of S3. Obtain the current shrimp population data and the target breeding results, and input the current shrimp population data and the target breeding results into the digital breeding model for training to obtain the target breeding data is as follows;
[0063] The current shrimp population data includes the shrimp population environment, the shrimp population genealogy, and the shrimp population trait data;
[0064] The target breeding result is the constraint condition of the digital breeding model;
[0065] If the breeding result still does not meet the constraint condition when the number of iterations reaches the main iteration number, continue the iteration until the breeding result meets the constraint condition and output the target breeding data;
[0066] If the breeding result meets the constraint condition when the number of iterations does not reach the main iteration number, directly output the target breeding data.
[0067] The target breeding data can include selecting shrimp that meet certain basic conditions from the existing shrimp population, or can include formulating a scientific and reasonable breeding plan according to the breeding goal and the characteristics of the basic shrimp population. It includes data such as feed feeding amount, feeding frequency, water quality management, water temperature control, and breeding density. For example, if the breeding goal is to improve the meat quality of shrimp, the nutritional components of the feed can be appropriately adjusted to increase the proportion of nutrients such as protein.
[0068] S4. Carry out breeding treatment on the current shrimp population according to the target breeding data, and use the visual processing system to conduct phased observation on the traits of the current shrimp population to obtain phased trait data;
[0069] Further, the specific content of S4. Carry out breeding treatment on the current shrimp population according to the target breeding data, and use the visual processing system to conduct phased observation on the traits of the current shrimp population to obtain phased trait data is as follows:
[0070] Set the installation positions of the breeding environment equipment according to the activities of the shrimp population, and install the image acquisition equipment at the installation positions to obtain the acquired images;
[0071] Select a suitable underwater camera or image acquisition equipment and install it at a suitable position in the breeding pond or breeding environment to ensure that the activities of the shrimp population can be clearly photographed. At the same time, equip corresponding lighting equipment to ensure that clear images can also be obtained in the underwater environment with relatively dim light.
[0072] Reasonably set the frequency of data collection according to the growth rate and breeding stage of shrimp. In the initial stage of shrimp growth, the data collection frequency can be appropriately higher to more timely understand the growth changes of the shrimp population; as the growth of shrimp gradually stabilizes, the data collection frequency can be correspondingly reduced.
[0073] Based on the deep learning algorithm, analyze and process the collected images to obtain stage trait data;
[0074] Adopt advanced image recognition and processing software, such as algorithms based on deep learning, to analyze and process the collected images. This software should be able to accurately identify the shrimp body and extract and analyze trait data such as the body length, body width, swimming speed, and behavior pattern of the shrimp.
[0075] The stage trait data includes growth rate, disease resistance ability, and individual size;
[0076] The stage observation is the observation of the growth stage of shrimp.
[0077] Determine the excellent traits of the shrimp to be selected, such as fast growth rate, strong disease resistance ability, large individual size, etc. For example, if the goal is to improve the growth rate of shrimp, it is necessary to pay attention to the individual characteristics with faster growth rate, such as indicators like body length and weight gain, during the selection process.
[0078] Shrimp that meet certain basic conditions can be selected from the existing shrimp population as the basic population, thereby providing an update basis for the digital selection and breeding model. These basic conditions can include good health status, no genetic diseases, etc. Ensure that the basic shrimp population has a certain quantity and genetic diversity to provide a sufficient genetic variation basis for subsequent selection and breeding.
[0079] Selection and breeding decision: Based on the analysis and evaluation results of the stage trait data, make a selection and breeding decision. Select those individuals with excellent traits as parents to breed and cultivate the next generation of shrimp population. At the same time, for those individuals with poor trait performance, consideration can be given to elimination or further improvement.
[0080] S5. Input the stage trait data into the digital selection and breeding model to adjust the target selection and breeding data to obtain optimized selection and breeding data;
[0081] Furthermore, S5. Input the stage trait data into the digital selection and breeding model to adjust the target selection and breeding data to obtain optimized selection and breeding data:
[0082] Take the stage trait data as stage information and input it into the corresponding process training layer and conduct different parts of training to obtain optimized selection and breeding data.
[0083] S6. Breed the shrimp population according to the optimized selection and breeding data to obtain the target shrimp body.
[0084] According to the breeding decision, artificial breeding or natural breeding is carried out on the selected parents. During the breeding process, a vision processing system is continuously used to monitor the traits of the offspring shrimp population and collect data, forming a continuous breeding cycle to continuously optimize the traits of the shrimp population.
[0085] A digital breeding system for shrimp farming, comprising:
[0086] Data acquisition unit: Obtain past breeding information, analyze the past breeding information to obtain past breeding data and past breeding results;
[0087] Model construction unit: Construct a digital breeding model, use the past breeding data as input and the past breeding results as output, and train the digital breeding model;
[0088] Breeding decision unit: The breeding decision unit includes an overall decision module and a partial decision module;
[0089] The overall decision module includes obtaining current shrimp population data and target breeding results, and bringing the current shrimp population data and target breeding results into the digital breeding model for training to obtain target breeding data;
[0090] The partial decision module includes performing aquaculture treatment on the current shrimp population according to the target breeding data, using a vision processing system to conduct phased observation of the traits of the current shrimp population to obtain phased trait data, and inputting the phased trait data into the digital breeding model to adjust the target breeding data to obtain optimized breeding data.
[0091] An electronic device, comprising a memory and a processor. A computer program is stored in the memory, and when the processor calls the computer program in the memory, the content of the digital breeding method for shrimp farming is implemented.
[0092] A storage medium stores computer-executable instructions, and when the computer-executable instructions are loaded and executed by a processor, the content of the digital breeding method for shrimp farming is implemented.
[0093] Finally, it should be noted that the above embodiments are only used to illustrate the technical method of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify the technical method of the present invention or make equivalent replacements, and these modifications or equivalent replacements cannot make the modified technical method deviate from the spirit and scope of the technical method of the present invention.
Claims
1. A digital breeding selection method for shrimp farming, characterized in that, It includes the following steps: S1. Obtain past breeding information, analyze the past breeding information to obtain past breeding data and past breeding results; S2. Construct a digital breeding model, use the past breeding data as input and the past breeding results as output to train the digital breeding model; S3. Obtain current shrimp population data and target breeding results, input the current shrimp population data and target breeding results into the digital breeding model for training to obtain target breeding data; S4. Perform aquaculture treatment on the current shrimp population according to the target breeding data, and use a vision processing system to conduct phased observation of the traits of the current shrimp population to obtain phased trait data; S5. Input the phased trait data into the digital breeding model to adjust the target breeding data to obtain optimized breeding data; S6. Perform aquaculture on the shrimp population according to the optimized breeding data to obtain target shrimp bodies.
2. The digital breeding and selection method for shrimp farming according to claim 1, wherein In S2, the digital breeding model is based on a neural network model and includes: Head training layer: The head training layer is used to receive initial information and complete overall training in combination with the process training layer and the tail training layer to generate target breeding data; Process training layer: The process training includes completing overall training in combination with the head training layer and the tail training layer or receiving phased information and performing partial training in combination with the tail training layer to obtain optimized breeding data; Tail training layer; The tail training layer is used to cooperate with the head training layer and the process training layer to complete overall training and partial training and output past breeding results.
3. The digital breeding and selection method for shrimps according to claim 2, characterized in that The specific content of receiving phased information and performing partial training in combination with the tail training layer to obtain optimized breeding data is: The process training layer includes several layers arranged in sequence; Different process training layers correspond to different growth stages of shrimp. Input the phased information into the corresponding process training layer, and the corresponding process training layer and the process training layers arranged after it combine with the tail training layer to perform different partial trainings.
4. A digital breeding and selection method for shrimp farming according to claim 3, characterized in that, During the training process of the digital breeding model, overall training is the main body, and overall training is carried out first and then partial training; The overall training and partial training are respectively set with main iteration times and partial iteration times, and the main iteration times are less than the partial iteration times of the tail training layer.
5. A digital breeding method for shrimp farming according to claim 4, characterized in that, The specific content of S3. Obtain current shrimp population data and target breeding results, input the current shrimp population data and target breeding results into the digital breeding model for training to obtain target breeding data is; The current shrimp population data includes shrimp population environment, shrimp population genealogy, and shrimp population trait data; The target breeding result is the constraint condition of the digital breeding model; If the breeding result does not meet the constraint condition when the iteration times reach the main iteration times, continue to iterate until the breeding result meets the constraint condition and output the target breeding data; If the breeding result meets the constraint condition when the iteration times do not reach the main iteration times, directly output the target breeding data.
6. The digital breeding and selection method for shrimp culture according to claim 5, characterized in that, The specific content of S4. Perform aquaculture treatment on the current shrimp population according to the target breeding data, and use a vision processing system to conduct phased observation of the traits of the current shrimp population to obtain phased trait data is: Set the installation sites of aquaculture environment equipment according to the activities of the shrimp population, install image acquisition equipment at the installation sites for acquisition to obtain acquisition images; Based on deep learning algorithms, the acquired images are analyzed and processed to obtain stage trait data; The stage trait data includes growth rate, disease resistance, and individual size; The stage observation is the observation of the growth stage of shrimp.
7. A digital breeding and selection method for shrimp culture according to claim 6, characterized in that, S5. Input the stage trait data into the digital breeding model to adjust the target breeding data to obtain optimized breeding data: Use the stage trait data as stage information to input into the corresponding process training layer and perform different parts of training to obtain optimized breeding data.
8. A digital breeding and selection system for shrimp farming, characterized in that, Including: Data acquisition unit: Obtain past breeding information, analyze the past breeding information to obtain past breeding data and past breeding results; Model construction unit: Construct a digital breeding model, use the past breeding data as input and the past breeding results as output to train the digital breeding model; Breeding decision unit: The breeding decision unit includes an overall decision module and a partial decision module; Overall decision module: Obtain the current shrimp population data and the target breeding results, input the current shrimp population data and the target breeding results into the digital breeding model for training to obtain the target breeding data; Partial decision module: Perform aquaculture treatment on the current shrimp population according to the target breeding data, use a vision processing system to perform stage observation of the traits of the current shrimp population to obtain stage trait data, and input the stage trait data into the digital breeding model to adjust the target breeding data to obtain optimized breeding data.
9. An electronic device, characterized in that, It includes a memory and a processor. A computer program is stored in the memory. When the processor calls the computer program in the memory, the content of the digital breeding method for shrimp aquaculture according to any one of claims 1 to 7 is implemented.
10. A storage medium, characterized in that, Computer-executable instructions are stored in the storage medium. When the computer-executable instructions are loaded and executed by the processor, the content of the digital breeding method for shrimp aquaculture according to any one of claims 1 to 7 is implemented.