Individual laying rate identification system for laying hen high-density breeding based on AI

Abstract

The invention discloses an individual laying rate recognition system for high-density breeding of laying hens based on AI in the technical field of laying hen breeding, and the system comprises a collection module which is used for collecting the number data of each hencoop in a farm, the egg production in each hencoop, and the appearance characteristics of each chicken in each hencoop; the analysis module is used for receiving an egg yield threshold value, set by a user, of a single coop, and marking the coop with the egg yield lower than the egg yield threshold value as a coop to be detected; the low-yield chicken recognition module is used for receiving set influence factors influencing the egg yield of the chickens, extracting real-time appearance characteristics of each chicken in the to-be-detected chicken coop, extracting the influence factors influencing the egg yield of the chickens from the appearance characteristics of each chicken in the to-be-detected chicken coop, and sending the extracted influence factors to the low-yield chicken recognition module; and outputting a low-yield chicken recognition result based on the number and degree of the extracted influence factors influencing the egg yield of the chickens. According to the invention, the low-yield chicken flocks are accurately identified through the acquisition module, the analysis module and the low-yield chicken identification module, so that traditional manual identification of the low-yield chicken flocks is replaced.

Description

Technical Field

[0001] The present invention belongs to the technical field of laying hen breeding, and specifically is an individual egg production rate identification system for high-density laying hen breeding based on AI. Background Art

[0002] Egg-laying chicken farming is an efficient, stable and sustainable agricultural industry. By providing egg-laying chickens with a suitable growth environment and nutritious feed, they can produce high-quality eggs regularly. Under suitable feeding conditions, egg-laying chickens can produce eggs continuously and stably, providing the market with a steady supply of eggs. Through scientific feeding management and breeding techniques, the egg production rate of egg-laying chickens will continue to increase, and the breeding benefits will be significant. In addition, egg-laying chicken farming can drive the development of related industries, such as feed processing, veterinary drug production, transportation, etc., forming a complete industrial chain and promoting the sustainable development of the local economy.

[0003] The egg production of laying hens is one of the main indicators to measure their production capacity. Laying hens with high egg production can bring higher economic benefits to farmers. The egg production of laying hens directly determines the economic benefits of farmers. The higher the egg production, the more income farmers will get. Therefore, increasing the egg production of laying hens is one of the important goals pursued by farmers.

[0004] At present, the existing methods to increase egg production of laying hens are mostly through optimizing feed formula and improving feeding environment. The identification of low-producing chicken flocks is mainly done through traditional manual methods. For low-producing chicken flocks, the accuracy of manual identification is low, resulting in the inability of the optimized feed formula and feeding environment to accurately target low-producing chicken flocks. Therefore, it is necessary to propose an AI-based individual egg production rate identification system for high-density breeding of laying hens. Summary of the invention

[0005] In order to solve the above problems, the purpose of the present invention is to provide an AI-based individual egg production rate identification system for high-density laying hen breeding, which accurately identifies low-producing chicken flocks through a collection module, an analysis module and a low-producing chicken identification module to replace the traditional manual identification of low-producing chicken flocks.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows: an AI-based individual egg production rate identification system for high-density breeding of laying hens, comprising:

[0007] The acquisition module is used to collect the number data of each chicken cage in the farm, the egg production in each chicken cage and the appearance characteristics of each chicken in the cage based on AI image recognition technology; and to collect environmental data in the farm through sensors.

[0008] The analysis module is used to receive the egg production threshold of a single chicken cage set by the user, and mark the chicken cages with egg production below the egg production threshold as chicken cages to be inspected.

[0009] The low-producing chicken identification module is used to receive the influencing factors that affect the egg production of chickens, extract the real-time appearance features of each chicken in the chicken cage to be inspected, extract the influencing factors that affect the egg production of chickens from the appearance features of each chicken in the chicken cage to be inspected, and output the low-producing chicken identification results based on the number and degree of the extracted influencing factors that affect the egg production of chickens.

[0010] The principle of the basic scheme is: the collection module collects the numbers of each chicken cage in the farm, the egg production in each chicken cage, the appearance characteristics of each chicken in the chicken cage, and the environmental data to provide a basis for subsequent data analysis. The analysis module will conduct a preliminary analysis of the data collected by the collection module, and compare the egg production information of each chicken cage with the preset egg production threshold. By comparing the number of eggs produced by different chicken cages, the analysis module marks the chicken cages with egg production lower than the egg production threshold as chicken cages to be inspected. In order to more accurately identify low-producing chickens, the low-producing chicken identification module compares the factors affecting the egg production of chickens with the factors affecting the egg production of chickens extracted from the appearance characteristics of each chicken in the chicken cage to be inspected. When the factors affecting the egg production of chickens extracted from the appearance characteristics of each chicken in the chicken cage to be inspected match the number and degree of the factors affecting the egg production of chickens preset, the output is identified as a low-producing chicken.

[0011] The beneficial effects of the basic solution are: 1. From the perspective of economic benefits, the use of the individual egg production rate identification system for high-density egg-laying chicken farming has significantly improved the egg production rate of the farm. According to actual cases, after the introduction of the individual egg production rate identification system for high-density egg-laying chicken farming, the accuracy rate of identifying low-producing chickens in a large farm was as high as 95%. Subsequently, through a targeted performance recovery plan, the overall egg production rate of the farm increased by about 20%. This improvement is directly converted into economic benefits. Calculated based on each chicken producing 10 more eggs per year and each egg sold for 0.5 yuan, the annual income of the farm can reach several million yuan.

[0012] 2. By analyzing the egg-laying and health status of each chicken, the system can guide the rational distribution of feed. Different nutrition strategies are adopted for high-producing chickens and low-producing chickens to ensure that each chicken can obtain the most suitable nutrition for its growth and egg-laying.

[0013] 3. By automatically identifying the cage number, egg production, chicken appearance and environmental data, the system can achieve accurate management of each chicken. The staff can quickly obtain the overall situation of the farm, identify problems in time and deal with them.

[0014] Furthermore, the acquisition module includes a binocular camera, the binocular camera signal is connected to a processing unit, the side wall of the breeding rack in the breeding farm is fixedly connected to a slide rail, the slide rail is slidably connected to a slide seat driven and controlled by the processing unit, and the binocular camera is fixedly connected to the slide seat.

[0015] The binocular camera is used to capture real-time images of the chicken cages on the opposite breeding racks, extract the number data of the chicken cages, obtain the location of each numbered chicken cage, and identify moving objects at the bottom of the chicken cages in real time. When a moving object appears at the bottom of the chicken cage, the moving object is tracked and the motion path of the moving object is obtained. Based on the motion path of the moving object, it is determined whether the object is an egg, the chicken cage to which the egg belongs, and the chicken to which the egg belongs.

[0016] The beneficial effect of the basic solution is that the binocular camera can slide in the chute to achieve all-round monitoring of every corner of the farm. This panoramic monitoring method helps to detect and deal with abnormal situations in a timely manner, and improve the overall management level of the farm. The binocular camera can dynamically identify moving objects at the bottom of the chicken cage, and accurately determine the egg production in each chicken cage based on the number of moving objects at the bottom of the chicken cage. The binocular camera has a distance measurement function, which helps to determine the specific location and number of the chicken cage by accurately measuring the distance between the binocular camera and the chicken cage.

[0017] Furthermore, the processing unit is used to obtain the serial number data of the chicken cage to be inspected, obtain the position of the chicken cage to be inspected based on the serial number data of the chicken cage to be inspected, control the slide to slide to the opposite side of the chicken cage to be inspected, and control the binocular camera to capture images of the chickens in the chicken cage to be inspected, and extract the influencing factors of each chicken in the chicken cage to be inspected that affect the egg production of the chicken.

[0018] The beneficial effect of the basic solution is: the position of the chicken cage to be inspected is obtained through the processing unit, and then the slide is controlled to slide to the opposite side of the chicken cage to be inspected, which improves the accuracy of the binocular camera in collecting images of the chickens in the chicken cage to be inspected, and is conducive to more comprehensive extraction of the influencing factors that affect the egg production of each chicken in the chicken cage to be inspected.

[0019] Furthermore, the low-yield chicken identification module is constructed based on a convolutional neural network. By inputting the factors that affect the egg production of chickens, it outputs whether the chicken is a low-yield chicken to train the low-yield chicken identification module.

[0020] The factors that affect the egg production of chickens include the appearance characteristics of different chickens and the environmental data in different farms; the appearance characteristics of chickens include the breed, age, body shape and feather characteristics of the chickens, and the environmental data in the farms include the temperature, humidity, light and air quality in the farms.

[0021] The beneficial effect of the basic scheme is that the convolutional neural network has a strong representation learning ability. By inputting the breed, age, body characteristics, feather characteristics of chickens that affect egg production, and the temperature, humidity, light and air quality information in the farm, it outputs whether the chicken is a low-producing chicken, thereby achieving high-precision identification.

[0022] Furthermore, the binocular camera is used to judge the moving object as an egg when the acquired motion path of the moving object contains a wavy motion trajectory; and obtain the area where the motion path of the moving object is located, judge the chicken cage to which the eggs belong according to the positions of the numbered chicken cages, and obtain the source direction of the moving object, and judge the chicken corresponding to the source direction as the chicken to which the eggs belong.

[0023] The basic solution has the following beneficial effects: by capturing and analyzing the motion path of moving objects, especially the wavy motion trajectory, the accuracy of egg recognition can be greatly improved. This specific motion pattern helps to distinguish eggs from other objects that may appear in the breeding environment. The binocular camera automatically tracks the source of the egg by accurately obtaining the motion path and source direction of the moving object, thereby determining which chicken laid the egg.

[0024] Furthermore, the body shape characteristics include the shin length and weight of laying hens. When the low-producing chicken identification module identifies that the shin length of the chickens in the cage to be inspected is less than 14 cm or the weight is less than 2 kg, they are identified as low-producing chickens.

[0025] The beneficial effect of the basic scheme is that the chicken's shank length and weight are related to its egg-laying performance. Shank length and weight are important indicators for measuring the growth and development of chickens, and these indicators are closely related to the egg-laying ability of chickens. By incorporating shank length and weight into the low-producing chicken identification module, the egg-laying potential of chickens can be more directly reflected, and the identification can be more targeted and accurate. By setting clear shank length and weight thresholds, it is possible to more objectively judge whether a chicken is a low-producing chicken, reduce interference from human factors, and improve the accuracy of identification.

[0026] Furthermore, feather features include feather density, feather color and feather shed amount. When the low-yield chicken identification module identifies that the feather density of the chicken in the chicken cage to be inspected is lower than the feather density threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the feather color vividness and neatness and brightness of the chicken in the chicken cage to be inspected are lower than the color threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the feather shed amount of the chicken in the chicken cage to be inspected is higher than the feather shed threshold, it is identified as a low-yield chicken.

[0027] The benefit of the basic scheme is that it can more accurately identify low-yielding chickens by taking into account three key features: feather density, color brightness and neatness, and feather shedding. This method is more reliable than single-feature identification and reduces the possibility of misjudgment.

[0028] Testing for feather density, feather color and feather shedding is a non-invasive method that does not cause any harm or stress to the birds. This helps maintain the health of the birds and their egg production while ensuring the safety of the testing process.

[0029] Furthermore, it also includes a data storage module, which is used to store the data generated by the collection module, the analysis module and the low-yield chicken identification module; the data generated by the collection module, the analysis module and the low-yield chicken identification module are used as historical data to optimize the training of the low-yield chicken identification module.

[0030] The beneficial effect of the basic solution is that the data storage module can ensure that all data generated from the collection to the judgment process are completely saved. This includes the chicken's body shape characteristics, feather characteristics, farm temperature, humidity and recognition results. This is conducive to subsequent staff to query and analyze the egg-laying situation of laying hens, as well as to optimize the training of the low-production chicken recognition module.

[0031] Furthermore, it also includes an alarm module, which is used to send out an alarm signal when the low-producing chicken identification module identifies a chicken as a low-producing chicken. The alarm signal includes the egg production of the low-producing chicken within a preset time period and the number data of the chicken cage where the low-producing chicken is located.

[0032] The beneficial effect of the basic solution is that the alarm module can be triggered after the low-yield chicken is identified, ensuring that the staff can immediately know the egg production of the low-yield chicken and the number of the chicken cage where the low-yield chicken is located. This helps the staff to quickly observe and inspect the low-yield chicken and implement special breeding plans for the low-yield chicken.

[0033] Furthermore, it also includes a scoring module, which is used to score the egg production rate of each chicken in the farm at a preset time interval, and output the egg production rate scoring result to the user's mobile terminal.

[0034] The beneficial effect of the basic solution is that the scoring module can score the egg production rate of each chicken based on the egg production data of the laying hens at preset time intervals. This helps staff to quickly identify high-producing and low-producing chickens and take corresponding management measures. Through the scoring module, staff can more reasonably allocate resources such as feed, water resources and breeding space. High-producing chickens can receive more attention and resource support, while low-producing laying hens can be specially bred according to actual conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a schematic diagram of the main structure of the individual egg production rate identification system for high-density breeding of laying hens based on AI in an embodiment of the present invention.

[0036] Figure 2 This is a schematic diagram of the data storage module structure of an individual egg production rate identification system for high-density breeding of laying hens based on AI in an embodiment of the present invention.

[0037] Figure 3 This is a schematic diagram of the entire process of the AI-based individual egg production rate identification system for high-density breeding of laying hens in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The following is further described in detail through specific implementation methods:

[0039] Embodiment 1:

[0040] Basically as attached Figure 1 -Attached Figure 3 As shown: An AI-based individual egg production rate recognition system for high-density laying hens, including a collection module for collecting the number data of each chicken cage in the farm, the egg production in each chicken cage and the appearance characteristics of each chicken in the cage based on AI image recognition technology; and collecting environmental data in the farm through sensors; the collection module includes a binocular camera.

[0041] The binocular camera signal is connected to a processing unit, the side wall of the breeding rack in the breeding farm is fixedly connected to a slide rail by bolts, a slide seat driven by the processing unit is slidably connected in the slide rail, and the binocular camera is fixedly connected to the slide seat.

[0042] The binocular camera is used to capture real-time images of the chicken cages on the opposite breeding racks, extract the number data of the chicken cages, obtain the location of each numbered chicken cage, and identify the moving objects at the bottom of the chicken cages in real time. When a moving object appears at the bottom of the chicken cage, the moving object is tracked and the motion path of the moving object is obtained. Based on the motion path of the moving object, the eggs, the chicken cage to which the eggs belong, and the chicken to which the eggs belong are identified.

[0043] The binocular camera is used to judge the moving object as an egg when the motion path of the moving object contains a wavy motion trajectory; and obtain the area where the motion path of the moving object is located, judge the chicken cage to which the eggs belong according to the positions of the numbered chicken cages, and obtain the source direction of the moving object, and judge the chicken corresponding to the source direction as the chicken to which the eggs belong.

[0044] The analysis module is used to receive the egg production threshold of a single chicken cage set by the user, and mark the chicken cages with egg production below the egg production threshold as chicken cages to be inspected.

[0045] The specific implementation process is as follows: first, the binocular camera is fixed to the slide by bolts, and the slide is driven by the processing unit to slide in the slide rail to monitor each chicken cage in the entire farm.

[0046] Because after laying hens in the chicken cage lay eggs, the eggs will slide to the lower layer of the chicken cage in a wave-like motion trajectory. The binocular camera has two cameras, and the binocular camera usually sets the distance between the two cameras to a certain value, which is also called the baseline of the camera. Each camera will capture an image from a different angle in the scene. By processing these two images, a "stereoscopic picture", that is, a three-dimensional depth image, can be obtained. The binocular camera can accurately determine the distance between the egg and the binocular camera when the egg slides through the three-dimensional depth image, so as to accurately determine the direction of the egg's source, the laying hen and the number of the chicken cage. In addition, the binocular camera uses dynamic recognition technology to accurately determine the egg production of each chicken in each chicken cage.

[0047] When the binocular camera and sensor collect information in the farm, the collection module will receive the data transmitted by the binocular camera and sensor.

[0048] The analysis module will compare the egg production of a single cage collected by the collection module with the egg production threshold set by the user, and mark the cages with egg production below the egg production threshold as cages to be inspected.

[0049] Embodiment 2:

[0050] The difference from the above embodiment is that the binocular camera signal is connected to a processing unit, and the processing unit is used to obtain the number data of the chicken cage to be inspected, obtain the position of the chicken cage to be inspected based on the number data of the chicken cage to be inspected, and control the slide to slide to the opposite side of the chicken cage to be inspected, and control the binocular camera to capture images of the chickens in the chicken cage to be inspected, and extract the influencing factors of each chicken in the chicken cage to be inspected that affect the egg production of the chicken.

[0051] Also includes a module for identifying low-producing chickens.

[0052] The low-producing chicken identification module is used to receive the influencing factors that affect the egg production of chickens, extract the real-time appearance features of each chicken in the chicken cage to be inspected, extract the influencing factors that affect the egg production of chickens from the appearance features of each chicken in the chicken cage to be inspected, and output the low-producing chicken identification results based on the number and degree of the extracted influencing factors that affect the egg production of chickens.

[0053] The low-yield chicken identification module is built based on a convolutional neural network. By inputting the factors that affect the egg production of chickens, it outputs whether the chicken is a low-yield chicken to train the low-yield chicken identification module.

[0054] The factors that affect the egg production of chickens include the appearance characteristics of different chickens and the environmental data in different farms; the appearance characteristics of chickens include the breed, age, body shape and feather characteristics of the chickens, and the environmental data in the farms include the temperature, humidity, light and air quality in the farms.

[0055] Body shape features include the shin length and weight of laying hens. When the low-yield chicken identification module identifies that the shin length of the chicken in the cage to be inspected is less than 14cm or the weight is less than 2kg, it is identified as a low-yield chicken. Feather features include feather density, feather color and feather shedding amount. When the low-yield chicken identification module identifies that the feather density of the chicken in the cage to be inspected is lower than the feather density threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the feather color brightness and neatness and brightness of the chicken in the cage to be inspected are lower than the color threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the feather shedding amount of the chicken in the cage to be inspected is higher than the feather shedding threshold, it is identified as a low-yield chicken.

[0056] It also includes a data storage module, which is used to store the data generated by the acquisition module, the analysis module and the low-yield chicken identification module; the data generated by the acquisition module, the analysis module and the low-yield chicken identification module are used as historical data to optimize the training of the low-yield chicken identification module. It also includes an alarm module, which is used to send an alarm signal when the low-yield chicken identification module identifies a chicken as a low-yield chicken. The alarm signal includes the egg production of the low-yield chicken within a preset time period and the number data of the chicken cage where the low-yield chicken is located. It also includes a scoring module, which is used to score the egg production rate of each chicken in the farm according to a preset time interval, and output the egg production rate score result to the user's mobile terminal.

[0057] The specific implementation process is as follows:

[0058] The low-yield chicken identification module is built based on a convolutional neural network. The chicken breed, age, appearance characteristics, and temperature, humidity, light, and air quality of the farm are input into the low-yield chicken identification module for training, and the low-yield chicken identification results are output.

[0059] After analyzing the chicken cage with low egg production, the binocular camera will drive the slider to move to the opposite side of the chicken cage under the control of the processing unit to take detailed photos and analyze each chicken in the cage.

[0060] Finally, the low-laying chicken identification module is used to analyze the chickens in the cages with low egg production. When the low-laying chicken identification module identifies that the laying hen has a shank length of less than 14 cm or a weight of less than 2 kg, it is identified as a low-laying chicken.

[0061] When the low-yield chicken identification module identifies that the feather density of the chicken in the cage to be inspected is lower than the feather density threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the color vividness and neatness and brightness of the feathers of the chicken in the cage to be inspected are lower than the color threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the amount of feather shedding of the chicken in the cage to be inspected is higher than the feather shedding threshold, it is identified as a low-yield chicken.

[0062] When the low-yield chicken identification module identifies a chicken as a low-yield chicken, the alarm module will quickly alert the staff, including the egg production of the low-yield chicken in the preset time period and the number of the chicken cage where the low-yield chicken is located. For example, when the low-yield chicken identification module identifies a chicken in cage B as a low-yield chicken, the alarm module triggers an alarm and sends an alarm message to the staff: "There is a low-yield chicken in cage B, with an egg production of 10. Please handle it in time."

[0063] The scoring module will use a binocular camera to score the chickens in the cages according to the number of eggs they lay at preset time intervals, and score them from high to low based on the amount of eggs they lay. This will help staff better understand the egg production rate of laying hens in the farm, so as to improve the feed and optimize the breeding environment in a targeted manner.

[0064] The data storage module will continuously store the data generated by the collection module, analysis module and low-yield chicken identification module, and use the data generated by the collection module, analysis module and low-yield chicken identification module as historical data to optimize the training of the low-yield chicken identification module, and when the staff needs to retrieve the data, it can quickly help the staff to find the data.

[0065] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0066] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. An AI-based individual egg production rate identification system for high-density breeding of laying hens, characterized in that: include: The acquisition module is used to collect the number data of each chicken cage in the farm, the egg production in each chicken cage, and the appearance characteristics of each chicken in the cage based on AI image recognition technology; And collect environmental data in the farm through sensors; An analysis module is used to receive an egg production threshold value of a single chicken cage set by a user, and mark a chicken cage with an egg production lower than the egg production threshold value as a chicken cage to be inspected; The low-producing chicken identification module is used to receive the influencing factors that affect the egg production of chickens, extract the real-time appearance features of each chicken in the chicken cage to be inspected, extract the influencing factors that affect the egg production of chickens from the appearance features of each chicken in the chicken cage to be inspected, and output the low-producing chicken identification results based on the number and degree of the extracted influencing factors that affect the egg production of chickens.

2. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 1 is characterized by: The acquisition module includes a binocular camera, the binocular camera signal is connected to a processing unit, the side wall of the breeding rack in the breeding farm is fixedly connected to a slide rail, a slide seat driven by the processing unit is slidably connected in the slide rail, and the binocular camera is fixedly connected to the slide seat; The binocular camera is used to capture real-time images of the chicken cages on the opposite breeding racks, extract the number data of the chicken cages, obtain the location of each numbered chicken cage, and identify moving objects at the bottom of the chicken cages in real time. When a moving object appears at the bottom of the chicken cage, the moving object is tracked and the motion path of the moving object is obtained. Based on the motion path of the moving object, it is determined whether the object is an egg, the chicken cage to which the egg belongs, and the chicken to which the egg belongs.

3. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 2 is characterized by: The processing unit is used to obtain the serial number data of the chicken cage to be inspected, obtain the position of the chicken cage to be inspected based on the serial number data of the chicken cage to be inspected, control the slide to slide to the opposite side of the chicken cage to be inspected, and control the binocular camera to collect images of the chickens in the chicken cage to be inspected, and extract the influencing factors that affect the egg production of each chicken in the chicken cage to be inspected.

4. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 1 is characterized by: The low-yield chicken identification module is built based on a convolutional neural network. By inputting the factors that affect the egg production of chickens, it outputs whether the chicken is a low-yield chicken to train the low-yield chicken identification module. The factors that affect the egg production of chickens include the appearance characteristics of different chickens and the environmental data in different farms; the appearance characteristics of chickens include the breed, age, body shape and feather characteristics of the chickens, and the environmental data in the farms include the temperature, humidity, light and air quality in the farms.

5. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 3 is characterized by: The binocular camera is used to judge the moving object as an egg when the motion path of the moving object contains a wavy motion trajectory; and obtain the area where the motion path of the moving object is located, judge the chicken cage to which the eggs belong according to the positions of the numbered chicken cages, and obtain the source direction of the moving object, and judge the chicken corresponding to the source direction as the chicken to which the eggs belong.

6. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 4 is characterized by: The physical characteristics include the shin length and weight of laying hens. When the low-yield chicken identification module identifies that the shin length of the chickens in the cage to be inspected is less than 14cm or the weight is less than 2kg, they are identified as low-yield chickens.

7. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 6 is characterized by: Feather features include feather density, feather color and feather shedding amount. When the low-yield chicken identification module identifies that the feather density of the chicken in the cage to be inspected is lower than the feather density threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the feather color vividness and neatness and brightness of the chicken in the cage to be inspected are lower than the color threshold, it is identified as a low-yield chicken. When the low-yield chicken identification module identifies that the feather shedding amount of the chicken in the cage to be inspected is higher than the feather shedding threshold, it is identified as a low-yield chicken.

8. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 7 is characterized by: It also includes a data storage module, which is used to store the data generated by the acquisition module, the analysis module and the low-yield chicken identification module; the data generated by the acquisition module, the analysis module and the low-yield chicken identification module are used as historical data to optimize the training of the low-yield chicken identification module.

9. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 8, characterized in that: It also includes an alarm module, which is used to send out an alarm signal when the low-production chicken identification module identifies a chicken as a low-production chicken. The alarm signal includes the egg production of the low-production chicken within a preset time period and the number data of the chicken cage where the low-production chicken is located.

10. The AI-based individual egg production rate identification system for high-density breeding of laying hens according to claim 9, characterized in that: It also includes a scoring module, which is used to score the egg production rate of each chicken in the farm at preset time intervals, and output the egg production rate scoring result to the user's mobile terminal.

Images