Intelligent management and control system and method for animal living place

Abstract

The invention discloses an intelligent management and control system and method for an animal living place, relates to the technical field of breeding management, and aims to solve the problem of poor monitoring effect on an animal growth process in the prior art. The system at least comprises a multi-source sensing module, a data transmission module, a multi-dimensional decision control module, a plurality of execution devices and a man-machine interaction device. The data transmission module is used for acquiring multi-source sensing data of a target animal living place and performing data preprocessing on the multi-source sensing data to obtain target multi-source sensing data; the multi-dimensional decision control module is used for generating a target control instruction for performing multi-dimensional control on the target animal living field based on the target multi-source sensing data; the execution devices are used for executing corresponding operations according to the target control instruction; the man-machine interaction device is used for visually displaying a target animal living place, generating prompt information based on an early warning mechanism and setting system parameters; and refined and automatic management of the target animal farm is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of breeding management, and in particular to an intelligent management and control system and method for an animal living field. BACKGROUND

[0002] As a product of the deep integration of modern agriculture and information technology, intelligent breeding systems are redefining traditional breeding models through technologies such as the Internet of Things, big data, and artificial intelligence, and are showing significant advantages in efficiency, management optimization, and sustainable development.

[0003] However, existing intelligent breeding systems still have many problems. Early health risks lack continuous monitoring and abnormal early warning of real-time physiological data. This mainly manifests in the following four aspects: First, existing intelligent breeding systems mainly focus on a single link, such as automatic adjustment of environmental temperature and humidity or timed feeding of feed, and lack multi-dimensional data linkage analysis of environment, physiology, and reproduction. Second, intelligent breeding systems rely on fixed threshold values to trigger device control (such as temperature > 28℃ to start the cooling device), without considering individual differences in animal breed, age, and health status; for example, in the process of intelligent breeding of rabbits, the optimal growth temperature range of different breeds of rabbits differs by 2-3℃. Third, in the aspect of reproduction management, intelligent breeding systems still mainly rely on manual observation; for example, estrus recognition mainly relies on observation by breeding personnel, with an accuracy rate of only 60%-70%, and the effective observation time is mainly concentrated in the morning and evening, which easily misses the best window, resulting in low reproduction efficiency. Fourth, health monitoring is lagging, and disease diagnosis relies on manual inspection, such as diarrhea or significant body temperature rise, at which point the animal's condition has developed to the middle or late stage, significantly increasing the mortality rate.

[0004] Therefore, there is an urgent need for a more effective animal living field monitoring solution to address the poor monitoring effect on animal growth in the prior art. SUMMARY

[0005] The present application aims to provide an intelligent management and control system and method for an animal living field, covering four core dimensions of "environment-feeding-health-reproduction", and achieving fine and automated management of the animal living field through multi-source data fusion and intelligent decision-making, thereby improving animal growth efficiency, reproduction efficiency, and reproduction quality, and solving the problem of poor monitoring effect on animal growth in the prior art.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions: In a first aspect, the present application provides an intelligent management and control system for an animal living field, which can at least include: a multi-source perception module, a data transmission module, a multi-dimensional decision-making control module, a plurality of execution devices, and a human-computer interaction device; The multi-source sensing module is connected to the data transmission module, the data transmission module is connected to the multi-dimensional decision control module, the multi-dimensional decision control module is connected to multiple execution devices, and the human-computer interaction device is connected to the multi-dimensional decision control module, the data transmission module, and multiple execution devices respectively. The data transmission module is used to acquire multi-source sensory data of the target animal's habitat and to preprocess the multi-source sensory data to obtain target multi-source sensory data; the multi-source sensory data includes environmental sensory data, physiological sensory data, and reproductive sensory data; The multi-dimensional decision control module is used to generate target control instructions for multi-dimensional management and control of the target animal's habitat based on the target multi-source perception data and according to task scheduling priority; the multi-dimensional management and control includes environmental adaptive management and control, precise feeding decision-making, health risk assessment and reproductive optimization; Multiple execution devices are used to perform corresponding operations according to the target control instructions issued by the multi-dimensional decision control module; The human-computer interaction device is used to visualize the target animal's habitat based on the target multi-source perception data, generate prompts based on the early warning mechanism, and set system parameters.

[0007] Preferably, the multi-source sensing module may include at least an environmental sensing submodule, a physiological sensing submodule, and a reproductive sensing submodule; The environmental sensing submodule is used to perform spatial grid-based monitoring of environmental parameters of animal habitats, based at least on distributed temperature and humidity sensors, light sensors, and ammonia or hydrogen sulfide composite sensors. The physiological sensing submodule is used to perform physiological monitoring of the target animal based on at least an embedded weight sensor, a feed trough pressure sensor, an acceleration sensor, and an infrared body temperature array. The reproductive sensing submodule is used to monitor the reproduction of target animals based at least on a visual camera, a pH sensor, and a sound acquisition microphone.

[0008] Preferably, the data transmission module may include a first network submodule and a second network submodule; The first network submodule is used to collect multi-source sensing data in the target animal's habitat and preprocess the collected multi-source sensing data using the 3σ principle to obtain the target multi-source sensing data. The second network submodule is used to remotely transmit the target multi-source sensing data, and to cache the target multi-source sensing data when the network is disconnected, and to automatically resend it after the network is connected.

[0009] Preferably, the multi-dimensional decision control module may include at least an environmental adaptive regulation submodule, a precision feeding decision submodule, a health risk assessment submodule, and a reproductive optimization submodule; The environmental adaptive control submodule is used to generate environmental control instructions based on environmental perception data and the animal's growth stage, using the DDPG algorithm to dynamically adjust the ventilation, humidity, and temperature in the target animal's living area; the growth stage includes at least the juvenile stage, adult stage, and reproductive stage. The precision feeding decision submodule is used to predict the target feed intake of animals at different growth stages and generate feeding control instructions based on physiological perception data and target matrix, using a random forest decision mechanism to carry out personalized feeding. The health risk assessment submodule is used to generate health regulation instructions based on physiological perception data, and, in conjunction with isolation pens and medication equipment, to monitor diseases in animals at different growth stages. The reproductive optimization submodule is used to generate reproductive control instructions, including the target mating time period, based on reproductive perception data and a preset neural network model; the preset neural network model is a network model obtained after model training based on multimodal animal data.

[0010] Preferably, the multi-dimensional decision control module may further include a task priority scheduling submodule; The task priority scheduling submodule is used to regulate the tasks corresponding to the regulation instructions generated by the environmental adaptive regulation submodule, the precision feeding decision submodule, the health risk assessment submodule, and the reproduction optimization submodule according to the task scheduling priority, which decreases in the order of health regulation, reproduction regulation, environmental regulation, and feeding regulation.

[0011] Preferably, the human-computer interaction device may include at least a visual management platform, an intelligent early warning subsystem, and a strategy configuration subsystem; The visualization management platform is at least used to display in real time, in a three-dimensional digital interface, the environmental heat map of the target animal's habitat, cage status, abnormal data, growth curves, health reports, and reproductive statistics. The intelligent early warning subsystem is at least used to issue early warnings for abnormal detection data in the target animal's habitat based on the early warning mechanism of the target level; The strategy configuration subsystem is used at least to set environmental thresholds, feeding schemes, and system parameters.

[0012] Secondly, the present invention provides an intelligent management and control method for animal habitats, wherein the intelligent management and control method for animal habitats is applied to the intelligent management and control system for animal habitats described in the first aspect, and the method may include: Acquire multi-source sensory data of the target animal's habitat; the multi-source sensory data includes environmental sensory data, physiological sensory data, and reproductive sensory data; The multi-source sensing data is preprocessed to obtain the target multi-source sensing data; Based on the target multi-source sensing data, target control instructions are generated according to task scheduling priority to manage the target animal's habitat in multiple dimensions; the multi-dimensional management includes environmental adaptive management, precise feeding decision-making, health risk assessment, and reproductive optimization. Based on the target control command, multiple execution devices are controlled to perform corresponding operations.

[0013] Preferably, acquiring multi-source sensory data of the target animal's habitat may include: When the animals in the target animal habitat are in a non-breeding period, data is collected using a multi-source sensing module based on a first data collection frequency to obtain the multi-source sensing data. When the animals in the target animal habitat are in their breeding season, data is collected using a multi-source sensing module based on a second data acquisition frequency to obtain the multi-source sensing data; the first data acquisition frequency is less than the second data acquisition frequency.

[0014] Preferably, the step of generating target control instructions for multi-dimensional management and control of the target animal's habitat based on the target multi-source sensing data and according to task scheduling priority may include: Based on the target multi-source sensing data, health regulation instructions, reproductive regulation instructions, environmental regulation instructions, and feeding regulation instructions for the target animal's habitat are generated respectively. The target control instructions are obtained by sorting the health control instructions, reproductive control instructions, environmental control instructions, and feeding control instructions according to the task scheduling priority in descending order of health control, reproductive control, environmental control, and feeding control. The generation of health regulation commands, reproductive regulation commands, environmental regulation commands, and feeding regulation commands for the target animal's habitat based on multi-source sensing data may include: Based on environmental perception data and combined with the animal's growth stage, the DDPG algorithm is used to generate the environmental control instructions, which include at least dynamically adjusting the ventilation, humidity and temperature in the target animal's living area. Based on physiological perception data and target matrix, a random forest decision-making mechanism is used to predict the target feed intake of animals at different growth stages and generate the feeding control instructions. The health regulation instructions are generated based on physiological perception data; Based on reproductive perception data, a preset neural network model is used to generate reproductive regulation instructions, including the target mating period; the preset neural network model is a network model obtained after model training based on multimodal animal data.

[0015] Preferably, the method may further include: Acquire target feedback data of the target animal habitat within a target time period; the target feedback data includes at least the mating success rate and feed conversion rate. The system parameters are updated based on the target feedback data, and an incremental learning algorithm is used to adapt to the changes in the growth cycle of the target animal, forming a closed loop of collection, decision-making, execution and feedback.

[0016] Compared with existing technologies, this invention provides an intelligent management and control system for animal habitats. This system comprises a multi-source sensing module, a data transmission module, a multi-dimensional decision control module, multiple execution devices, and a human-computer interaction device. The multi-source sensing module is connected to the data transmission module, which in turn is connected to the multi-dimensional decision control module. The multi-dimensional decision control module is connected to the multiple execution devices, and the human-computer interaction device is connected to each of these modules. This allows the system to acquire multi-source sensing data from the target animal habitat using the data transmission module, perform data preprocessing to obtain target multi-source sensing data, and then use the multi-dimensional decision control module to generate target control commands for multi-dimensional management and control of the target animal habitat based on the target multi-source sensing data and according to task scheduling priorities. In this system, multi-source sensing data includes environmental sensing data, physiological sensing data, and reproductive sensing data; multi-dimensional management includes environmental adaptive management, precise feeding decisions, health risk assessment, and reproductive optimization; furthermore, multiple execution devices can be used to perform corresponding operations based on target control commands issued by the multi-dimensional decision control module; simultaneously, human-computer interaction devices are used to visualize the target animal habitat based on the target multi-source sensing data, generate prompts based on the early warning mechanism, and set system parameters; based on this, it covers the four core dimensions of the target animal habitat: "environment-feeding-health-reproduction". Through the fusion of multi-dimensional data and intelligent decision-making, it achieves refined and automated management of the target animal habitat, improves the reproductive efficiency and quality of the target animals, and solves the problem of poor monitoring of animal growth processes in existing technologies. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1A schematic diagram of the main system architecture of an intelligent management and control system for an animal habitat provided by the present invention; Figure 2 This is a schematic diagram of the main process of an intelligent management and control method for animal habitats provided by the present invention.

[0018] Figure reference numerals: 100-Multi-source sensing module, 110-Environmental sensing sub-module, 120-Physiological sensing sub-module, 130-Reproduction sensing sub-module, 200-Data transmission module, 210-First network sub-module, 220-Second network sub-module, 300-Human-computer interaction device, 310-Visual management platform, 320-Intelligent early warning subsystem, 330-Strategy configuration subsystem, 400-Multi-dimensional decision control module, 410-Environmental adaptive regulation sub-module, 420-Precision feeding decision sub-module, 430-Health risk assessment sub-module, 440-Reproduction optimization sub-module, 450-Task priority scheduling sub-module, 500-Multiple execution devices, 510-Environmental regulation device, 520-Feeding device, 530-Health device, 540-Reproduction auxiliary device. Detailed Implementation

[0019] To facilitate a clear description of the technical solutions in the embodiments of the present invention, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect. For example, the first threshold and the second threshold are merely used to distinguish different thresholds and do not limit their order. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0020] It should be noted that in this invention, the terms "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0021] In this invention, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding related objects have an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, a combination of a and b, a combination of a and c, a combination of b and c, or a, b, and c, where a, b, and c can be single or multiple.

[0022] In recent years, researchers have explored the use of various wearable sensors and cameras to collect animal data (color images, infrared images, 3D images, temporal motion data, etc.) to extract abnormal physiological and behavioral characteristics of animal health, achieving some progress. For example, researchers have used YOLOv4 networks to recognize pig drinking behavior in images from different perspectives; researchers have proposed a deep video recognition algorithm based on a Hidden Markov Model to address the difficulty in recognizing high-risk actions in lactating sows; researchers have proposed an improved Ostu algorithm based on thermal infrared images to detect the ear root region of pigs; and researchers have borrowed facial landmark detection technology to automatically identify key parts of pig faces in infrared images. However, the above research only obtains a single physiological or behavioral characteristic of pigs, focusing on a single aspect, and does not consider the overall health status of pigs. Furthermore, the susceptibility of fixed equipment to damage and falling makes it not significantly more cost-effective than manual labor, hindering its widespread application. Moreover, there are no more advanced technical solutions for environmental monitoring, reproductive monitoring, and health monitoring, and no solution has yet emerged that can comprehensively consider these aspects to improve the efficiency of animal habitat monitoring.

[0023] In view of this, the present invention provides an intelligent management and control system and method for animal habitats, covering four core dimensions: environment, feeding, health and reproduction. It effectively integrates environmental monitoring, feeding monitoring, reproduction monitoring and health monitoring, realizes refined and automated management of animal habitats, and improves animal growth efficiency, reproduction efficiency and reproduction quality.

[0024] The technical solution of the present invention will now be described in detail with reference to the accompanying drawings: Firstly, this invention provides an intelligent management and control system for animal habitats; please refer to [reference needed]. Figure 1 , Figure 1 This is a schematic diagram of the main system architecture of an intelligent management and control system for animal habitats provided by the present invention.

[0025] exist Figure 1 In this system, the intelligent control system may include at least: a multi-source sensing module 100, a data transmission module 200, a multi-dimensional decision control module 400, multiple execution devices 500, and a human-machine interaction device 300.

[0026] Specifically, the multi-source sensing module 100 can be connected to the data transmission module 200, the data transmission module 200 can be connected to the multi-dimensional decision control module 400, the multi-dimensional decision control module 400 can be connected to multiple execution devices 500, and the human-computer interaction device 300 can be connected to the multi-dimensional decision control module 400, the data transmission module 200, and the multiple execution devices 500 respectively. Thus, the data transmission module 200 can be used to acquire multi-source sensing data of the target animal's habitat, and the multi-source sensing data can be preprocessed to obtain target multi-source sensing data. The multi-source sensing data can include environmental sensing data, physiological sensing data, and reproductive sensing data. The system utilizes a multi-dimensional decision control module 400 to generate target control instructions for multi-dimensional management of the target animal habitat based on multi-source sensing data and according to task scheduling priorities. These multi-dimensional management instructions include environmental adaptive management, precise feeding decisions, health risk assessment, and reproductive optimization. Furthermore, multiple execution devices 500 can execute corresponding operations based on the target control instructions issued by the multi-dimensional decision control module 400. Simultaneously, a human-computer interaction device 300 is used to visualize the target animal habitat based on the multi-source sensing data, generate alerts based on an early warning mechanism, and set system parameters.

[0027] Based on this, the intelligent management and control system for animal habitats provided by this invention can cover the four core dimensions of the target animal habitat: "environment, feeding, health, and reproduction". Through the fusion of multi-dimensional data and intelligent decision-making, it realizes the refined and automated management of the target animal habitat, improves the reproductive efficiency and quality of the target animals, and solves the problem of poor monitoring of the animal growth process in the prior art.

[0028] As an alternative embodiment, please continue to refer to... Figure 1 The multi-source sensing module 100 may include at least an environmental sensing submodule 110, a physiological sensing submodule 120, and a reproductive sensing submodule 130.

[0029] Specifically, the environmental sensing submodule 110 can be used to conduct spatial gridded monitoring of environmental parameters in animal habitats, based at least on distributed temperature and humidity sensors, light sensors, and ammonia or hydrogen sulfide composite sensors.

[0030] And by using the physiological sensing submodule 120, based at least on an embedded weight sensor, a feed trough pressure sensor, an acceleration sensor and an infrared body temperature array, to perform physiological monitoring on the target animal.

[0031] And the reproductive sensing submodule 130 is used to monitor the reproduction of the target animal based at least on a visual camera, a pH sensor and a sound acquisition microphone.

[0032] As a specific embodiment, the construction of the multi-source sensing module 100 provided by the present invention will be illustrated by taking an animal rabbit habitat as an example.

[0033] 1. Regarding the environmental sensing submodule 110: Distributed deployment of temperature and humidity sensors, light sensors, and ammonia / hydrogen sulfide composite sensors can be used for environmental sensing, thereby achieving spatial grid-based monitoring of rabbitry environmental parameters. The optimal arrangement of environmental sensing nodes is 3-5 nodes per 50㎡. Specifically, the sensor accuracy is preferably ±0.5℃, ±3% RH; the light sensor's sensitivity range is preferably 0-10000 lux; and the composite sensor's resolution is preferably 1ppm to ensure the accuracy of the environmental sensing data.

[0034] 2. Regarding the physiological sensing submodule 120: An embedded weight sensor (accuracy ±10g) can be installed in the rabbit cage, and a pressure sensor can be installed in the feed trough to accurately record the rabbit's weight changes and food intake; a triaxial accelerometer can also be installed to collect the rabbit's behavioral data, such as movement trajectory and resting time; an infrared body temperature array sensor can also be installed to scan the body temperature of the entire group or individuals at fixed or variable intervals; thus obtaining more accurate animal physiological sensing data. Specifically, the accuracy of the weight sensor is preferably ±10g, the resolution of the pressure sensor is preferably ±5g, the battery life of the triaxial accelerometer is preferably not less than 12 months, the accuracy of the infrared body temperature array sensor is preferably ±0.3℃, and the fixed time interval is preferably 2 hours.

[0035] 3. Regarding the reproductive perception submodule 130: A binocular vision camera can be deployed to capture estrus-related behaviors such as chasing and mounting in real time using an estrus behavior recognition algorithm. In practical applications, the recognition accuracy is greater than 92%. Simultaneously, a secretion pH sensor and a sound acquisition microphone can be deployed to dynamically detect rabbit secretions to obtain physiological indicators, and to detect and identify the rabbit's unique calls during estrus. Multimodal data fusion of reproductive perception data can further improve the accuracy of estrus monitoring, achieving an accuracy greater than 95% in practical applications. Specifically, the frame rate of the binocular vision camera is preferably 25fps; the pH sensor is preferably a disposable sterile probe with a measurement range of 3.5-8.0 and an accuracy of ±0.1pH; and the frequency band of the sound acquisition microphone is preferably 2-5kHz. Based on this, more accurate animal reproductive perception data can be obtained.

[0036] It should be noted that, in order to more clearly illustrate the technical solution of the present invention, a rabbit is used as an example for animal monitoring in the specific embodiments. It is understood that the system and method of the present invention can also monitor other mammals, such as pigs and sheep.

[0037] As an optional embodiment, the data transmission module 200 may include a first network submodule 210 and a second network submodule 220; the first network submodule 210 is used to collect multi-source sensing data in the target animal's habitat and preprocess the collected multi-source sensing data using the 3σ principle to obtain target multi-source sensing data; the second network submodule 220 is used to remotely transmit the target multi-source sensing data and cache the target multi-source sensing data when the network is disconnected, and automatically resend it after the network is connected.

[0038] Specifically, taking a rabbit hutch as an example, the first network submodule 210 can preferably use a LoRaMesh network to collect multi-source sensing data within the target animal's living area, achieving low-power data aggregation and solving the signal shielding problem caused by metal cages. Simultaneously, the first network submodule 210 can use the 3σ principle to remove outliers from the multi-source data, obtaining the target multi-source sensing data. Then, the second network submodule 220 is used for remote data transmission. For example, the target multi-source sensing data can be normalized and then encrypted and uploaded to the cloud via a 5G network. The LoRaMesh network has a single gateway coverage of 200m and supports 200+ nodes. The second network submodule 220 preferably uses an NB-IoT module, which supports seamless mobile network access. When the network is down, the edge gateway caches data for ≥24 hours and automatically retransmits it after reconnection, with data transmission reliability ≥99.9%.

[0039] This can greatly improve the accuracy of data transmission.

[0040] As an optional embodiment, the multi-dimensional decision control module 400 may include at least an environmental adaptive regulation submodule 410, a precision feeding decision submodule 420, a health risk assessment submodule 430, and a reproductive optimization submodule 440. The environmental adaptive regulation submodule 410 is used to generate environmental regulation instructions based on environmental perception data and the animal's growth stage, using the DDPG algorithm to dynamically adjust the ventilation, humidity, and temperature in the target animal's living area. The growth stage includes at least the juvenile stage, adult stage, and reproductive stage. The precision feeding decision submodule 420 is used to predict the target feed intake of animals at different growth stages and generate feeding regulation instructions based on physiological perception data and a target matrix, using a random forest decision mechanism for personalized feeding. The health risk assessment submodule 430 is used to generate health regulation instructions based on physiological perception data, and, in conjunction with isolation pens and medication equipment, monitor diseases in animals at different growth stages. The reproductive optimization submodule 440 is used to generate reproductive regulation instructions, including the target mating period, using a preset neural network model based on reproductive perception data; the preset neural network model is a network model obtained after training on multimodal animal data.

[0041] Specifically, let's continue with the example of a rabbit hutch: In terms of environmental adaptive control, parameters such as the rabbit's growth stage (young rabbit / adult rabbit / breeding rabbit) and external weather can be preset. Then, by using environmental sensing data and combining the animal's growth stage, the DDPG algorithm can be used to generate environmental control instructions and dynamically output the optimal combination of ventilation volume, temperature setpoint, and light duration.

[0042] In terms of precision feeding decisions: a target matrix of "breed-age-weight-feed intake" can be pre-constructed. Combining the rabbit's physiological perception data with the target matrix, a random forest decision-making mechanism can be used to predict the target feed intake of animals at different growth stages and generate feeding control instructions for personalized feeding. For example, the feed intake of rabbits the next day can be predicted, and the feed intake of underweight individuals can be automatically increased by 10%-15%.

[0043] In terms of health risk assessment: Through time-series correlation analysis of body temperature, activity level, and food intake, an early warning is triggered when body temperature >39.2℃ and activity level drops sharply by 30%, which in turn links the isolation barrier and medication delivery device, thereby improving the early detection rate of diseases.

[0044] In terms of breeding optimization, behavioral characteristics (number of mountings), physiological indicators (pH 5.5-6.5), and vocal characteristics (2-5 kHz calls lasting >4 hours) can be integrated to predict the optimal mating time.

[0045] Preferably, the multi-dimensional decision control module 400 may further include a task priority scheduling submodule 450. The task priority scheduling submodule 450 is used to regulate the tasks corresponding to the regulation instructions generated by the environmental adaptive regulation submodule 410, the precision feeding decision submodule 420, the health risk assessment submodule 430, and the reproduction optimization submodule 440 according to the task scheduling priority, in descending order of task scheduling priority: health regulation, reproduction regulation, environmental regulation, and feeding regulation.

[0046] In practical applications, adopting a four-level scheduling strategy of "health > reproduction > environment > feeding" can ensure that high-priority tasks (such as health warnings) respond within 1 minute, avoid conflicts of equipment resources, and more effectively guarantee the healthy growth and efficient reproduction of animals.

[0047] As an optional embodiment, the plurality of execution devices 500 may include environmental control devices 510, feeding devices 520, health devices 530, and reproductive assistance devices 540, etc.

[0048] The environmental control equipment 510 can use a variable frequency fan to achieve automatic adjustment of ventilation volume, a heat pump unit to achieve automatic temperature adjustment, and a lighting system to achieve automatic lighting adjustment; among them, the temperature control range of the heat pump unit is preferably 10-30℃, the accuracy is preferably ±1℃, and the lighting system is preferably 0-100% brightness adjustment.

[0049] The feeding equipment 520 can use a spiral feeder to achieve differentiated feeding for each rabbit, and an intelligent waterer to monitor water quality and automatically switch water sources if the water quality exceeds the standard. Among them, the spiral feeder can achieve feed dispensing accuracy of ±1g by recognizing rabbit ear tags through NFC, and the intelligent waterer can achieve recording accuracy of ±5mL of daily water consumption per cage.

[0050] The health equipment 530 can use an automatic drug delivery device to achieve precise drug administration, and an electric isolation fence can isolate sick rabbits within 10 seconds to avoid cross-infection; among them, the automatic drug delivery device is preferably a solid-liquid dual-channel automatic drug delivery device.

[0051] The breeding assistance equipment 540 can use an estrus cage LED marker, a mating record terminal, and automatically link to rabbit files to record mating time, thereby improving breeding efficiency.

[0052] As an optional embodiment, the human-computer interaction device 300 may include at least a visualization management platform 310, an intelligent early warning subsystem 320, and a strategy configuration subsystem 330; wherein, the visualization management platform 310 may be used to display in real time, in a three-dimensional digital interface, the environmental heat map of the target animal's habitat, cage status, abnormal data, growth curves, health reports, and reproductive statistics; the intelligent early warning subsystem 320 may be used to issue early warnings for abnormal detection data in the target animal's habitat according to the early warning mechanism of the target level; and the strategy configuration subsystem 330 may be used to set environmental thresholds, feeding programs, and system parameters.

[0053] Specifically, the 310 visualization management platform can display real-time heat maps of the rabbit hutch environment and cage status, such as feeding, resting, or estrus, through a 3D digital interface; monitor abnormal data, such as flashing red when the ammonia concentration is greater than 15ppm; automatically generate growth curves, such as weekly trend curves of rabbit weight or food intake; automatically generate health reports, such as the proportion of abnormal body temperature or disease incidence; and reproductive statistics reports, such as breeding success rate or monthly comparison reports of breeding numbers.

[0054] The intelligent early warning subsystem 320 is preferably set up with a three-level early warning mechanism. The first level of early warning mechanism is an APP pop-up window (yellow), the second level of early warning mechanism is an SMS notification (orange), and the third level of early warning mechanism is an on-site audible and visual alarm (red). For example, when an animal's body temperature is greater than 39.5℃, the third level of early warning mechanism is triggered, a red alert is issued, and an isolation recommendation is automatically generated.

[0055] The strategy configuration subsystem 330 supports custom environmental thresholds, such as a minimum temperature of 16°C in the adult rabbit area during winter; the feeding plan is to feed the young rabbits 4 times a day, and the parameters are synchronized to the multi-dimensional decision-making layer control module in seconds.

[0056] Based on this, the intelligent management and control system for animal habitats provided by this invention follows the design logic of "perception-transmission-decision-control-application" to ensure a closed-loop data chain. This includes forward data flow and reverse control flow. The core of the forward data flow (data acquisition - user display) is to transform the physical information of the animal habitat into analyzable digital information. The process involves: perception layer acquisition (data is collected by various perception subsystems), transmission layer uploading (sensor data fusion processing, outlier removal, and uploading to the cloud), decision layer processing (different models output different instructions), and application layer analysis (models trained on historical data are called to analyze real-time data, and the analysis results are pushed to the visualization management platform). The core of the reverse data flow (user instruction - device execution) is to convert user or system control instructions into device actions. The process involves: application layer initiating instructions (the administrator issues the "feed rabbit hutch 1" instruction via the APP), decision layer processing instructions (precise feeding decisions are made based on the number of rabbits in hutch 1, age, and feed quantity, generating control instructions), and execution layer executing instructions (the feeding equipment receives the instruction and starts feeding).

[0057] Secondly, the present invention provides an intelligent management and control method for animal habitats, wherein the intelligent management and control method for animal habitats is applied to the intelligent management and control system for animal habitats described in the first aspect; please refer to the following references. Figure 2 This is a schematic diagram of the main process of an intelligent management and control method for animal habitats provided by the present invention. The main body executing the method is a service platform or intelligent animal farm equipped with the intelligent management and control method for animal habitats provided by the present invention.

[0058] exist Figure 2 In this context, the method may include: Step 210: Acquire multi-source sensory data of the target animal's habitat; the multi-source sensory data includes environmental sensory data, physiological sensory data, and reproductive sensory data.

[0059] Step 220: Perform data preprocessing on the multi-source sensing data to obtain target multi-source sensing data.

[0060] In steps 210 to 220, the system activates multi-source sensor devices within the target animal farm to collect environmental perception data, animal physiological perception data, and animal reproductive perception data. Then, the system uses the first network submodule of the data aggregation terminal to preprocess the multi-source perception data to obtain target multi-source perception data. The second network submodule then transmits the target multi-source perception data to the human-computer interaction device and the multi-dimensional decision control module. Finally, step 230 is executed to process the multi-source perception data to obtain control commands for the target animal farm.

[0061] Step 230: Based on the target multi-source sensing data, generate target control instructions for multi-dimensional management and control of the target animal's habitat according to task scheduling priority; the multi-dimensional management and control includes environmental adaptive management and control, precise feeding decision-making, health risk assessment and reproductive optimization.

[0062] Step 240: Based on the target control command, control multiple execution devices to perform corresponding operations.

[0063] In steps 230 to 240, the task scheduling priority is optimized using a four-level scheduling strategy of "health > reproduction > environment > feeding" to ensure that high-priority tasks (such as health warnings) respond within 1 minute. That is, after the multi-dimensional decision control module generates control instructions for environmental adaptive management, precise feeding decision, health risk assessment and reproduction optimization based on the target multi-source perception data, it needs to be sorted according to the task scheduling priority to obtain timely target control instructions. Then, the corresponding execution equipment is operated according to the target control instructions in sequence to perform the corresponding operations, thereby achieving efficient monitoring of the target animal farm.

[0064] Based on this, the intelligent management and control method for animal habitats provided by this invention acquires multi-source sensing data of the target animal habitat; the multi-source sensing data includes environmental sensing data, physiological sensing data, and reproductive sensing data; the multi-source sensing data is preprocessed to obtain target multi-source sensing data; based on the target multi-source sensing data, target control instructions for multi-dimensional management and control of the target animal habitat are generated according to task scheduling priority; the multi-dimensional management and control includes environmental adaptive management and control, precise feeding decision-making, health risk assessment, and reproductive optimization; based on the target control instructions, multiple execution devices are controlled to perform corresponding operations; thereby realizing refined and automated management of the animal habitat from four core dimensions: environment, feeding, health, and reproduction, improving the reproductive efficiency and quality of the target animals; and solving the problem of poor monitoring of animal growth processes in existing technologies.

[0065] As an optional embodiment, in step 210, multi-source sensing data of the target animal's habitat can be acquired in the following two ways.

[0066] Method 1: When the animals in the target animal's habitat are not in their breeding season, data is collected using a multi-source sensing module based on a primary data acquisition frequency to obtain multi-source sensing data. For example, data can be collected at 2-5 minute intervals using the multi-source sensing module, thereby reducing the power consumption of the sensing equipment.

[0067] Method 2: When the animals in the target animal's habitat are in their breeding season, data is collected using a multi-source sensing module based on a second data collection frequency to obtain multi-source sensing data; the first data collection frequency is lower than the second data collection frequency. For example, the data collection frequency can be increased to 1 minute / time during 6:00-8:00 and 18:00-20:00 daily to improve the accuracy of data collection for breeding animals.

[0068] It should be noted that multi-source sensing data needs to include unique identifiers such as timestamps, cage IDs, and animal ear tags to distinguish which individual animal the multi-source sensing data originated from and the time of the data.

[0069] As an optional embodiment, in step 220, the edge gateway of the first network submodule can be used to process the data by the 3σ principle to remove outliers in the collected data and normalize the processed data; then the second network submodule can be used to encrypt and upload the data to the cloud via the 5G network.

[0070] Specifically, continuing with the rabbit example, the application steps of the 3σ principle can include: data collection and preprocessing (such as the daily feed intake of a single rabbit, temperature monitoring data, etc.), eliminating obvious errors (such as negative numbers, values ​​exceeding physical limits); calculating the mean μ and standard deviation σ (mean μ = sum of data / number of data points, standard deviation σ = sum of data points / number of data points). (where n is the number of data points); determine the 3σ interval (calculate " " and" ", as the boundary for judging outliers"; identify and handle outliers (if the data falls within ""). "Outliers are marked as outliers; the cause of the outlier is analyzed, and if it is a data error, it is directly removed. After removing outliers using the 3σ principle, Z-score normalization can be used to map the data to the 0-1 range, obtaining target multi-source sensing data. This improves the accuracy of the data and ensures the security of data transmission."

[0071] As an optional embodiment, in step 230, based on the target multi-source sensing data, a target control instruction for multi-dimensional management and control of the target animal's habitat is generated according to the task scheduling priority. The target control instruction can be obtained according to the methods corresponding to S1 to S2.

[0072] S1: Based on the target multi-source sensing data, generate health regulation instructions, reproductive regulation instructions, environmental regulation instructions, and feeding regulation instructions for the target animal's living space.

[0073] S2: Sort the health control instructions, reproductive control instructions, environmental control instructions and feeding control instructions according to the task scheduling priority in descending order of health control, reproductive control, environmental control and feeding control, and obtain the target control instructions.

[0074] In S2, based on the target multi-source sensing data, health regulation instructions, reproductive regulation instructions, environmental regulation instructions, and feeding regulation instructions for the target animal's living space are generated respectively. The regulation instructions can be obtained according to the methods corresponding to S21 to S24.

[0075] S21: Based on environmental perception data and combined with the animal's growth stage, the DDPG algorithm is used to generate environmental control instructions. The environmental control instructions include at least dynamically adjusting the ventilation, humidity, light and temperature in the target animal's living area.

[0076] Specifically, continuing with the rabbit example, dynamic adjustment based on the DDPG algorithm can include the following three aspects: a. Precise environmental adaptation during the breeding season. Input data includes: the doe's physiological stage (estrus / pregnancy / lactation), real-time temperature and humidity, ammonia concentration, and light intensity. Adjustment strategy: During pregnancy, when the temperature deviation is > +1℃, the fan power increases in a 0.5% / minute gradient (to avoid sudden cooling); when the humidity is < 50% RH, the humidifier setting automatically adjusts dynamically to match the doe's water consumption (a 10% increase in water consumption corresponds to a 20% increase in humidification intensity). During lactation, the light cycle is adjusted to 14L:10D (to promote lactation), and weather forecast data is used to preheat the oven to 21℃.

[0077] b. Extreme weather emergency response. Status input: real-time temperature and humidity, ammonia concentration, and weather forecast data (temperature / humidity for the next 24 hours). Response strategy: High temperature warning (30℃≤T<32℃), increase fan power to 80%, start pre-cooling cycle, if humidity>60%RH, prioritize dehumidification-ventilation linkage mode.

[0078] c. Refined control during the rabbit rearing period. Input parameters: rabbit age (days), real-time temperature and humidity, and light intensity. Control strategy: During lactation, when the mother rabbit is nursing, the light intensity automatically drops to 50 lux (simulating a nighttime environment), and local heating is activated. If nursing time is less than 5 minutes per session, environmental compensation is triggered: humidity is increased by 3% RH, and white noise (40-50 dB) is played to soothe the rabbits. Weaning transition period: When the rabbits are older than 20 days, the temperature is gradually reduced to 22℃ in a 0.5℃ / day gradient, while the duration of light exposure is gradually reduced (by 1 hour per day).

[0079] S22: Based on physiological perception data and target matrix, a random forest decision-making mechanism is used to predict the target feed intake of animals at different growth stages and generate feeding control instructions.

[0080] Utilizing random forests for decision-making mechanisms driven by multi-dimensional data can include the following two aspects: a. Nutritional regulation of breeding does: Adjust protein content, feed formula, and pellet feed amount according to the gestation stage (weeks 1-3 / 4-5) and weight change rate (±5% / week). For example, if the protein content increases by 20% in early pregnancy, adjust the feed formula to 25% soybean meal + 15% alfalfa meal, and reduce the pellet feed amount by 5%. b. Nutritional regulation during the rearing period of young rabbits: Adjust feed particle size and feed formula according to the age of the young rabbits (15-30 days old) and weaning weight (300-500g). During the weaning transition period, if feed intake is less than 80% of the target value, add 1% protease preparation; during the rapid growth period, when the daily weight gain is greater than 25g, the protein requirement increases by 15%, and the feeding frequency increases from 3 times to 4 times a day.

[0081] S23: Generate health regulation instructions based on physiological perception data.

[0082] Dynamic early warning of risks such as diseases and stress can be achieved through temporal anomaly detection by analyzing the temporal sequence characteristics of sensor data, behavioral trajectories, and physiological indicators. Specifically, this can include the following two aspects: a. Early disease warning and transmission interruption: Early symptoms of infectious diseases in rabbit populations are often subtle. Time-series anomaly detection, achieved through multi-module data fusion, enables early detection and intervention. Data acquisition uses accelerometers to record activity levels, combined with feed trough pressure sensors to calculate the feed intake ratio (target dwell time / total time). An infrared thermometer measures the overall body temperature. If a single rabbit's feed intake ratio is less than 60% for 24 consecutive hours and its daily activity level is less than 180 minutes, the system automatically isolates and relocates it, and sends it to a veterinarian for examination. If three or more rabbits in the same rabbitry show abnormalities, ultraviolet disinfection and air filtration are activated to block the source of transmission.

[0083] b. Health management during the rearing period of young rabbits: Young rabbits (0-30 days old) have a fragile immune system, and diarrhea and respiratory diseases are the main causes of death. Time-series anomaly detection utilizes growth curve tracking and environmental compensation to dynamically adjust health risks. For example, if the daily weight gain of young rabbits is set at 15-20g, individuals deviating by more than 15% for three consecutive days are identified as having diarrhea events. When the diarrhea incidence rate is >5%, antibiotics are administered at a dose of 0.5mg / kg body weight.

[0084] S24: Based on reproductive perception data, use a preset neural network model to generate reproductive control instructions including the target mating period; the preset neural network model is a network model obtained after model training based on multimodal animal data.

[0085] Neural networks achieve intelligent optimization of reproductive decisions through multi-dimensional data fusion and dynamic modeling. Specifically, they enable accurate prediction of the estrus cycle and optimization of mating timing in female rabbits. Given the short estrus cycle (4-7 days) and subtle symptoms in female rabbits, neural networks use physiological and behavioral multimodal data fusion to predict the estrus cycle and accurately pinpoint the mating window. A pH sensor collects physiological indicators from secretions, and a visual camera records behavioral characteristics such as mounting frequency (5 times / hour), pH value (5.8), and vocalization frequency (3kHz for 6 hours). The system then pushes the optimal mating window (current time + 1.5 hours) and illuminates the corresponding cage LED light.

[0086] Based on this, health control commands, reproductive control commands, environmental control commands, and feeding control commands for the target animal habitat are generated based on multi-source sensing data. Then, the health control commands, reproductive control commands, environmental control commands, and feeding control commands are sorted according to the task scheduling priority in descending order, to obtain the target control commands. These commands are then sent to multiple execution devices for sequential execution, improving the accuracy and timeliness of monitoring the animal habitat.

[0087] As an optional embodiment, after step 240, the intelligent management and control method for animal habitats provided by the present invention may further include: acquiring target feedback data of the target animal habitat within a target time period; the target feedback data includes at least the mating success rate and feed conversion rate; updating system parameters based on the target feedback data, and using an incremental learning algorithm to adapt to changes in the growth cycle of the target animal, forming a closed loop of acquisition, decision-making, execution, and feedback.

[0088] Specifically, after the execution layer device completes the instruction, it is preferred to provide feedback within 30 seconds, such as "medication completed" or "isolation fence closed in place" within 30 seconds. The multi-dimensional decision control module can update the model parameters weekly based on feedback data (such as mating success rate and feed conversion rate), and use incremental learning algorithms to adapt to changes in the rabbit's growth cycle, forming a closed loop of "collection-decision-execution-feedback".

[0089] Based on this, the intelligent management and control method for animal habitats provided by this invention synchronously collects environmental perception data, physiological perception data, and reproductive perception data of the target animal habitat. First, outliers are removed, and Z-score normalization is applied. Then, multi-dimensional decision-making is achieved based on four types of algorithms: environmental control (DDPG algorithm), dynamically outputting the optimal combination of fan speed, heat pump temperature, and light duration; precise feeding (random forest algorithm), predicting the daily feed intake of a single rabbit; health assessment (temporal anomaly detection algorithm), performing temporal analysis of body temperature, activity level, and feed intake to determine whether a health warning is triggered; and reproductive optimization (neural network model), integrating behavioral, physiological, and auditory features to predict the optimal mating window. Execution resources are allocated according to the priority order: health warning > reproductive needs > environmental control > feeding management. After receiving instructions, the execution layer devices act precisely and provide real-time feedback on the execution status. The feedback data is then used for model iteration and optimization. This achieves a closed-loop management system for target animal farms, encompassing "data acquisition, preprocessing, intelligent decision-making, priority scheduling, instruction execution, and feedback iteration." It enables refined and automated management of animal farms across four core dimensions: environment, feeding, health, and reproduction, thereby improving the reproductive efficiency and quality of target animals. This also solves the problem of poor monitoring of animal growth processes in existing technologies.

[0090] Although the invention has been described herein in conjunction with various embodiments, those skilled in the art will understand and implement other variations of the disclosed embodiments by reviewing the accompanying drawings, the disclosure, and the appended claims in carrying out the claimed invention. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0091] Although the invention has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made therein without departing from the spirit and scope of the invention. Accordingly, this specification and drawings are merely exemplary descriptions of the invention as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its spirit and scope. Thus, if such modifications and modifications of the invention fall within the scope of the claims and their equivalents, the invention is also intended to include such modifications and modifications.

Claims

1. An intelligent management and control system for an animal living field, characterized in that, At least comprising: Multi-source perception module, data transmission module, multi-dimensional decision control module, multiple execution devices and human-computer interaction device; The multi-source perception module is connected with the data transmission module, the data transmission module is connected with the multi-dimensional decision control module, the multi-dimensional decision control module is connected with multiple execution devices, and the human-computer interaction device is connected with the multi-dimensional decision control module, the data transmission module and multiple execution devices respectively; The data transmission module is used for acquiring multi-source perception data of a target animal living field, and performing data preprocessing on the multi-source perception data to obtain target multi-source perception data; the multi-source perception data includes environmental perception data, physiological perception data and reproduction perception data; The multi-dimensional decision control module is used for generating a target control instruction for multi-dimensional control of the target animal living field based on the target multi-source perception data and according to task scheduling priority; the multi-dimensional control includes environmental self-adaptive control, precise feeding decision, health risk assessment and reproduction optimization; Multiple execution devices are used for executing corresponding operations according to the target control instruction issued by the multi-dimensional decision control module; The human-computer interaction device is used for visually displaying the target animal living field based on the target multi-source perception data, generating prompt information based on an early warning mechanism and setting system parameters.

2. The intelligent management and control system of the animal living yard according to claim 1, characterized in that, The multi-source perception module at least includes an environmental perception submodule, a physiological perception submodule and a reproduction perception submodule; The environmental perception submodule is used for at least based on distributed deployment of temperature and humidity sensors, light sensors and ammonia or hydrogen sulfide composite sensors to perform spatial grid monitoring on animal living field environmental parameters; The physiological perception submodule is used for at least based on embedded weight sensors, feed trough pressure sensors, acceleration sensors and infrared temperature arrays to perform physiological monitoring on target animals; The reproduction perception submodule is used for at least based on visual cameras, pH sensors and sound collection microphones to perform reproduction monitoring on target animals. 3.The intelligent management and control system of the animal living yard according to claim 1, characterized in that, The data transmission module includes a first network submodule and a second network submodule; The first network submodule is used for collecting multi-source perception data in the target animal living field, and performing data preprocessing on the collected multi-source perception data using the 3σ principle to obtain target multi-source perception data; The second network submodule is used for remotely transmitting the target multi-source perception data, and buffering the target multi-source perception data when the network is disconnected, and automatically resending after networking.

4. The intelligent management and control system of the animal life field according to claim 1, wherein, The multi-dimensional decision control module at least includes an environmental self-adaptive control submodule, a precise feeding decision submodule, a health risk assessment submodule and a reproduction optimization submodule; The environmental self-adaptive control submodule is used for generating environmental control instructions based on environmental perception data and combining animal growth stages, and dynamically adjusting the ventilation, humidity and temperature in the target animal living field using the DDPG algorithm; the growth stage at least includes a juvenile stage, an adult stage and a reproduction stage; The precise feeding decision submodule is configured to predict target feed intake of animals at different growth stages and generate feeding control instructions by using a random forest decision mechanism based on physiological sensing data and a target matrix, and to perform individualized feeding; The health risk assessment submodule is configured to generate health control instructions based on physiological sensing data, and to perform disease monitoring on animals at different growth stages in combination with isolation pens and drug delivery equipment; The breeding optimization submodule is configured to generate breeding control instructions including a target breeding period by using a preset neural network model based on breeding sensing data; the preset neural network model is a network model obtained by model training based on multi-modal animal data.

5. The intelligent management and control system of the animal life field according to claim 4, wherein, The multi-dimensional decision control module further comprises a task priority scheduling submodule; The task priority scheduling submodule is configured to control tasks corresponding to control instructions generated by the environment self-adaptive control submodule, the precise feeding decision submodule, the health risk assessment submodule, and the breeding optimization submodule in order of decreasing task scheduling priority of health control, breeding control, environment control, and feeding control.

6. The intelligent management and control system of the animal life field according to claim 1, wherein, The human-computer interaction device at least comprises a visual management platform, an intelligent early warning subsystem, and a strategy configuration subsystem; The visual management platform is at least configured to display an environmental heat map, a cage status, abnormal data, a growth curve, a health report, and breeding statistics of a target animal living farm in real time in a three-dimensional digital interface; The intelligent early warning subsystem is at least configured to perform early warning on abnormal detection data in the target animal living farm according to a target level early warning mechanism; The strategy configuration subsystem is at least configured to set environment thresholds, feeding schemes, and system parameters.

7. An intelligent management method for an animal living field, characterized in that, The intelligent management and control method of the animal living farm is applied to the intelligent management and control system of the animal living farm according to any one of claims 1 to 6, and the method comprises: Obtaining multi-source sensing data of a target animal living farm; the multi-source sensing data comprises environmental sensing data, physiological sensing data, and breeding sensing data; Data preprocessing is performed on the multi-source sensing data to obtain target multi-source sensing data; Based on the target multi-source sensing data, target control instructions for multi-dimensional management and control of the target animal living farm are generated according to task scheduling priority; the multi-dimensional management and control comprises environment self-adaptive management and control, precise feeding decision, health risk assessment, and breeding optimization; Based on the target control instructions, a plurality of execution devices are controlled to perform corresponding operations. 8.The intelligent management method of the animal life field according to claim 7, wherein, The multi-source sensing data of the target animal living farm is obtained by: When the animals in the target animal living farm are in a non-breeding period, data is collected by a multi-source sensing module based on a first data collection frequency to obtain the multi-source sensing data; When the animals in the target animal living farm are in a breeding period, data is collected by a multi-source sensing module based on a second data collection frequency to obtain the multi-source sensing data; the first data collection frequency is less than the second data collection frequency. 9.The intelligent management method of the animal living field according to claim 7, wherein, Based on the target multi-source sensing data, target control instructions for multi-dimensional management and control of the target animal living farm are generated according to task scheduling priority, which comprises: generate, based on the target multi-source perception data, health regulation instructions, breeding regulation instructions, environment regulation instructions, and feeding regulation instructions for the target animal living field; sort the health regulation instructions, breeding regulation instructions, environment regulation instructions, and feeding regulation instructions according to task scheduling priorities in descending order of health regulation, breeding regulation, environment regulation, and feeding regulation to obtain the target control instructions; The method further comprises: generate, based on the environment perception data and in combination with the growth stage of the animal, the environment regulation instructions by using a DDPG algorithm, wherein the environment regulation instructions at least include dynamically adjusting the ventilation, humidity, and temperature in the target animal living field; generate, based on the physiological perception data and the target matrix, the feeding regulation instructions by using a random forest decision mechanism to predict the target feeding amount of the animal at different growth stages; generate the health regulation instructions based on the physiological perception data; generate, based on the breeding perception data, the breeding regulation instructions including the target mating period by using a preset neural network model, wherein the preset neural network model is a network model obtained after model training based on multi-modal animal data. 10.The intelligent management method of the animal living field according to claim 7, wherein, The method further comprises: acquire target feedback data of the target animal living field within a target time period, wherein the target feedback data at least includes a mating success rate and a feed conversion rate; update system parameters based on the target feedback data, and adapt to changes in the growth cycle of the target animal by using an incremental learning algorithm to form a closed loop of collection, decision-making, execution, and feedback.

Images