An optimization method for laying hen breeding system based on a large model of artificial intelligence

Abstract

The present application provides a method for optimizing a laying hen breeding system based on a large model of artificial intelligence, which belongs to the field of breeding technology. The method includes: determining a first target individual from a plurality of breeding individuals based on an image, and obtaining target historical feature data based on a target historical image; constructing a first neural network training model, importing a training set into the first neural network training model and training; obtaining a plurality of individual feature data based on a historical data image of a second cycle, inputting the individual feature data into the trained first neural network training model, and obtaining a plurality of predicted individual feature data; comparing a plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and according to the comparison result, confirming a second target individual from the plurality of individual feature data. The present application provides a device for optimizing a laying hen breeding system based on a large model of artificial intelligence and an electronic device.

Description

Technical Field

[0001] The present application relates to the field of breeding technology, and in particular to an optimization method for laying hen breeding systems based on a large model of artificial intelligence. Background Art

[0002] In the process of centralized breeding, individuals with unstable factors will appear in the farms, such as injured individuals, sick individuals and abnormal individuals. When there are too many of these individuals, it will lead to a reduction in breeding output. In the prior art, usually only image recognition is used to extract key information in the image, such as edges, corners, textures, etc. through algorithms, make decisions based on the output results of the classifier, achieve target screening, and mark these individuals that will cause unstable factors from the monitoring screen or other visual images. However, since image recognition requires a large amount of computing power, the image recognition method will occupy more resources. However, since illness and other conditions have a certain delay, when individuals with unstable factors are discovered, it is proved that more individuals who are prone to become unstable have appeared. Therefore, how to make timely adjustments in advance is an urgent problem to be solved. Summary of the invention

[0003] The embodiment of the present application provides an optimization method for laying hen breeding systems based on a large model of artificial intelligence to improve the above-mentioned problems.

[0004] In order to achieve the above objectives, this application adopts the following technical solutions:

[0005] In the first aspect, the present application embodiment proposes a method for optimizing a laying hen breeding system based on a large model of artificial intelligence, the method comprising:

[0006] Determine a first target individual from a plurality of farmed individuals based on the image, and determine comprehensive individual data used to characterize the behavior, action, and image features of the first target individual as first target individual feature data, wherein the first target individual includes an injured individual, a diseased individual, and an abnormal individual;

[0007] Acquire a target historical image based on the historical data image of the first cycle, acquire target historical feature data based on the target historical image, and acquire a training set based on the first target individual feature data and the target historical feature data;

[0008] Constructing a first neural network training model, importing the training set into the first neural network training model and training it, the first neural network training model is used to obtain predicted individual feature data based on the input individual feature data;

[0009] Acquire a plurality of individual feature data based on the historical data image of the second period, input the individual feature data into the trained first neural network training model, and acquire a plurality of predicted individual feature data;

[0010] Comparing the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifying a second target individual from the plurality of individual feature data according to the comparison result;

[0011] The breeding system is optimized according to the quantitative relationship between the second target individuals and the breeding individuals.

[0012] In combination with the first aspect, in some embodiments, a first target individual is determined from a plurality of farmed individuals based on an image, and comprehensive individual data used to characterize the behavior, action, and image features of the first target individual is determined as first target individual feature data, and the first target individual includes an injured individual, a diseased individual, and an abnormal individual, including:

[0013] Acquire multiple images of breeding individuals based on the image;

[0014] In response to a selection instruction from a user, a target individual image is selected from a plurality of farmed individual images, and the farmed individual corresponding to the target individual image is determined as a first target individual;

[0015] Extract features from target individual images.

[0016] In combination with the first aspect, in some embodiments, before acquiring a plurality of individual feature data based on the historical data image of the second period, inputting the individual feature data into the first neural network training model, and acquiring a plurality of predicted individual feature data, the method includes:

[0017] Importing the training set into the first neural network training model and performing iterative calculations, and obtaining predicted individual feature data;

[0018] Comparing the predicted individual characteristic data with the first target individual characteristic data and obtaining error data;

[0019] When the error data is less than the first preset data, the training process of the first neural network training model is stopped and the trained first neural network training model is obtained.

[0020] In combination with the first aspect, in some embodiments, a plurality of predicted individual feature data obtained from the first neural network training model is compared with the first target individual feature data, and a second target individual is identified from the plurality of individual feature data according to the comparison result, including:

[0021] Compare the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and determine that the predicted individual feature data whose difference value with the first target individual feature data is less than the second preset data is the second target individual feature data;

[0022] An individual corresponding to the second target individual characteristic data is determined as the second target individual.

[0023] In combination with the first aspect, in some embodiments, the breeding system is optimized according to the quantitative relationship between the second target individuals and the breeding individuals, including:

[0024] Obtain historical environmental data of the farm, including temperature data, humidity data and breeding density data;

[0025] Determine variable data of breeding individuals of the second target individual within multiple breeding cycles based on historical data of the quantitative relationship between the second target individual and the individual characteristic data in multiple breeding cycles;

[0026] Constructing a second neural network training model, wherein temperature data, humidity data and breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data and variable data, and output predicted variable data according to the input humidity data, breeding density data and variable data;

[0027] Acquire multiple sets of preset temperature data, preset humidity data, and preset breeding density data, and input the preset temperature data, preset humidity data, and preset breeding density data into a second neural network training model;

[0028] The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data.

[0029] In combination with the first aspect, in some embodiments, optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data, including:

[0030] Determine the optimal environmental data according to the variable data output by the second neural network training model, the optimal environmental data being the environmental data in which the second target individuals account for the least proportion among the breeding individuals;

[0031] Determine a set of preset temperature data, preset humidity data, and preset breeding density data corresponding to the optimal environment data as target control data;

[0032] Carry out environmental control on the farm based on target control data.

[0033] In combination with the first aspect, in some embodiments, a second neural network training model is constructed, wherein temperature data, humidity data, and breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data, and variable data, and output predicted variable data based on the input humidity data, breeding density data, and variable data, including:

[0034] Comparing the predicted variable data output by the second neural network training model based on the historical environmental data with the variable data confirmed in multiple cycles, and determining an error value;

[0035] When the error data is less than the third preset data, the training process of the second neural network training model is stopped and the trained second neural network training model is obtained.

[0036] In combination with the first aspect, in some implementations, comparing the predicted variable data output by the second neural network training model according to the historical environmental data with the variable data confirmed in multiple cycles, and determining the error value, includes:

[0037] The relationship between the error value, the predicted variable data and the variable data satisfies:

[0038] T = (y1-yt)y1

[0039] Among them, the error value is T, the value of the predicted variable data is yt, and the value of the variable data is y1.

[0040] The second aspect of the embodiment of the present application proposes a large-scale model-based laying hen breeding system optimization device based on artificial intelligence, and the device is configured as follows:

[0041] Determine a first target individual from a plurality of farmed individuals based on the image, and determine comprehensive individual data used to characterize the behavior, action, and image features of the first target individual as first target individual feature data, wherein the first target individual includes an injured individual, a diseased individual, and an abnormal individual;

[0042] Acquire a target historical image based on the historical data image of the first cycle, acquire target historical feature data based on the target historical image, and acquire a training set based on the first target individual feature data and the target historical feature data;

[0043] Constructing a first neural network training model, importing the training set into the first neural network training model and training it, the first neural network training model is used to obtain predicted individual feature data based on the input individual feature data;

[0044] Acquire a plurality of individual feature data based on the historical data image of the second period, input the individual feature data into the trained first neural network training model, and acquire a plurality of predicted individual feature data;

[0045] Comparing the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifying a second target individual from the plurality of individual feature data according to the comparison result;

[0046] The breeding system is optimized according to the quantitative relationship between the second target individuals and the breeding individuals.

[0047] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0048] A first target individual is determined from a plurality of farmed individuals based on the image, and comprehensive individual data used to characterize the behavior, action, and image features of the first target individual is determined as first target individual feature data. The first target individual includes injured individuals, diseased individuals, and abnormal individuals, including:

[0049] Acquire multiple images of breeding individuals based on the image;

[0050] In response to a selection instruction from a user, a target individual image is selected from a plurality of farmed individual images, and the farmed individual corresponding to the target individual image is determined as a first target individual;

[0051] Extract features from target individual images.

[0052] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0053] Before obtaining a plurality of individual feature data based on the historical data image of the second cycle, inputting the individual feature data into the first neural network training model, and obtaining a plurality of predicted individual feature data, the method includes:

[0054] Importing the training set into the first neural network training model and performing iterative calculations, and obtaining predicted individual feature data;

[0055] Comparing the predicted individual characteristic data with the first target individual characteristic data and obtaining error data;

[0056] When the error data is less than the first preset data, the training process of the first neural network training model is stopped and the trained first neural network training model is obtained.

[0057] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0058] Comparing the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifying the second target individual from the plurality of individual feature data according to the comparison result, including:

[0059] Compare the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and determine that the predicted individual feature data whose difference value with the first target individual feature data is less than the second preset data is the second target individual feature data;

[0060] An individual corresponding to the second target individual characteristic data is determined as the second target individual.

[0061] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0062] According to the quantitative relationship between the second target individuals and the breeding individuals, the breeding system is optimized, including:

[0063] Obtain historical environmental data of the farm, including temperature data, humidity data and breeding density data;

[0064] Determine variable data of breeding individuals of the second target individual within multiple breeding cycles based on historical data of the quantitative relationship between the second target individual and the individual characteristic data in multiple breeding cycles;

[0065] Constructing a second neural network training model, wherein temperature data, humidity data and breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data and variable data, and output predicted variable data according to the input humidity data, breeding density data and variable data;

[0066] Acquire multiple sets of preset temperature data, preset humidity data, and preset breeding density data, and input the preset temperature data, preset humidity data, and preset breeding density data into a second neural network training model;

[0067] The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data.

[0068] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0069] The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data, including:

[0070] Determine the optimal environmental data according to the variable data output by the second neural network training model, the optimal environmental data being the environmental data in which the second target individuals account for the least proportion among the breeding individuals;

[0071] Determine a set of preset temperature data, preset humidity data, and preset breeding density data corresponding to the optimal environment data as target control data;

[0072] Carry out environmental control on the farm based on target control data.

[0073] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0074] Construct a second neural network training model, wherein temperature data, humidity data and breeding density data are used as three different dimensions, and the three dimensions constitute the characteristic vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data and variable data, and output the predicted variable data according to the input humidity data, breeding density data and variable data, including:

[0075] Comparing the predicted variable data output by the second neural network training model based on the historical environmental data with the variable data confirmed in multiple cycles, and determining an error value;

[0076] When the error data is less than the third preset data, the training process of the second neural network training model is stopped and the trained second neural network training model is obtained.

[0077] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0078] The predicted variable data output by the second neural network training model according to the historical environmental data are compared with the variable data confirmed in multiple cycles, and the error value is determined, including:

[0079] The relationship between the error value, the predicted variable data and the variable data satisfies:

[0080] T = (y1-yt)y1

[0081] Among them, the error value is T, the value of the predicted variable data is yt, and the value of the variable data is y1.

[0082] A third aspect of an embodiment of the present invention provides an electronic device, the electronic device comprising:

[0083] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method proposed in the first aspect of the embodiment of the present invention.

[0084] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided in the first aspect of the embodiment of the present invention.

[0085] In summary, the above method and device have the following technical effects:

[0086] The present application proposes an optimization method for laying hen farming systems based on a large model of artificial intelligence. First, individuals with unstable factors, such as injured individuals, diseased individuals and abnormal individuals, are identified among multiple farming individuals, and their characteristic data are extracted, and their historical characteristic data in the first cycle are obtained. A neural network model is trained with these data so that the neural network model can use the input data to predict the predicted characteristic data of the individuals in the second cycle, and predict the data of unstable individuals in the second cycle. The farming environment is adjusted in advance according to the prediction results, thereby improving the farming output. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 A schematic flow chart of an optimization method for laying hen farming systems based on a large artificial intelligence model proposed in an embodiment of the present application. DETAILED DESCRIPTION

[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0089] This application proposes an optimization method for laying hen breeding system based on a large model of artificial intelligence. Figure 1 , including the following steps:

[0090] S101: Determine a first target individual from a plurality of farmed individuals based on an image, and determine comprehensive individual data used to characterize the behavior, action and image features of the first target individual as first target individual feature data, wherein the first target individual includes injured individuals, diseased individuals and abnormal individuals.

[0091] It can be understood that the first target individual in the present application refers to individuals that have unstable factors in the breeding production process, which affect egg production and meat production, such as injured individuals, diseased individuals, and abnormal individuals. These individuals have certain image features. As for how to characterize each feature, for example, the weight value of each feature can be determined separately, and then the weight value of all features of each individual can be calculated to obtain a comprehensive individual data for characterizing the behavior, action and image features of the first target individual, that is, the first target individual feature data. The larger the value of the data, the higher the probability of proving that it is an unstable individual. In this embodiment, a threshold value can be set. When the value is greater than the threshold value, it is proved that it is an unstable individual. Of course, in other embodiments, other methods can also be used to obtain the first target individual feature data, which are not limited here.

[0092] In this embodiment, the image can be an image taken as a whole, or an image taken separately for each breeding unit, which is not limited here. As for how to determine the first target individual, as an implementation method, first, multiple breeding individual images can be obtained based on the image, so that it is easy to mark. Then, manual labeling can be used to select the target individual image from multiple breeding individual images in response to the user's selection instruction, determine the breeding individual corresponding to the target individual image as the first target individual, and finally, perform feature extraction.

[0093] S102: Acquire a target historical image based on the historical data image of the first cycle, acquire target historical feature data based on the target historical image, and acquire a training set based on the first target individual feature data and the target historical feature data.

[0094] S103: construct a first neural network training model, import the training set into the first neural network training model and perform training, the first neural network training model is used to obtain predicted individual feature data according to the input individual feature data.

[0095] In this embodiment, the first cycle can be a fixed period of time, such as a week, a day, etc., which is not limited here. The historical image can be an image of a single individual at different time points within the cycle. Through multiple historical images, multiple feature data of an individual at multiple time points within the first cycle can be obtained. It is understandable that, for example, before an individual is determined to be a sick individual, its image features will change, but these changes are not sufficient to determine that the individual is a sick individual at the current time point. Therefore, by forming a training set with the feature data corresponding to an individual and its feature data group after the end of the cycle, artificial intelligence can be used to timely predict its subsequent features when individual features change.

[0096] In some embodiments, in order to make the prediction result more accurate, during the training process, the training set can also be imported into the first neural network training model and iterative calculation can be performed to obtain predicted individual feature data. Then, the predicted individual feature data is compared with the first target individual feature data, and error data is obtained. Finally, when the error data is less than the first preset data, the training process of the first neural network training model is stopped and the trained first neural network training model is obtained.

[0097] S104: Acquire a plurality of individual feature data based on the historical data image of the second cycle, input the individual feature data into the trained first neural network training model, and acquire a plurality of predicted individual feature data.

[0098] It is understandable that after obtaining the trained first neural network training model, multiple individual feature data in the historical data image of the second cycle can be imported into the neural network training model, so that multiple predicted individual feature data can be obtained. It is understandable that the first cycle and the second cycle are different breeding cycles. Therefore, through multiple predicted individual feature data, the individual feature data of each individual at the current time after the second cycle breeding is completed or during the second cycle breeding can be predicted.

[0099] S105: Compare the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identify the second target individual from the plurality of individual feature data according to the comparison result.

[0100] It can be understood that after obtaining the predicted characteristic data of each individual at the current time, the data is compared with the first target individual characteristic data, that is, the predicted characteristic data is the same or similar to the first target individual characteristic data, which is the second target individual, that is, the individual with a higher probability of unstable situation in the future. As an implementation method, first, a plurality of predicted individual characteristic data obtained from the first neural network training model are compared with the first target individual characteristic data, and the predicted individual characteristic data whose difference value with the first target individual characteristic data is less than the second preset data is determined to be the second target individual characteristic data, and then the individual corresponding to the second target individual characteristic data is determined to be the second target individual. By analyzing the predicted individual characteristic data of each farmed individual, the current status of the farmed individual can be obtained without manual labeling or determination.

[0101] S106: Optimizing the breeding system according to the quantitative relationship between the second target individuals and the breeding individuals.

[0102] After obtaining the prediction data of each individual in the second cycle, the current environment can be regulated according to the prediction data to change the final prediction data.

[0103] Specifically, as an implementation, step S106 may include the following steps:

[0104] S1061: Obtain farm environmental data, which includes temperature data, humidity data, and breeding density data.

[0105] It is understandable that, for the breeding environment of the breeding farm, temperature data, humidity data and breeding density data are three data with great influence and easy to control. Of course, in other embodiments, it can also be some other data, such as oxygen content, etc., which is not limited here.

[0106] S1062: Based on the historical data of the quantitative relationship between the second target individual and the individual characteristic data in multiple breeding cycles, determine the variable data of the breeding individuals of the second target individual in multiple breeding cycles.

[0107] It can be understood that in this embodiment, the variable data is the data of the increase or decrease of unstable individuals in multiple cycles.

[0108] S1063: Construct a second neural network training model, wherein temperature data, humidity data and breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data and variable data, and output the predicted variable data based on the input humidity data, breeding density data and variable data.

[0109] S1064: Acquire multiple sets of preset temperature data, preset humidity data, and preset breeding density data, and input the preset temperature data, preset humidity data, and preset breeding density data into a second neural network training model.

[0110] It is understandable that, in the process of training, in order to improve the accuracy of the prediction results, the predicted variable data output by the second neural network training model based on the historical environmental data can also be compared with the variable data confirmed in multiple cycles, and the error value can be determined. When the error data is less than the third preset data, the training process of the second neural network training model is stopped and the trained second neural network training model is obtained. For the determination of the error value, the relationship between the error value, the predicted variable data and the variable data satisfies:

[0111] T = (y1-yt)y1

[0112] Among them, the error value is T, the value of the predicted variable data is yt, and the value of the variable data is y1.

[0113] Of course, in other implementations, the error value may be determined in other ways, which are not limited here.

[0114] S1065: Determine optimal environmental data based on the variable data output by the second neural network training model, and perform environmental control on the farm based on the optimal environmental data.

[0115] Specifically, the optimal environmental data can be determined based on the variable data output by the second neural network training model. The optimal environmental data is the environmental data in which the second target individuals account for the least proportion among the farmed individuals. Then, a set of preset temperature data, preset humidity data, and preset breeding density data corresponding to the optimal environmental data are determined as target control data. Finally, the environment of the farm is controlled based on the target control data.

[0116] The present application proposes an optimization method for laying hen farming systems based on a large model of artificial intelligence. First, individuals with unstable factors, such as injured individuals, diseased individuals and abnormal individuals, are identified among multiple farming individuals, their feature data are extracted, and their historical feature data in the first cycle are obtained. A neural network model is trained with these data so that the neural network model can use the input data to predict the predicted feature data of individuals in the second cycle, predict the data of unstable individuals in the second cycle, and timely regulate the farming environment according to the prediction results, thereby improving the farming output.

[0117] Based on the same inventive concept, the embodiment of the present application also proposes an optimization device for laying hen breeding system based on a large model of artificial intelligence, and the device is configured as follows:

[0118] Determine a first target individual from a plurality of farmed individuals based on the image, and determine comprehensive individual data used to characterize the behavior, action, and image features of the first target individual as first target individual feature data, wherein the first target individual includes an injured individual, a diseased individual, and an abnormal individual;

[0119] Acquire a target historical image based on the historical data image of the first cycle, acquire target historical feature data based on the target historical image, and acquire a training set based on the first target individual feature data and the target historical feature data;

[0120] Constructing a first neural network training model, importing the training set into the first neural network training model and training it, the first neural network training model is used to obtain predicted individual feature data based on the input individual feature data;

[0121] Acquire a plurality of individual feature data based on the historical data image of the second period, input the individual feature data into the trained first neural network training model, and acquire a plurality of predicted individual feature data;

[0122] Comparing the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifying a second target individual from the plurality of individual feature data according to the comparison result;

[0123] The breeding system is optimized according to the quantitative relationship between the second target individuals and the breeding individuals.

[0124] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0125] A first target individual is determined from a plurality of farmed individuals based on the image, and comprehensive individual data used to characterize the behavior, action, and image features of the first target individual is determined as first target individual feature data, wherein the first target individual includes injured individuals, diseased individuals, and abnormal individuals, including:

[0126] Acquire multiple images of breeding individuals based on the image;

[0127] In response to a selection instruction from a user, a target individual image is selected from a plurality of farmed individual images, and the farmed individual corresponding to the target individual image is determined as a first target individual;

[0128] Extract features from target individual images.

[0129] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0130] Before obtaining a plurality of individual feature data based on the historical data image of the second cycle, inputting the individual feature data into the first neural network training model, and obtaining a plurality of predicted individual feature data, the method includes:

[0131] Importing the training set into the first neural network training model and performing iterative calculations, and obtaining predicted individual feature data;

[0132] Comparing the predicted individual characteristic data with the first target individual characteristic data and obtaining error data;

[0133] When the error data is less than the first preset data, the training process of the first neural network training model is stopped and the trained first neural network training model is obtained.

[0134] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0135] Comparing the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifying the second target individual from the plurality of individual feature data according to the comparison result, including:

[0136] Compare the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and determine that the predicted individual feature data whose difference value with the first target individual feature data is less than the second preset data is the second target individual feature data;

[0137] An individual corresponding to the second target individual characteristic data is determined as the second target individual.

[0138] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0139] According to the quantitative relationship between the second target individuals and the breeding individuals, the breeding system is optimized, including:

[0140] Obtain historical environmental data of the farm, including temperature data, humidity data and breeding density data;

[0141] Determine variable data of breeding individuals of the second target individual within multiple breeding cycles based on historical data of the quantitative relationship between the second target individual and the individual characteristic data in multiple breeding cycles;

[0142] Constructing a second neural network training model, wherein temperature data, humidity data and breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data and variable data, and output predicted variable data according to the input humidity data, breeding density data and variable data;

[0143] Acquire multiple sets of preset temperature data, preset humidity data, and preset breeding density data, and input the preset temperature data, preset humidity data, and preset breeding density data into a second neural network training model;

[0144] The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data.

[0145] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0146] The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data, including:

[0147] Determine the optimal environmental data according to the variable data output by the second neural network training model, the optimal environmental data being the environmental data in which the second target individuals account for the least proportion among the breeding individuals;

[0148] Determine a set of preset temperature data, preset humidity data, and preset breeding density data corresponding to the optimal environment data as target control data;

[0149] Carry out environmental control on the farm based on target control data.

[0150] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0151] Construct a second neural network training model, wherein temperature data, humidity data and breeding density data are used as three different dimensions, and the three dimensions constitute the characteristic vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, humidity data, breeding density data and variable data, and output the predicted variable data according to the input humidity data, breeding density data and variable data, including:

[0152] Comparing the predicted variable data output by the second neural network training model based on the historical environmental data with the variable data confirmed in multiple cycles, and determining an error value;

[0153] When the error data is less than the third preset data, the training process of the second neural network training model is stopped and the trained second neural network training model is obtained.

[0154] In conjunction with the second aspect, in some embodiments, the apparatus is configured to:

[0155] The predicted variable data output by the second neural network training model according to the historical environmental data are compared with the variable data confirmed in multiple cycles, and the error value is determined, including:

[0156] The relationship between the error value, the predicted variable data and the variable data satisfies:

[0157] T = (y1-yt)y1

[0158] Among them, the error value is T, the value of the predicted variable data is yt, and the value of the variable data is y1.

[0159] The present application proposes an optimization device for laying hen breeding systems based on a large model of artificial intelligence. First, individuals with unstable factors, such as injured individuals, diseased individuals and abnormal individuals, are identified among multiple breeding individuals, their characteristic data are extracted, and their historical characteristic data in the first cycle are obtained. A neural network model is trained with these data so that the neural network model can use the input data to predict the predicted characteristic data of individuals in the second cycle, predict the data of unstable individuals in the second cycle, and timely adjust the breeding environment according to the prediction results, thereby improving breeding output.

[0160] Based on the same inventive concept, an embodiment of the present application further proposes an electronic device, the electronic device comprising:

[0161] At least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method for optimizing the laying hen breeding system based on a large model of artificial intelligence according to an embodiment of the present application.

[0162] In addition, to achieve the above-mentioned purpose, an embodiment of the present application also proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the egg-laying chicken breeding system optimization method based on a large model of artificial intelligence of the embodiment of the present application is implemented.

[0163] The following is a detailed introduction to the various components of electronic equipment:

[0164] The processor is the control center of the electronic device, which can be a processor or a general term for multiple processing elements. For example, the processor is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (field programmable gate arrays, FPGAs).

[0165] Optionally, the processor can perform various functions of the electronic device by running or executing a software program stored in the memory, and calling data stored in the memory.

[0166] The memory is used to store the software program for executing the solution of the present invention, and the execution is controlled by the processor. The specific implementation method can refer to the above method embodiment and will not be repeated here.

[0167] Optionally, the memory may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may be integrated with the processor, or may exist independently and be coupled to the processor through an interface circuit of the electronic device, which is not specifically limited in the embodiments of the present invention.

[0168] A transceiver is used to communicate with a network device or a terminal device.

[0169] Optionally, the transceiver may include a receiver and a transmitter, wherein the receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0170] Optionally, the transceiver may be integrated with the processor, or may exist independently and be coupled to the processor via an interface circuit of the router, which is not specifically limited in the embodiment of the present invention.

[0171] In addition, the technical effects of the electronic device can refer to the technical effects of the data transmission method in the above method embodiment, which will not be repeated here.

[0172] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0173] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0174] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0175] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0176] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0177] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0178] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

Claims

1. A method for optimizing laying hen breeding system based on a large model of artificial intelligence, characterized in that: The method comprises: A first target individual is determined from a plurality of farmed individuals based on an image, and comprehensive individual data for characterizing the behavior, action, and image features of the first target individual is determined as first target individual feature data, wherein the first target individual includes an injured individual, a diseased individual, and an abnormal individual, and includes: Acquire multiple images of breeding individuals based on the image; In response to a selection instruction of a user, a target individual image is selected from the plurality of farmed individual images, and the farmed individual corresponding to the target individual image is determined as the first target individual; Performing feature extraction on the target individual image; Acquire a target historical image based on the historical data image of the first period, acquire target historical feature data based on the target historical image, and acquire a training set based on the first target individual feature data and the target historical feature data; Constructing a first neural network training model, importing the training set into the first neural network training model and training it, wherein the first neural network training model is used to obtain predicted individual characteristic data based on the input individual characteristic data; Importing the training set into the first neural network training model and performing iterative calculations, and obtaining the predicted individual feature data; Comparing the predicted individual characteristic data with the first target individual characteristic data, and acquiring error data; When the error data is less than the first preset data, the training process of the first neural network training model is stopped and the trained first neural network training model is obtained; a plurality of the individual feature data are obtained based on the historical data image of the second period, the individual feature data are input into the trained first neural network training model, and a plurality of the predicted individual feature data are obtained; The method compares the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifies a second target individual from the plurality of individual feature data according to the comparison result, and includes: Compare the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and determine that the predicted individual feature data whose difference value with the first target individual feature data is less than the second preset data is the second target individual feature data; Determining the individual corresponding to the second target individual characteristic data as the second target individual; According to the quantitative relationship between the second target individuals and the cultured individuals, the culture system is optimized, and the method includes: Acquire farm environmental data, including temperature data, humidity data, and breeding density data; Determine variable data of breeding individuals of the second target individual within the plurality of breeding cycles based on historical data of the quantitative relationship between the second target individual and the individual characteristic data in the plurality of breeding cycles; Constructing a second neural network training model, wherein the temperature data, the humidity data and the breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of the variable data, and the second neural network training model is used to iteratively calculate the temperature data, the humidity data, the breeding density data and the variable data, and output predicted variable data according to the input humidity data, the breeding density data and the variable data; Acquire multiple sets of preset temperature data, preset humidity data, and preset breeding density data, and input the preset temperature data, the preset humidity data, and the preset breeding density data into the second neural network training model; The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data.

2. The method for optimizing laying hen breeding system based on a large model of artificial intelligence according to claim 1, characterized in that: Determining optimal environmental data according to the variable data output by the second neural network training model, and performing environmental control on the farm according to the optimal environmental data, including: Determine optimal environmental data according to the variable data output by the second neural network training model, wherein the optimal environmental data is the environmental data in which the second target individuals account for the least proportion among the farmed individuals; Determine a set of the preset temperature data, the preset humidity data, and the preset breeding density data corresponding to the optimal environment data as target control data; The environment of the farm is regulated based on the target regulation data.

3. The method for optimizing laying hen breeding system based on a large model of artificial intelligence according to claim 2, characterized in that: A second neural network training model is constructed, wherein the temperature data, the humidity data and the breeding density data are used as three different dimensions, and the three dimensions constitute the characteristic vector of the variable data. The second neural network training model is used to iteratively calculate the temperature data, the humidity data, the breeding density data and the variable data, and output the predicted variable data according to the input humidity data, the breeding density data and the variable data, including: Comparing the predicted variable data output by the second neural network training model according to the historical environmental data with the variable data confirmed in a plurality of the cycles, and determining an error value; When the error data is less than the third preset data, the training process of the second neural network training model is stopped and the trained second neural network training model is obtained.

4. The method for optimizing laying hen breeding system based on a large model of artificial intelligence according to claim 3, characterized in that: Comparing the predicted variable data output by the second neural network training model according to the historical environmental data with the variable data confirmed in a plurality of the cycles and determining an error value, including: The relationship between the error value, the predicted variable data and the variable data satisfies: T = (y1-yt)y1 Among them, the error value is T, the value of the predicted variable data is yt, and the value of the variable data is y1.

5. An optimization device for laying hen breeding system based on a large model of artificial intelligence, characterized in that: The device is configured to: A first target individual is determined from a plurality of farmed individuals based on an image, and comprehensive individual data for characterizing the behavior, action, and image features of the first target individual is determined as first target individual feature data, wherein the first target individual includes an injured individual, a diseased individual, and an abnormal individual, and includes: Acquire multiple images of breeding individuals based on the image; In response to a selection instruction of a user, a target individual image is selected from the plurality of farmed individual images, and the farmed individual corresponding to the target individual image is determined as the first target individual; Performing feature extraction on the target individual image; Acquire a target historical image based on the historical data image of the first period, acquire target historical feature data based on the target historical image, and acquire a training set based on the first target individual feature data and the target historical feature data; Constructing a first neural network training model, importing the training set into the first neural network training model and training it, wherein the first neural network training model is used to obtain predicted individual characteristic data based on the input individual characteristic data; Importing the training set into the first neural network training model and performing iterative calculations, and obtaining the predicted individual feature data; Comparing the predicted individual characteristic data with the first target individual characteristic data, and acquiring error data; When the error data is less than the first preset data, the training process of the first neural network training model is stopped and the trained first neural network training model is obtained; a plurality of the individual feature data are obtained based on the historical data image of the second period, the individual feature data are input into the trained first neural network training model, and a plurality of the predicted individual feature data are obtained; The method compares the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and identifies a second target individual from the plurality of individual feature data according to the comparison result, and includes: Compare the plurality of predicted individual feature data obtained from the first neural network training model with the first target individual feature data, and determine that the predicted individual feature data whose difference value with the first target individual feature data is less than the second preset data is the second target individual feature data; Determining the individual corresponding to the second target individual characteristic data as the second target individual; According to the quantitative relationship between the second target individuals and the cultured individuals, the culture system is optimized, and the method includes: Acquire farm environmental data, including temperature data, humidity data, and breeding density data; Determine variable data of breeding individuals of the second target individual within the plurality of breeding cycles based on historical data of the quantitative relationship between the second target individual and the individual characteristic data in the plurality of breeding cycles; Constructing a second neural network training model, wherein the temperature data, the humidity data and the breeding density data are used as three different dimensions, and the three dimensions constitute a feature vector of the variable data, and the second neural network training model is used to iteratively calculate the temperature data, the humidity data, the breeding density data and the variable data, and output predicted variable data according to the input humidity data, the breeding density data and the variable data; Acquire multiple sets of preset temperature data, preset humidity data, and preset breeding density data, and input the preset temperature data, the preset humidity data, and the preset breeding density data into the second neural network training model; The optimal environmental data is determined according to the variable data output by the second neural network training model, and the environment of the farm is regulated according to the optimal environmental data.

6. An electronic device, characterized in that: include: at least one processor; And, a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by at least one of the processors, and the instructions are executed by at least one of the processors so that at least one of the processors can execute the method as set forth in any one of claims 1-4.

Images