Parrot breeding auxiliary system

Abstract

The invention relates to the technical field of intelligent breeding, and discloses a parrot breeding auxiliary system, which comprises a knowledge base construction module, a data acquisition module, a digital portrait module, a strategy generation module, a strategy execution module and a closed-loop optimization module, and is characterized in that the knowledge base comprises a rule base, a case base and a model base; the strategy execution module comprises an environment regulation and control unit, an automatic feeding unit and a behavior training unit; according to the method, paradigm transformation from experience driving to data driving is realized, breeding scientificity is improved, personalized requirements of parrots are met, animal welfare is remarkably improved, a knowledge core with sustainable evolution ability is constructed, time cost and professional knowledge threshold of feeders are greatly reduced, the intelligent upper limit of the system is broken through, and the method is suitable for large-scale popularization and application. And through intelligent cooperation of multiple modules, integrated fine management of the whole breeding process is realized.

Description

Technical Field

[0001] This invention relates to the field of intelligent aquaculture technology, and more specifically to a parrot aquaculture auxiliary system. Background Technology

[0002] Parrots are highly intelligent and social birds that are kept as ornamental and companion birds. Their breeding process is complex and requires a lot of energy and expertise from the breeder.

[0003] Currently, parrot breeding mainly relies on a combination of artificial management and basic equipment. In terms of environmental control, constant temperature boxes or air conditioning equipment are usually used to maintain the basic temperature requirements according to the parrot species, and artificial lighting is used to simulate the day and night cycle. Feeding management is mainly accomplished by the keepers feeding the parrots the formulated feed at regular times and changing the drinking water regularly. Some farms use simple automatic feeders to assist. In terms of behavior management, perches, chew toys and other facilities are mainly provided to meet the parrots' basic activity needs. During the breeding process, the keepers mainly rely on daily observation records to grasp the parrots' health status and behavior.

[0004] Current technology still has the following shortcomings:

[0005] Existing technologies rely heavily on human experience, lack objective data support, and are not scientifically sound. Breeding decisions are based entirely on the subjective observation and personal experience of the breeders, with vague judgment criteria and significant differences in operation among different breeders, making it difficult to form a standardized and replicable scientific breeding process.

[0006] Existing technologies employ extensive management models that cannot meet the individual differences in needs; uniform environmental settings, feeding standards, and a lack of psychological intervention cannot respond to the special needs of parrots of different species, ages, and health conditions, which can easily lead to stress, malnutrition, or psychological problems in parrots, resulting in low levels of animal welfare.

[0007] Existing automated equipment has fixed functions, and the system's intelligence limit is locked at the factory. Traditional automated equipment has a closed and static rule system, which cannot learn from practice, cannot adapt to new breeding discoveries or handle unknown situations, and does not have the ability to self-optimize.

[0008] In existing technologies, the various links are isolated from each other, forming data silos; environmental controllers, feeders and other equipment work independently, data cannot be shared, decision-making cannot be linked, and keepers need to act as information relay stations, making it difficult to carry out collaborative management from a global perspective, resulting in low efficiency and easy to miss problems.

[0009] Existing technologies are time-consuming and have high professional barriers, making them difficult to popularize. Refined breeding requires keepers to invest a lot of time in daily operations and have considerable professional knowledge to judge the parrots' condition, which makes it difficult to effectively promote scientific breeding among ordinary enthusiasts.

[0010] Therefore, scientific, personalized, continuously evolving, integrated management, and methods that lower the professional threshold are needed to solve the above problems. Summary of the Invention

[0011] In order to overcome the above-mentioned defects of the prior art, the present invention provides a parrot breeding auxiliary system to solve the problems existing in the background art.

[0012] To achieve the above objectives, the present invention provides the following technical solution: a parrot breeding auxiliary system, comprising:

[0013] Knowledge base construction module: Used to build and maintain a parrot breeding knowledge base based on anonymized data from a group of parrots and external authoritative knowledge sources. The knowledge base specifically includes:

[0014] Rule base: Used to store expert experience in "IF-THEN" format and the discovered association rules;

[0015] Case Library: Used to store typical parrot digital profiles, the breeding strategies applied, and successful / failed cases of their implementation effects;

[0016] Model library: Used to store pre-trained and continuously optimized machine learning model parameters for building digital images of parrots;

[0017] Data acquisition module: used to collect multi-source data of the target parrot, including individual habit data, environmental data, historical feeding data, and behavioral performance data;

[0018] Digital profiling module: Based on the multi-source data, continuously calls the model library in the knowledge base to construct a digital profile of the parrot that describes the personalized needs and behavioral characteristics of the target parrot;

[0019] Strategy generation module: used to match and intelligently analyze the digital profile of the parrot with the rule base and case base in the parrot breeding knowledge base to generate personalized breeding assistance strategies for the target parrot;

[0020] Strategy execution module: used to receive and execute the personalized aquaculture assistance strategy, the strategy execution module specifically includes:

[0021] Environmental control unit: used to regulate aquaculture environmental parameters;

[0022] Automatic feeding unit: used for automatically dispensing food and water;

[0023] Behavior training unit: used for automated behavior induction and training of parrots;

[0024] Closed-loop optimization module: Used to collect feedback data after strategy execution, optimize the parrot digital profile, and send the feedback data back to the knowledge base to drive the continuous iteration and evolution of the knowledge base.

[0025] The technical effects and advantages of this invention are as follows:

[0026] 1. This invention constructs a digital profile of parrots, transforming vague breeding experience into quantifiable multi-dimensional indicators, enabling breeding decisions to be based on precise data analysis, eliminating absolute dependence on the personal experience of breeders, and significantly improving the scientific nature and repeatability of the breeding process.

[0027] 2. This invention can identify and respond to the unique habits, health status and behavioral preferences of each parrot, generate and implement customized environment, feeding and training strategies for them, accurately meet their physiological and psychological needs, effectively prevent behavioral problems such as feather pecking, and thus elevate animal welfare to a new level.

[0028] 3. By introducing a closed-loop optimization mechanism, this invention enables the system to learn from the effects of each strategy execution, continuously feeding back successful cases and validated new rules into the knowledge base, thus becoming an expert system that can iterate with data accumulation and become smarter with use.

[0029] 4. This invention integrates the three previously isolated functions of environmental control, feeding management, and behavioral training through a unified data profile and strategy engine for deep integration and intelligent scheduling. The modules no longer operate independently but work collaboratively to achieve common breeding goals, thus realizing truly integrated, refined, and intelligent breeding.

[0030] 5. This invention automates repetitive tasks such as daily environmental control, timed feeding, and behavioral training, and transforms complex health and behavioral judgments into intuitive warnings and suggestions, greatly freeing up the time and energy of breeders, enabling even non-professional enthusiasts to carry out scientific and meticulous parrot breeding. Attached Figure Description

[0031] Figure 1 This is a structural block diagram of the present invention.

[0032] Figure 2 This is a flowchart of the present invention. Detailed Implementation

[0033] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. In addition, the forms of the various structures described in the following embodiments are merely illustrative. The parrot breeding auxiliary system involved in the present invention is not limited to the structures described in the following embodiments. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0034] Reference Figure 1 This invention provides a parrot breeding auxiliary system, including a knowledge base construction module, a data acquisition module, a digital profiling module, a strategy generation module, a strategy execution module, and a closed-loop optimization module. The knowledge base includes a rule base, a case base, and a model base, and the strategy execution module includes an environmental control unit, an automatic feeding unit, and a behavior training unit.

[0035] Reference Figure 2 The specific implementation steps of the present invention include the following steps:

[0036] S1. Based on anonymized data from a group of parrots and authoritative external knowledge sources, construct and maintain a parrot breeding knowledge base.

[0037] It should be specifically noted that the knowledge base includes:

[0038] Rule base: Used to store expert experience in the form of "IF-THEN" and the discovered association rules; and through association rule mining algorithms, such as the Apriori algorithm, to automatically discover new and potential causal relationships from the population data. New rules are added to the database after confidence evaluation, enabling the system to have knowledge discovery capabilities.

[0039] For example, there is a strong correlation between "persistently low ambient humidity and insufficient light" and "increased incidence of feather pecking";

[0040] Case Library: Used to store typical parrot digital profiles, the breeding strategies applied, and successful / failed cases of their implementation effects;

[0041] It should be explained that each case study is a complete data structure, including a problem profile, application strategy, and effect evaluation; for example:

[0042] Problem description: "Lovebird, exhibiting restlessness, with a 30% decrease in activity level";

[0043] Application strategy: "Introduce educational foraging toys and train twice a day";

[0044] Effect assessment: "Activity returned to normal after one week";

[0045] Model library: Used to store pre-trained and continuously optimized machine learning model parameters for building digital images of parrots;

[0046] For example, pre-trained convolutional neural network models for identifying specific parrot behaviors such as preening, singing, and plucking feathers from video streams; and classification model parameters for predicting health risks.

[0047] It should be specifically noted that the construction steps of the parrot breeding knowledge base are as follows:

[0048] A1. Knowledge Acquisition: Extract structured knowledge from authoritative external sources, such as species guidelines and veterinary literature, and receive anonymized population data uploaded from multiple networked individual breeding support systems;

[0049] The group data includes: parrot species A, digital profile indicators (comfort index, activity level, etc.), implemented strategies (environmental parameters, feeding plan, etc.), and strategy effects (feeding completion rate, behavior improvement, etc.).

[0050] A2. Knowledge Mining: Perform association rule mining, cluster analysis, and pattern recognition on the group data to discover new breeding patterns;

[0051] For example, the new breeding pattern is {Species = Macaw, Behavior = Frequent nodding, Environment = Insufficient light} → {Health risk = Vitamin deficiency}, with a confidence level of 85%.

[0052] A3. Knowledge Integration and Verification: Integrate newly discovered patterns with existing knowledge and detect conflicts, and verify them through statistical significance tests or expert review mechanisms;

[0053] The system marks the new breeding patterns as to be observed and actively pushes them to some users who are raising macaws and whose lighting is insufficient, suggesting that they observe whether their parrots are nodding frequently.

[0054] A4. Knowledge Update: Injecting verified new knowledge into the rule base, case base, or model base of the knowledge base.

[0055] Once the amount of feedback data exceeds a certain threshold and the new breeding pattern is confirmed, its confidence level is increased, and it is formally incorporated into the rule base of the knowledge base.

[0056] S2. Collect multi-source data of the target parrot, including individual habit data, environmental data, historical feeding data, and behavioral performance data.

[0057] It should be specifically noted that the individual behavior data is obtained through user input and / or image recognition. In addition to the basic parrot species, age and sex, the individual preferences, such as a preference for a certain food or toy, can be learned through user interaction interface or long-term behavioral observation.

[0058] The environmental data is acquired through environmental sensors, including temperature, humidity, light intensity, noise, and air quality.

[0059] The historical feeding data is obtained through automatic feeding records and user logs, including feeding duration, food type, food amount, and the parrot's feeding response;

[0060] The behavioral data is acquired through image / sound acquisition devices and analyzed using behavior recognition algorithms, including activity level, frequency and type of chirping, and specific behaviors such as the frequency and duration of preening, feather pecking, and playing.

[0061] S3. Based on the multi-source data, continuously call the model library in the knowledge base to construct a digital profile of the parrot that describes the personalized needs and behavioral characteristics of the target parrot.

[0062] It should be specifically explained that the multi-source data is cleaned, normalized, and time-series aligned. Algorithms in the model library are called, and multiple dedicated models in the model library perform dimensional calculations and then fuse them to finally output a multi-dimensional indicator system, the Parrot Digital Profile. The dimensional calculations specifically include physiological comfort index, mental health score, nutrition and health prediction, and training response.

[0063] The physiological comfort index is calculated as follows:

[0064] Model to use: Lightweight time-series classification model or rule engine;

[0065] Input: Real-time environmental data, including temperature, humidity, and wind speed, and real-time attitude characteristics;

[0066] Processing: The model compares the input features with the optimal comfort range for the species in the knowledge base. For temperature, humidity, wind speed and posture factors, it calculates the degree of deviation of the current value from the optimal comfort range and maps it to a score of 0-100, where 100 represents being in the optimal state.

[0067] The temperature factor fraction is calculated as follows:

[0068] Retrieve the optimal temperature range [Tol, Toh] and tolerable range [Ttl, Tth] for this species from the knowledge base, and calculate the score using the membership function:

[0069] If the current temperature Tc is in [Tol, Toh], then St = 100; if Tc is in [Ttl, Tool) or (Toh, Tth], then St decreases linearly from 100 to 0; if Tc exceeds the tolerable range, then St = 0.

[0070] The humidity factor and wind speed factor are calculated in the same way as the temperature factor.

[0071] The attitude factor score is calculated as follows:

[0072] The system retrieves a series of uncomfortable postures and their thresholds from the knowledge base, such as: mouth breathing rate Pb, which deducts points if it exceeds the threshold Thb; and tufting coefficient Pf, which deducts points if it exceeds the threshold Thf.

[0073] Sp = max(0, 100 - (Pb / Thb*50) - (Pf / Thf*50)), where each of the two attitudes accounts for 50 points, ensuring that Sp is not less than 0.

[0074] The scores of each factor are weighted and fused to finally output a continuous score from 0 to 100.

[0075] The mental health score is calculated as follows:

[0076] Invoking models: Abnormal behavior detection models, such as unsupervised models based on isolated forests or autoencoders;

[0077] Input: The frequency of abnormal behavior within a certain time window, such as the number of stereotypical pacing steps and the duration of feather pecking;

[0078] Processing: The model learns the baseline data of the parrot's normal behavior pattern, calculates the degree of deviation of the current behavior data from the baseline. The greater the deviation, the lower the mental health score. At the same time, the case library in the knowledge base provides prior knowledge, such as doubling the weight of the mental health score if feather pecking behavior is observed for three consecutive days.

[0079] The nutrition and health prediction calculation is specifically as follows:

[0080] Models to use: classification models, such as logistic regression and gradient boosting trees, or time series prediction models;

[0081] Input: Historical eating patterns include trends in average daily food intake and eating speed, weight change curves, and changes in activity levels.

[0082] Processing: The model outputs one or more risk probabilities based on the input features. For example, the probability of being underweight within the next two weeks is 30%. The model's weight parameters are derived from the training results of a large amount of similar parrot data in the knowledge base.

[0083] The training response rate calculation is specifically as follows:

[0084] Model to be used: value function in reinforcement learning or simple linear regression model;

[0085] Input: Historical training records, such as success rate and average response time for different instructions;

[0086] Processing: For different training items, such as standing on a pole and recognizing colors, the model calculates a response score. The response score is calculated as: Base success rate × Weight 1 + (1 / Average reaction time) × Weight 2. The weights are defined by the knowledge base. For example, for newly learned instructions, the weight of reaction time is higher.

[0087] It should be explained that the calculation results of the above four dimensions, together with the parrot's static attributes, species, age and sex, are combined into a structured feature vector, which is the parrot's digital profile at the current moment. The parrot's digital profile is a structured data object that, in addition to containing the quantitative indicators of the above four dimensions, also contains metadata such as timestamps and data confidence levels, which together constitute the unique and unified digital basis for the system to make decisions.

[0088] S4. Match and intelligently analyze the digital image of the parrot with the rule base and case base in the parrot breeding knowledge base to generate a personalized breeding assistance strategy for the target parrot.

[0089] It should be specifically noted that the formulation of the personalized aquaculture assistance strategy adopts a hybrid intelligent reasoning mechanism, specifically as follows:

[0090] First, case matching is performed: the similarity between the current digital profile and the problem profiles in the case library is calculated. When the similarity value is higher than a preset threshold, the historical strategy is directly adopted or fine-tuned to achieve experience reuse. The similarity calculation can use algorithms such as cosine similarity or Euclidean distance to search in the feature space of the case library.

[0091] Secondly, if there is no matching case, rule reasoning is initiated: the current profile data is matched with the conditions in the rule base, and a new execution strategy is derived through the logic chain to achieve logical fallback; the rule reasoning engine is used to match the current profile data with the conditions in the rule base, trigger all rules that meet the conditions, and execute the strategy defined in their conclusions.

[0092] The final personalized aquaculture support strategy is an executable set of instructions, such as: "Environmental control unit: Set the temperature to 26℃ and simulate the sunrise pattern for the light cycle"; "Behavioral training unit: Start the 'target pole recognition' training course at 3 pm for 10 minutes".

[0093] S5. Receive and execute the personalized aquaculture assistance strategy.

[0094] The strategy execution module specifically includes:

[0095] Environmental control unit: used to regulate aquaculture environmental parameters;

[0096] The integrated control system, which combines a thermostat, humidifier, dehumidifier, and full-spectrum LED lighting system, can simulate the climate characteristics of different regions and seasons.

[0097] Automatic feeding unit: used for automatically dispensing food and water;

[0098] It features a multi-compartment precision feeding mechanism that supports the feeding of various materials such as dry food, nuts, and nutritional supplements. It can achieve precise feeding at fixed times, in fixed quantities, and with fixed ratios according to strategies, and also has a low-feed alarm function.

[0099] Behavior training unit: used for automated behavior induction and training of parrots; includes:

[0100] Driveable training equipment: such as a movable target stick controlled by a servo motor, or an instruction button with illuminated prompts;

[0101] Automatic reward mechanism: A small precision feeder that works in conjunction with the automatic feeding unit to provide positive reinforcement immediately after the parrot completes a specified action.

[0102] For example, the specific steps of target pole training in the aforementioned behavior training are as follows:

[0103] B1. The behavior training unit is activated, and its drive mechanism begins to slowly move the target rod.

[0104] B2. The built-in camera starts working to identify whether the parrot has stood on the target pole;

[0105] B3. Once a successful behavior is identified by the image recognition algorithm, the strategy execution module immediately triggers the automatic reward mechanism and distributes reward food.

[0106] B4. The results of this training, such as the number of successful attempts and reaction time, are recorded as feedback data.

[0107] S6. Collect feedback data after the strategy is executed, optimize the parrot digital profile, and send the feedback data back to the knowledge base to drive the knowledge base to continuously iterate and evolve.

[0108] It should be specifically noted that the closed-loop optimization module includes short-term optimization and long-term evolution, specifically:

[0109] Short-term optimization: Feedback data is directly used to optimize the digital profile of the target parrot, enabling real-time updates of individual models;

[0110] Long-term evolution: The "strategy-effect" data package after desensitizing all individuals is sent back to the knowledge base construction module;

[0111] Successful cases are stored in the case library; universally effective patterns are extracted into new rules and added to the rule library; massive amounts of data are used to retrain models in the model library in the cloud, improving their prediction and recognition accuracy.

[0112] Through the above description of the embodiments, those skilled in the art can clearly understand that the various embodiments of this application can be implemented by means of software or software combined with necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware functions. Based on this understanding, the technical solution of this application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. The software product is stored in a storage medium and includes several instructions to cause a computer device, such as including but not limited to a personal computer, server, or network device, to execute all or part of the steps of the method described in any embodiment of this application.

[0113] The foregoing has described exemplary embodiments of this application. It should be understood that the above exemplary embodiments are not restrictive but illustrative, and the scope of protection of this application is not limited thereto. It should be understood that those skilled in the art can make modifications and variations to the embodiments of this application without departing from the spirit and scope of this application, and such modifications and variations should be within the scope of protection of this application.

Claims

1. A parrot breeding assistance system characterized by, Specifically comprising: Knowledge base construction module: for constructing and maintaining a parrot breeding knowledge base based on anonymized data from the parrot community and external authoritative knowledge sources, which specifically includes: Rule base: for storing expert experience and mined association rules in the form of "IF-THEN"; Case base: for storing typical parrot digital portraits, applied breeding strategies and their success / failure cases; Model base: for storing pre-trained and continuously optimized machine learning model parameters for building parrot digital portraits; Data collection module: for collecting multi-source data of target parrots, including individual habit data, environment data, historical feeding data and behavior performance data; Digital portrait module: for continuously calling the model base in the knowledge base based on the multi-source data to construct a parrot digital portrait describing the individualized needs and behavior characteristics of the target parrot; Strategy generation module: for matching and intelligently analyzing the parrot digital portrait with the rule base and case base in the parrot breeding knowledge base to generate individualized breeding assistance strategies for the target parrot; Strategy execution module: for receiving and executing the individualized breeding assistance strategies, which specifically includes: Environment control unit: for controlling breeding environment parameters; Automatic feeding unit: for automatically dispensing food and water; Behavior training unit: for automatically inducing and training parrots; Closed-loop optimization module: for collecting feedback data after strategy execution, optimizing the parrot digital portrait, and feeding the feedback data back to the knowledge base to drive continuous iteration and evolution of the knowledge base.

2. A parrot breeding assistance system according to claim 1, characterized in that: The construction steps of the parrot breeding knowledge base are specifically: A1, knowledge acquisition: extract structured knowledge from external authoritative materials such as species guides and veterinary literature, and receive anonymized group data uploaded from multiple individual breeding assistance systems connected online; the group data content includes: parrot species, digital portrait indicators, executed strategies and strategy effects; A2, knowledge mining: correlation rule mining, clustering analysis and pattern recognition are performed on the group data to discover new breeding rules; A3, knowledge fusion and verification: fuse the newly discovered rules with existing knowledge and detect conflicts, and verify through statistical significance test or expert review mechanism; A4, knowledge update: inject the verified new knowledge into the rule base, case base or model base of the knowledge base.

3. The parrot breeding assistance system of claim 1, wherein: The individual habit data is obtained through user input and / or image recognition. In addition to basic parrot species, age and gender, individual preferences can be entered through a user interface or learned through long-term behavior observation; The environment data is obtained through environmental sensors, including temperature, humidity, light intensity, noise and air quality; The historical feeding data is obtained through automatic feeding records and user logs, including feeding duration, food types, food intake and parrot feeding response; The behavior performance data is obtained by image / sound acquisition equipment and analyzed by a behavior recognition algorithm, including activity amount, calling frequency and type, and specific behavior.

4. A parrot breeding assistance system according to claim 1, characterized in that: The multi-source data is cleaned, normalized, and time series aligned, an algorithm in the model library is called, a plurality of special models in the model library are used for dimension calculation, and then fusion is performed, and finally a multi-dimensional index system parrot digital portrait is output; the dimension calculation specifically includes a physiological comfort index, a psychological health score, a nutrition and health prediction, and a training responsiveness.

5. The parrot breeding auxiliary system according to claim 4, characterized in that: The physiological comfort index calculation specifically comprises: calling a lightweight time series classification model or a rule engine, inputting real-time environmental data including temperature, humidity and wind speed, and real-time posture features, the model comparing the input features with the best comfort interval of the species in the knowledge base, calculating the deviation of the current value of the temperature factor, humidity factor, wind speed factor and posture factor from the best comfort interval, and mapping it into a score of 0-100, with 100 points representing the best state; The psychological health score calculation specifically comprises: calling an abnormal behavior detection model such as an unsupervised model based on isolation forest or autoencoder, inputting abnormal behavior frequency in a time window such as the number of stereotyped pacing and the duration of feather picking, the model learning the baseline data of the parrot in the normal behavior mode, calculating the deviation of the current behavior data from the baseline, the greater the deviation, the lower the psychological health score, and the case library in the knowledge base provides prior knowledge, such as observing feather picking behavior for three consecutive days, the psychological health score weight is doubled; The nutrition and health prediction calculation specifically comprises: calling a classification model such as logistic regression and gradient boosting tree, or a time series prediction model, inputting historical feeding rules including daily average food intake and feeding speed trend, body weight change curve and activity change, the model outputting one or more risk probabilities according to the input features, for example, the probability of developing underweight risk in the next two weeks is 30%, and the weight parameters of the model come from the training results of a large number of similar parrot data in the knowledge base; The training responsiveness calculation specifically comprises: calling a value function in reinforcement learning or a simple linear regression model, inputting historical training records such as success rate and average reaction delay for different instructions, the model calculating a response score for different training items such as standing on a pole and recognizing colors, the response score = basic success rate x weight 1 + (1 / average reaction time) x weight 2, and the weights are defined by the knowledge base, for example, the weight of reaction time is higher for new learning instructions; The calculation results of the above four dimensions, together with the static attributes of the parrot, species, age and gender, are combined into a structured feature vector, which is the parrot digital portrait at the current time, the parrot digital portrait is a structured data object, in addition to the four dimension quantitative indicators, it also contains timestamp, data confidence and other metadata, which together constitute the unique and unified digital basis for system decision-making.

6. A parrot breeding assistance system according to claim 5, characterized in that: The temperature factor score calculation specifically comprises: The optimal temperature interval [Tol, Toh] and tolerable range [Ttl, Tth] of the species are obtained from the knowledge base, and the score is calculated using the membership function: St = 100 if the current temperature Tc is in [Tol, Toh]; St linearly decreases from 100 to 0 if Tc is in [Ttl, Tol) or (Toh, Tth]; St = 0 if Tc is out of the tolerable range; the humidity factor and wind speed factor are calculated in the same way as the temperature factor; The posture factor score is calculated as follows: the system obtains a series of uncomfortable postures and their thresholds from the knowledge base, for example: mouth opening breathing frequency Pb, deduct points if it exceeds the threshold Thb; puffy hair coefficient Pf, deduct points if it exceeds the threshold Thf; Sp = max(0, 100-(Pb / Thb*50)-(Pf / Thf*50)), each posture accounts for 50 points, ensuring that Sp is not less than 0; the factor scores are weighted and fused to output a continuous score of 0-100.

7. A parrot breeding assistance system according to claim 1, characterized in that: The personalized breeding assistance strategy is formulated using a hybrid intelligent reasoning mechanism, specifically: First, case matching: calculate the similarity between the current digital portrait and the problem portrait in the case library. When the similarity calculation value is higher than the preset threshold, directly adopt or fine-tune the historical strategy to realize experience reuse; Second, if there is no matching case, start rule reasoning: match the current portrait data with the conditions in the rule base, and deduce a new execution strategy through logical chain to realize logical backup; The final generated personalized breeding assistance strategy is an executable instruction set.

8. A parrot breeding assistance system according to claim 1, characterized in that: The strategy execution module specifically includes: Environment control unit: used to control the breeding environment parameters; integrated with a cooperative control system of thermostat, humidifier, dehumidifier and full-spectrum LED lighting system, which can simulate the climate characteristics of different regions and different seasons; Automatic feeding unit: used for automatic food and water feeding; with a precise feeding mechanism of multiple bins, supporting the feeding of dry food, nuts and nutritional supplements, etc., which can realize precise feeding of timing, quantity and ratio according to the strategy, and has a material shortage alarm function; Behavior training unit: used for automatic behavior induction and training of parrots; including: drivable training tools: such as movable target rods controlled by steering wheels, instruction buttons with light prompting; automatic reward mechanism: small precise bait feeder linked with the automatic feeding unit, used to give positive reinforcement to the parrot immediately after completing the specified action.

9. A parrot breeding assistance system according to claim 1, characterized in that: The system adopts a cloud-edge collaborative architecture; wherein the knowledge base construction module and knowledge mining and updating engine are deployed on the cloud platform for global knowledge aggregation and evolution; the digital portrait module, strategy generation module and strategy execution module are deployed on the edge computing node for real-time processing and rapid response of individual parrot data.

Images