A high-quality black pig precise feeding amount analysis method and system based on big data

Abstract

The application discloses a high-quality black pig precise feeding amount analysis method and system based on big data, relates to the intelligentized technical field of livestock breeding, and forms a feeding amount analysis data set by acquiring black pig object data, pig house environment data, wet mixing material batch data, feeding operation data, trough residual material data, feeding feedback data and historical records, generates a wet mixing material effective feeding window based on the pig house environment data, the wet mixing material batch data and the feeding operation data, and the residual material of the current meal is first subjected to time boundary and state boundary constraints, and then enters the residual material state result and effective feeding amount record, so that the residual wet mixing material in the trough can be distinguished into effective residual material and non-countable residual material, and the deviation caused by the participation of overtime, dilution, layered caking or suspected deteriorated residual material in the next feeding cycle feeding amount analysis is reduced.

Description

Technical Field

[0001] This invention relates to the field of intelligent livestock farming technology, and in particular to a method and system for analyzing the precise feeding amount of high-quality black pigs based on big data. Background Technology

[0002] The fattening cycle of high-quality black pigs is relatively long, and their feed intake varies with age, weight range, feeding stage, pigsty environment, and feeding method. In semi-open or naturally ventilated pigsties in southern summers, high temperature and humidity, insufficient ventilation, and changes in drinking conditions can affect the feeding time distribution of high-quality black pigs. To improve palatability and reduce dust, some pig farms use wet-mixed feed, porridge feed, or a combination of wet and dry feeding methods.

[0003] Existing precise feeding rate analysis methods typically generate the feeding amount for the next feeding cycle based on age, body weight, stage diet, historical feed intake, environmental temperature and humidity, feed trough residue, image recognition results, or feeding records. Some methods also determine the appetite status of the pig herd based on the feeding completion time, the first feed shortage time, the amount of feed remaining at fixed times, feeding frequency, or feeding duration, and adjust subsequent feeding plans accordingly.

[0004] Under conditions of high temperature and humidity in summer and when wet feed is used, high-quality black pigs may reduce their feed intake during the hottest part of the day and compensate by eating more in the evening, at night, or in the early morning. If the appetite is judged solely based on the amount of feed left at a fixed time or the time of the first feed shortage, it is easy to mistake the shift in feeding time caused by heat stress for a true decrease in appetite, leading to a continuous reduction in subsequent feeding or abnormal meal distribution.

[0005] Meanwhile, the amount of wet feed entering the feed trough is affected by the feed-to-water ratio, the time the feed trough remains in the trough, the temperature and humidity of the pigsty, ventilation conditions, water output from the drinkers, and the pigs' rooting behavior. Remaining wet feed in the trough may exhibit conditions such as prolonged residence time, stratification and clumping, dilution by drinking water, or suspected spoilage. Visual identification or weight detection of remaining feed in the trough only indicates the presence of residual material; it does not directly indicate whether this residual material can still be counted as effective feed intake. If unfeedable residual feed is included in subsequent feed intake or effective feed intake, historical feeding records will deviate from actual intake and affect the feeding analysis results for the next feeding cycle.

[0006] Therefore, how to distinguish between effective feed intake, unaccountable feed intake, and feed intake shifts caused by heat stress within the effective feeding window of wet feed under high temperature and humidity conditions and wet feed feeding conditions in summer, and to form an effective feed intake record that can be used for analysis in the next feeding cycle, has become a technical problem that needs to be solved in the precision feeding analysis of high-quality black pigs. Summary of the Invention

[0007] This application provides a method and system for analyzing the precise feeding amount of high-quality black pigs based on big data, which solves the problem of feeding amount analysis deviation caused by miscounting of leftover feed in wet-mixed feed and misjudgment of feed migration due to heat stress.

[0008] This invention provides a method for analyzing the precise feeding amount of high-quality black pigs based on big data, which includes:

[0009] Acquire data on black pigs, pigsty environment, wet feed batches, feeding operations, feed trough residue, feed intake feedback, and historical records to form a feed intake analysis dataset. Historical records include historical feed intake feedback data, historical effective feed intake records, historical feed trough cleaning records, historical residue status results, and historical anomaly notes.

[0010] Based on pigsty environmental data, wet feed batch data, and feeding operation data, an effective feed intake window for wet feed is generated, representing the time and state boundaries at which the effective feed intake of wet feed can be counted. The start time of the window is determined based on the feeding time, and the end time of the window is determined based on the feed-to-water ratio, raw material composition, feed trough dwell time, temperature, humidity, ventilation, and historical residual feed status. The current meal is then matched with historical trough cleaning records, historical residual feed status results, and historical anomaly notes to obtain the window type.

[0011] Based on the data of residual material in the feed trough and the effective feeding window of the wet-mixed feed, the residual material status result is generated, and the effective residual material and the residual material that cannot be counted are distinguished.

[0012] Based on feeding operation data, leftover feed status results, and feeding feedback data, an effective feed intake record is generated;

[0013] Based on historical feeding feedback data, current pigsty environment data, and the time period of the current feeding session, determine the feeding time migration and output the feeding time migration marker;

[0014] Based on data on black pigs, historical effective feed intake records, feed residue status results, and feed intake period migration markers, a precise feed intake analysis result for the next feeding cycle is generated.

[0015] In some embodiments, the black pig object data corresponds to an individual black pig or a black pig pen object, and includes breed identification, age record, weight range record, feeding stage record, health marker and commercial grade target;

[0016] The pigsty environmental data includes temperature, humidity, ventilation, and time period markers;

[0017] The batch data of the wet feed includes feed batch identifier, feed-to-water ratio grade, raw material composition, preparation time, feeding time and dwell time in the feed trough;

[0018] The feeding operation data includes planned feeding amount, actual feeding amount, meal identification, replenishment record and trough cleaning record;

[0019] The data on residual material in the trough includes at least one of the following: the state of residual material presence, the appearance of residual material, the weight of residual material, and the state of manual inspection.

[0020] The feeding feedback data includes at least one of the following: feeding time period, feeding duration, water consumption status, changes in leftover feed, and abnormal remarks.

[0021] In some embodiments, when forming the feeding amount analysis dataset, object identifiers, meal identifiers, feed batch identifiers, and time identifiers are added to the data from automatic feeding equipment, environmental acquisition equipment, feed trough identification equipment, weighing records, drinking records, and manual inspection records.

[0022] When inconsistent data exists for the same meal, a data credibility tag is generated according to the data source, collection time, cleaning record, and manual confirmation status. The data credibility tag is then associated with and stored with the corresponding data. The data credibility tag includes at least one of the following: equipment consistency tag, equipment conflict tag, manual confirmation tag, and tag pending verification.

[0023] In some embodiments, when generating the residual material status result, a window status is generated based on whether the residual material is within the effective feeding window of the wet-mixed material, and a material status is generated based on at least one of the obtained residual material appearance status, residual material weight status, drinking water status, and cleaning record. The window status and the material status are combined to generate the residual material status result.

[0024] The status results of the remaining materials include at least one of the following: valid remaining materials, overdue remaining materials, suspected deteriorated remaining materials, drinking water diluted remaining materials, stratified and clumped remaining materials, and unreliable remaining materials. The remaining materials that cannot be counted include at least one of the following: overdue remaining materials, suspected deteriorated remaining materials, drinking water diluted remaining materials, and stratified and clumped remaining materials. The unreliable remaining materials are subject to manual confirmation.

[0025] In some embodiments, when generating the effective feed intake record, the actual feeding amount is used as the nominal feeding amount benchmark, the effective leftover feed is registered as the subsequent feed intake, the leftover feed that cannot be counted is registered as invalid residue, the material corresponding to the trough cleaning record is registered as trough cleaning loss, and the effective feed intake record is formed according to the meal identifier.

[0026] The effective feed intake record includes nominal feed intake, subsequent feed intake, ineffective residue, trough cleaning loss, effective feed intake, data reliability marker, and calculation time.

[0027] In some embodiments, when generating the feeding period migration marker, the historical feeding rhythm is used to characterize the temporal distribution of feeding occurrence time, feeding duration, feed residue change and water drinking status of the same black pig subject in multiple feeding cycles.

[0028] Extract the historical feeding rhythm of the same black pig subject under the same feeding stage and similar pig house environment conditions, and compare the feeding time, leftover feed status and drinking status of the current meal with the historical feeding rhythm to obtain the correlation between the current meal's incomplete effective feeding demand and pig house environment data, drinking status, health markers or abnormal remarks.

[0029] In some embodiments, the subsequent suitable meal is the meal after the current meal and the corresponding pig house environmental data does not indicate high temperature and humidity and insufficient ventilation.

[0030] When the comparison results show that the unfulfilled effective feeding demand of the current meal is concentrated in the high temperature and high humidity period, and there are compensatory feeding records or the drinking status is marked as normal in the subsequent suitable meal, a high temperature temporary feeding mark is generated.

[0031] An abnormal appetite marker is generated when the comparison results show an increase in uneaten food in consecutive meals, accompanied by abnormal drinking status, abnormal health markers, or abnormal remarks.

[0032] In some embodiments, when generating the precise feeding amount analysis results, a basic feeding amount is generated based on black pig object data, feeding stage, historical effective feed intake records and commercial grade targets. Unaccountable feed is deducted based on the feed status results, effective feed is retained and trough cleaning losses are recorded, and meal allocation results are adjusted based on feeding time migration markers.

[0033] Unfulfilled effective feeding needs corresponding to the high temperature-induced temporary withholding of feeding mark will be included in subsequent suitable meals, while unfulfilled effective feeding needs corresponding to the abnormal appetite mark will be subject to conservative feeding treatment; the results of the precise feeding amount analysis include recommended feeding amount, meal allocation results, uneaten feeding treatment results, trough clearing prompts, supplementary feeding prompts, and recalculation trigger conditions.

[0034] Secondly, embodiments of the present invention provide a high-quality black pig precision feeding analysis system based on big data, including a data access unit, a feeding data warehouse, a wet feed mixing window generation unit, a leftover feed status analysis unit, an effective feed intake calculation unit, a feeding time period migration analysis unit, and a feeding result generation unit.

[0035] The data access unit is used to receive black pig object data, pig house environment data, wet feed batch data, feeding operation data, feed trough residue data, feeding feedback data and historical records, and form a feeding amount analysis dataset;

[0036] The feeding data warehouse stores historical records, historical feeding rhythms, and commodity grade targets. The historical records include historical feeding feedback data, historical effective feed intake records, historical feed clearing records, historical leftover feed status results, and historical anomaly notes. The wet feed window generation unit generates a wet feed effective feeding window. The leftover feed status analysis unit generates leftover feed status results. The effective feed intake calculation unit generates effective feed intake records. The feeding period migration analysis unit generates feeding period migration markers. The feeding amount result generation unit generates accurate feeding amount analysis results for the next feeding cycle.

[0037] In some embodiments, a traceability output unit is further included. The traceability output unit is used to associate and output the precise feeding amount analysis results, adjustment reason codes, data sources, object identifiers, meal identifiers, and feed batch identifiers, and write them back to the feeding data warehouse. The adjustment reason codes include at least one of the following: wet feed window reason codes, leftover feed status reason codes, feeding period migration reason codes, trough cleaning loss reason codes, manual verification reason codes, and health abnormality reason codes. The feeding data warehouse will use the written-back data as part of the historical record to participate in the next feeding cycle update.

[0038] When the write-back data contains unaccountable leftover feed, pending verification marks, or recalculation trigger conditions, the feeding result generation unit reads the write-back data and regenerates the recommended feeding amount for each subsequent meal.

[0039] Through the above technical solution, the present invention can achieve at least the following beneficial effects:

[0040] This invention generates a feeding amount analysis dataset by acquiring data on black pigs, pig house environment, wet feed batches, feeding operations, feed trough residue, feeding feedback, and historical records. Based on the pig house environment data, wet feed batch data, and feeding operation data, an effective feeding window for wet feed is generated. The wet feed residue of the current meal is first constrained by time and state boundaries, and then enters the residue status result and effective feeding amount record. This allows the remaining wet feed in the feed trough to be distinguished into effective residue and non-recordable residue, reducing the deviation caused by the inclusion of overdue, diluted, stratified, or suspected spoiled residue in the feeding amount analysis of the next feeding cycle.

[0041] By matching the current meal with historical trough cleaning records, historical leftover feed status results, and historical abnormal notes, the window type is obtained. The effective feeding window of wet-mixed feed can change accordingly with the feed-to-water ratio, feed trough dwell time, pig house temperature and humidity, ventilation status, and historical leftover feed abnormal status. This makes the accounting attributes of the leftover feed in the current meal correspond to the actual feedable status of the wet-mixed feed, reducing the error of judging the effectiveness of leftover feed based solely on a fixed duration or a fixed amount of leftover feed.

[0042] By generating effective feed intake records based on feeding operation data, residual feed status results, and feeding feedback data, the nominal feed intake, effective residual feed, non-recordable residual feed, and trough cleaning losses are recorded separately. This allows the historical effective feed intake records to reflect the actual feed intake calculation results and reduces the impact of non-recordable residual feed being mixed into historical feeding data on subsequent recommended feed intake and meal allocation results.

[0043] By analyzing historical feeding feedback data, current pigsty environmental data, and the feeding time of the current meal, the shift in feeding time is determined and a shift in feeding time marker is output. The failure to meet the effective feeding demand in the current meal can be distinguished from the delayed feeding during periods of high temperature and humidity. This allows the accurate feeding amount analysis results for the next feeding cycle to be generated by combining the results of the remaining feed status and the shift in feeding time marker, reducing the possibility of identifying the shift in feeding time caused by heat stress as a continuous reduction in feed due to decreased appetite. Attached Figure Description

[0044] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation on the scope of this application.

[0045] Figure 1 This is a flowchart illustrating the method for analyzing the precise feeding amount of high-quality black pigs based on big data in the embodiments.

[0046] Figure 2 This is a framework diagram of the high-quality black pig precision feeding analysis system based on big data in the embodiment. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0048] All terms used in this application (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0049] To facilitate understanding, the relevant terms and concepts involved in the embodiments of this application will be introduced below:

[0050] High-quality black pigs refer to individual black pigs or black pig pens that possess breed identification, age records, weight range records, feeding stage records, and commercial grade targets. Wet feed refers to feed made by mixing feed ingredients and water and then feeding it into the feed trough. Feed quantity analysis datasets refer to data sets stored in association with object identifiers, meal times, feed batches, and time periods.

[0051] Furthermore, the time stamp is used to characterize the chronological order of feeding, data collection, inspection, trough cleaning, and data write-back within the same meal. Feeding time is based on the actual completion time recorded in the feeding operation data record; data collection time is based on the collection time corresponding to the environmental data collection equipment, feed trough identification equipment, weighing records, and drinking water records; manual inspection time is based on the submission time of the manual inspection record; and trough cleaning time is based on the confirmation time of the trough cleaning record. When data from multiple sources belongs to the same meal, they are aligned according to the object identifier, meal identifier, feed batch identifier, and time stamp. When multiple records exist from the same source within the same meal, the record closest to the current calculation time and with a valid data reliability marker is used in the analysis of the current meal.

[0052] Pig house environmental data is used to characterize the current temperature, humidity, ventilation, and time period of the pig house. Wet feed batch data characterizes the batch status of the same wet feed from preparation and feeding to its residence in the feed trough. Specifically, the feed-to-water ratio level is a level record in the wet feed batch data used to characterize the range of feed ingredient to water ratio. The feed-to-water ratio level is generated based on preparation records, feeding operation data, or manual confirmation records. Raw material composition characterizes the main feed ingredient categories and stage diet categories used in the same wet feed batch. The batch grouping identifier is an alternative matching field generated based on the raw material composition and feed-to-water ratio level, used to group wet feed records with similar material characteristics when the same feed batch identifier is missing in the historical records. Feed trough residence time is the length of time between the feeding time and the current observation time or trough cleaning time; when a trough cleaning record has been generated, the trough cleaning time is used as the end time of residence; when a trough cleaning record has not yet been generated, the current accounting time is used as the end time of residence.

[0053] Feed trough residue data is used to characterize the presence, appearance, weight, and manual inspection status of remaining material in the feed trough after a meal or at the time of observation. Feed intake feedback data is used to characterize the timing of feeding, duration of feeding, drinking status, changes in leftover feed, and any abnormal remarks of high-quality black pigs during the corresponding meal.

[0054] Example 1:

[0055] like Figure 1 As shown, this embodiment employs a method for analyzing the precise feeding amount of high-quality black pigs based on big data, including:

[0056] Step S1: Obtain black pig object data, pig house environment data, wet feed batch data, feeding operation data, feed trough residue data, feeding feedback data and historical records to form a feeding amount analysis dataset. The historical records include historical feeding feedback data, historical effective feed intake records, historical feed trough cleaning records, historical residue status results and historical abnormal remarks.

[0057] Step S2: Based on the pig house environmental data, wet feed batch data, and feeding operation data, generate the effective feeding window for wet feed, which represents the time boundary and state boundary of the effective feed intake for the current meal. The start time of the window is determined based on the feeding time in the wet feed batch data, and the end time of the window is determined based on the feed-to-water ratio, raw material composition, feed trough dwell time in the wet feed batch data, as well as the temperature, humidity, ventilation, and historical feed status results in the pig house environmental data. The current meal is then matched with historical trough cleaning records, historical feed status results, and historical anomaly notes to obtain the window type for the current meal.

[0058] Step S3: Generate the residual material status result based on the residual material data in the feed trough and the effective feeding window of the wet-mixed material, and distinguish between effective residual material and residual material that cannot be counted.

[0059] Step S4: Generate an effective feed intake record based on feeding operation data, leftover feed status results, and feeding feedback data;

[0060] Step S5: Determine the feeding time migration based on historical feeding feedback data, current pigsty environment data, and the feeding time of the current meal, and output the feeding time migration marker;

[0061] Step S6: Based on the black pig data, historical effective feed intake records, leftover feed status results, and feed period migration markers, generate the precise feed intake analysis results for the next feeding cycle.

[0062] The effective feed intake window for wet-mixed feed is used to define the time and state boundaries for counting the effective feed intake of wet-mixed feed in the same meal. The start time of the window is determined based on the feeding time, and the end time is determined based on the feed-to-water ratio, raw material composition, temperature, humidity, ventilation, feed trough dwell time, and historical residual feed status results. The residual feed status allowable set is used to define the residual feed status that can be counted in the effective feed intake. The residual feed status is jointly determined by the residual feed appearance status, residual feed weight status, drinking status, and trough cleaning records. The window generation reason codes include environmental humidity and heat reason codes, feed-to-water ratio reason codes, feed trough dwell time reason codes, historical anomaly reason codes, and manual verification reason codes. When matching the current meal with historical records, similar meal records are extracted according to feed batch identifier, meal identifier, time period marker, and pig house environmental data. The window type is determined based on the trough cleaning records, historical residual feed status results, and historical anomaly remarks in the similar meal records.

[0063] In one implementation, the black pig object data corresponds to individual black pigs or black pig pens, and includes breed identification, age records, weight range records, feeding stage records, health markers, and commercial grade targets; pig house environment data includes temperature status, humidity status, ventilation status, and time period markers; wet feed batch data includes feed batch identification, feed-to-water ratio level, raw material composition, preparation time, feeding time, and feed trough dwell time; feeding operation data includes planned feeding amount, actual feeding amount, meal identification, supplementary feeding record, and feed trough cleaning record; feed trough residual data includes at least one of the following: residual feed existence status, residual feed appearance status, residual feed weight status, and manual inspection status; feeding feedback data includes at least one of the following: feeding occurrence time, feeding continuity status, drinking status, residual feed change status, and abnormal remarks.

[0064] In one implementation, when forming a feed quantity analysis dataset, object identifiers, meal identifiers, feed batch identifiers, and time identifiers are added to data from automatic feeding equipment, environmental acquisition equipment, feed trough identification equipment, weighing records, drinking records, and manual inspection records. When inconsistent data exists for the same meal, a data credibility tag is generated according to the data source, acquisition time, trough cleaning record, and manual confirmation status. The data credibility tag is then associated with and stored with the corresponding data. The data credibility tag includes at least one of the following: equipment consistency tag, equipment conflict tag, manual confirmation tag, and tag pending verification.

[0065] Data reliability tags are used to characterize the degree of consistency among data from multiple sources within the same feeding period. The data used to generate data reliability tags includes data source category, collection time, equipment status, trough cleaning record, manual verification status, and historical conflict records. Equipment consistency tags correspond to situations where trough identification data, weighing records, and feeding operation data match within the same feeding period; equipment conflict tags correspond to situations where at least two types of equipment data conflict within the same feeding period; manual verification tags correspond to situations where manual inspection records confirm the status of leftover feed or trough cleaning; and tags pending verification correspond to situations where equipment data is missing, collection time is misaligned, or manual inspection records are missing. Data reliability tags are included in the feed intake analysis dataset along with the corresponding feeding period data and participate in the generation of leftover feed status results and effective feed intake records.

[0066] For example, the data credibility weight is the numerical record corresponding to the data credibility tag when participating in the weighted statistics, with a value range of 0 to 1. The weight corresponding to the manual confirmation tag is higher than that of the equipment consistency tag, the weight corresponding to the equipment consistency tag is higher than that of the equipment conflict tag, and the weight corresponding to the equipment conflict tag is higher than that of the pending review tag. The data corresponding to the equipment conflict tag and the pending review tag can participate in risk collection and review prompt generation; when used for threshold recalibration, only the data corresponding to the manual confirmation tag or the equipment consistency tag is used. When both the manual confirmation tag and the equipment conflict tag exist in the current meal, the residual material status or cleaning status corresponding to the manual confirmation tag is used as the accounting basis for the current meal.

[0067] In one embodiment, when generating the residual material status result, a window status is generated based on whether the residual material is within the effective feeding window of the wet-mixed material, and a material status is generated based on at least one of the obtained residual material appearance status, residual material weight status, drinking water status and cleaning record. The window status and the material status are combined to generate the residual material status result.

[0068] Unacceptable surplus materials include at least one of the following: overdue surplus materials, suspected deteriorated surplus materials, surplus materials diluted with drinking water, and stratified / clumped surplus materials. Unreliable surplus materials will be subject to manual verification.

[0069] In one implementation, the residual material status result is used to characterize the accounting attributes of the remaining wet-mixed feed in the feed trough during the corresponding meal. Valid residual material corresponds to the remaining material that is within the valid feeding window of the wet-mixed feed and whose appearance, weight, water consumption, and trough cleaning records meet the allowed set of residual material status. Expired residual material corresponds to the remaining material that remains in the feed trough after the window's end time. Suspected spoiled residual material corresponds to the remaining material that, according to manual inspection or image recognition results, shows signs of sourness, stickiness, mold, or off-odor. Water-diluted residual material corresponds to the remaining material associated with abnormal water consumption, water accumulation in the feed trough, or sudden changes in weight. Stratified and clumped residual material corresponds to the remaining material whose appearance shows water-material separation, lumpy aggregation, or hardened surface. Unreliable residual material corresponds to the remaining material whose data reliability is marked with an equipment conflict marker or a pending verification marker.

[0070] In one implementation, when generating an effective feed intake record, the actual feeding amount is used as the nominal feeding amount benchmark. Effective leftover feed is registered as subsequent feedable amount, and feed that cannot be counted as invalid residue is registered. The material corresponding to the trough cleaning record is registered as trough cleaning loss, and an effective feed intake record is formed according to the meal identifier. The effective feed intake record includes nominal feeding amount, subsequent feedable amount, invalid residue, trough cleaning loss, effective feed intake, data reliability marker, and calculation time.

[0071] In one implementation, the effective feed intake record is used to distinguish between the nominal feed intake, subsequent feed intake, ineffective residue, cleaning loss, and effective feed intake within the same meal. The nominal feed intake corresponds to the actual feed intake recorded by automatic feeding equipment or manual feeding. Subsequent feed intake corresponds to the remaining material marked as effective feed. Ineffective residue corresponds to the remaining material marked as expired feed, suspected spoiled feed, feed diluted with drinking water, or stratified / clumped feed. Cleaning loss corresponds to the material removed from the cleaning record. The effective feed intake is generated based on the record relationships between the nominal feed intake, subsequent feed intake, ineffective residue, and cleaning loss, and is stored in association with the object identifier, meal identifier, feed batch identifier, and calculation time.

[0072] When invalid residue and cleaning loss refer to the same remaining material, the amount of material confirmed by the cleaning record is recorded as cleaning loss, and the corresponding remaining material status mark is retained in invalid residue; the subsequent feedable amount is recorded as the remaining material amount that can be fed in the next meal, and is not included in the effective feedable amount of the current meal.

[0073] In one implementation, when generating precise feeding amount analysis results, a basic feeding amount is generated based on black pig target data, feeding stage, historical effective feed intake records and commercial grade target. Unaccountable feed is deducted based on the feed status results, effective feed is retained and trough cleaning loss is recorded. Meal allocation results are adjusted based on feeding time migration markers.

[0074] The next suitable meal is the meal after the current meal that does not show high temperature and humidity and insufficient ventilation in the corresponding pig house environment data. The unfulfilled effective feeding needs corresponding to the high temperature temporarily withholding feeding mark are included in the next suitable meal. The unfulfilled effective feeding needs corresponding to the abnormal appetite mark are included in the conservative feeding treatment. The results of the precise feeding amount analysis include the recommended feeding amount, meal allocation results, uneaten feed treatment results, trough cleaning prompts, supplementary feeding prompts, and recalculation trigger conditions.

[0075] Furthermore, conservative feeding is a process of limiting, retaining, or reviewing the incomplete effective feeding needs of meals corresponding to abnormal appetite markers, pending verification markers, or abnormal manual inspections. The inputs to conservative feeding include object identifier, meal identifier, incomplete effective feeding needs, drinking status, health marker, abnormal remarks, leftover feed status results, and data reliability marker; the outputs include conservative feeding results, review prompts, and write-back data. The conservative feeding results are used to temporarily exclude incomplete effective feeding needs from the increased feeding amount of subsequent suitable meals; the review prompts are used to correlate with manual inspection records or health marker updates; and the write-back data is used to update historical effective feeding records, historical leftover feed status results, and historical feeding rhythms.

[0076] Precise feed quantity analysis results are used to output feeding control data for the next feeding cycle. Recommended feed quantity is generated based on black pig population data, feeding stage, historical effective feed intake records, and commercial grade targets. Meal allocation results are used to determine the distribution of the recommended feed quantity across meals in the next feeding cycle. Uneaten feed handling results are used to record the transfer, retention, or conservative feeding treatment of unmet effective feed intake needs. Trough cleaning prompts are generated in association with ineffective residue and trough cleaning losses. Refeeding prompts are generated in association with subsequent available feed quantity, feeding period migration markers, and meal allocation results. Recalculation trigger conditions include changes in pig house environmental conditions, changes in uneaten feed status results, changes in data reliability markers, updates to trough cleaning records, and updates to anomaly notes.

[0077] In an optional implementation of Example 1, when the wet feed window generation unit matches the reading results with historical trough cleaning records, historical residual feed status results, and historical anomaly notes, it uses the feed batch identifier, meal identifier, and time period marker of the current meal as the main index to extract candidate records of the same black pig object within the historical backtracking period from the feeding data warehouse. The black pig object refers to an individual black pig or a black pig pen object. When there is no completely identical feed batch identifier in the historical records, a batch group identifier generated by the raw material composition and feed-water ratio level is used for substitution matching, so that wet feed records from different batches but with similar material characteristics can enter the candidate range. The candidate records also retain the corresponding historical trough cleaning records, historical residual feed status results, historical anomaly notes, and data reliability markers.

[0078] The similarity of operating conditions between historical candidate records and the current meal is calculated based on temperature, humidity, ventilation, and feed-to-water ratio. The calculation method for operating condition similarity is as follows: ,in, Indicates the first Similarity between historical candidate records and the current meal's operating conditions; subscript This represents the index of historical candidate records; index 0 indicates the current meal. Indicates the matching weight of temperature state; Indicates the matching weight for humidity status; Indicates the matching weight of ventilation status; Indicates the matching weight of the feed-to-water ratio level; Indicates historical temperature status Compared with the current temperature state The matching coefficient between them; Indicates historical humidity status Compared with the current humidity status The matching coefficient between them; Indicates historical ventilation status With current ventilation status The matching coefficient between them; Indicates the historical material water ratio level Compared with the current feed-to-water ratio level The matching coefficients between the two. The value range of each matching coefficient is greater than or equal to 0 and less than or equal to 1. The higher value is used for the same state, the middle value is used for adjacent level states, and the lower value is used for conflicting states. The value range of each matching weight is greater than or equal to 0 and less than or equal to 1, and the sum of the weights is 1. The weights can be calibrated based on the contribution of the manual confirmation records in the most recent feeding stage to the window type determination result.

[0079] A set of similar working conditions is generated based on working condition similarity, feed batch identifier, batch group identifier, meal identifier, time period marker, and historical backtracking period. The method for generating the set of similar working conditions is as follows: ,in, Represents a set of similar working conditions; Indicates the first One historical candidate record; This represents the set of historical records in the feeding data warehouse that can be used for window type matching; Indicates the first Feed batch identifiers for each historical candidate record; This indicates the feed batch number for the current meal. Indicates the first Batch grouping identifier for each historical candidate record; This indicates the batch grouping identifier for the current meal. Indicates the first Meal identifiers for each historical candidate record; The meal identifier indicating the current meal; Indicates the first Time period markers for each historical candidate record; Indicates the time period marker for the current meal; Indicates the feeding time for the current meal; Indicates the first Feeding time for each historical candidate record; The historical backtracking period is the range of feeding cycles used to extract historical candidate records. It is determined according to the continuous feeding cycles that have formed historical records within the same feeding stage. When there are insufficient historical records within the same feeding stage, historical records from the end of the previous feeding stage with valid data credibility labels are used to participate in the extraction of candidate records. This represents the similarity threshold. The higher the similarity threshold, the closer the historical records entering the similar work condition set are to the current meal; the lower the similarity threshold, the more historical records enter the similar work condition set, but the greater the dispersion of the work conditions.

[0080] The similarity threshold is determined based on the window type misjudgment rate of historical manually confirmed samples, with a value range of greater than or equal to 0 and less than or equal to 1. When the window type misjudgment rate caused by the similar working condition set increases in the most recent feeding stage, the similarity threshold is raised. When the data credibility-weighted effective sample size of the similar working condition set is lower than the lower limit of the sample size, the similarity threshold is lowered to expand the range of candidate records. However, only one reduction is performed in the same meal, and the reduced similarity threshold must not be lower than the lower limit of the similarity threshold configured by the pig farm according to the feeding stage. If the similar working condition set regenerated after the reduction is still lower than the lower limit of the sample size, the current meal enters the review window. When the number of manually confirmed samples is insufficient, the similarity threshold of the previous feeding cycle is used.

[0081] After obtaining the set of similar working conditions, events are aggregated from the historical cleaning records, historical residual material status results, and historical anomaly notes in the set. An event indicating early cleaning occurs when the historical cleaning time is earlier than the end time of the corresponding window for that historical meal, and the corresponding material in the cleaning record is registered as cleaning loss. An event indicating suspected spoiled residual material occurs when the historical residual material status results show suspected spoiled residual material or when historical anomaly notes describe rancidity, off-odor, heat generation, or mold. An event indicating residual material diluted with drinking water occurs when the historical residual material status results show residual material diluted with drinking water, or when the drinking water status and the residual material weight increase abnormally occur simultaneously. An event indicating stratified and clumped residual material occurs when the historical residual material status results show stratified and clumped residual material, or when the manual inspection status is recorded as stratified, clumped, or lumpy.

[0082] The frequency of occurrence of various events is statistically weighted according to data reliability, and the statistical method is as follows: ,in, Indicates event type Weighted frequency of occurrence in a set of similar working conditions; Indicates the first The data credibility weight of each historical candidate record; Indicates the first Event types in historical candidate records The occurrence flag is set to 1 if it appears, and 0 if it does not appear. Indicates event type Smoothing factor; This indicates that the cleaning process occurred prematurely. This indicates a suspected incident involving spoiled leftover materials; This indicates an incident involving the dilution of leftover materials with drinking water; This indicates a layered clumping of residual material event. Data reliability weights are configured from high to low according to the reliability of manual confirmation markers, equipment consistency markers, equipment conflict markers, and pending verification markers, with values ​​ranging from greater than or equal to 0 to less than or equal to 1; the smoothing factor takes a value greater than 0 to reduce the excessive influence of a single abnormal record on the window type when the sample size is small.

[0083] The threshold values ​​for each event type are calibrated based on historical samples that have been calculated and manually verified within the same feeding stage, and are periodically recalibrated according to the feeding cycle. When the number of similar samples in a feeding cycle does not reach the preset sample size, the threshold value of the previous feeding cycle is used. The recalibration method for the threshold values ​​is as follows: ,in, Indicates event type The threshold used to determine the current meal window type; Indicates event type The upper limit of the judgment threshold; Indicates event type The lower limit of the judgment threshold; Indicates event type Threshold recalibration coefficients; Indicates event type The judgment threshold used in the previous feeding cycle; Indicates event type The weighted frequency of occurrence in historical samples is determined manually, with the quartile values ​​being the highest. The threshold recalibration coefficient ranges from 0 to 1. The lower and upper limits of the judgment threshold are used to limit the magnitude of threshold changes, ensuring that concentrated occurrences of anomalies within a single feeding cycle do not cause excessive shifts in window type judgment. Samples corresponding to the pending review marker and equipment conflict marker can participate in the risk statistics for the current meal, but are not included in the judgment threshold recalibration.

[0084] Based on the weighted frequency of events such as premature tank cleaning, suspected spoiled leftover materials, leftover materials diluted with drinking water, and stratified / clumped leftover materials, combined with the data reliability and environmental risk status of the current meal, the window type for the current meal is output. The window type is determined as follows: ,in, Indicates the window type for the current meal; This indicates that windows cannot be counted. This indicates the review window; This indicates a shortened edible window; This indicates a normal, ready-to-eat window; Indicates the environmental risk status of the current meal; This indicates an environmental risk state characterized by high temperature and humidity and insufficient ventilation.

[0085] Specifically, the environmental risk status is generated based on a combination of temperature, humidity, and ventilation status. Temperature, humidity, and ventilation status are each generated from the collected data records corresponding to the pig house environmental data, and are categorized according to the status levels configured in the pig farm at the corresponding feeding stage. A high-temperature, high-humidity, and insufficient-ventilation environmental risk status corresponds to a combination of a high-temperature level, a high-humidity level, and an insufficient-ventilation level. When there are missing data records, equipment conflict markers, or pending-verification markers for temperature, humidity, or ventilation status, the data reliability weight for the current meal is processed according to the corresponding marker, and the current meal is placed in the verification window or the processing path corresponding to the pending-verification marker.

[0086] This represents the lower limit of the weighted effective sample size for data reliability in a set of similar working conditions, and is a value greater than 0. This indicates the data credibility weight for the current meal. The data credibility weight is based on the data credibility label for the current meal and follows the formula... Generate with the same weight configuration caliber; This represents the lower limit of data reliability for the current meal, with a value range of greater than or equal to 0 and less than or equal to 1, and is calibrated based on the acceptable range of window type misjudgment rate in manually confirmed samples;

[0087] Furthermore, the window type misclassification rate is the proportion of samples with inconsistent automatically output window types and manually confirmed window types within the same feeding stage, out of the total number of manually confirmed samples. Manually confirmed samples must include at least the object identifier, meal identifier, feed batch identifier, automatically output window type, manually confirmed window type, and manual confirmation time. Samples used for window type misclassification rate statistics are based on records under the same parameter version identifier; when the parameter version identifier is updated, the updated samples are re-entered into the statistical scope of the corresponding feeding cycle.

[0088] This indicates the threshold used to determine the type of suspected spoiled leftover food in the current meal window. This indicates the threshold used in determining the type of meal window for the event of diluting leftover food with drinking water; This indicates the threshold used in determining the current meal window type for the layered clumping and leftover food event. This indicates the threshold used in determining the window type for the current meal if the early clearing event occurs. If the sample size is insufficient or the data reliability for the current meal is insufficient, the current meal will enter the review window. If the sample size and data reliability meet the determination conditions, and multiple risk conditions are met simultaneously, the window type for the current meal will be output according to the priority of non-inclusion window, review window, shortened feedable window, and normal feedable window.

[0089] After the window type is output, the wet feed window generation unit converts the window type into a window record. When the current meal is determined to be a normal feedable window, the window start time and end time are generated based on the feeding time and the duration of stay in the feed trough. The allowed set of residual feed status includes valid residual feed, as well as residual feed that can be classified as valid residual feed after manual confirmation. When the current meal is determined to be a shortened feedable window, the window end time is shortened based on the historical time distribution of early trough clearing events in the set of similar working conditions, and the window generation reason code is recorded as early trough clearing risk, drinking water dilution risk, or a combination of the two. When the current meal is determined to be a review window, the window start time and the provisional window end time are retained, but the allowed set of residual feed status is restricted to valid residual feed after manual confirmation, and a manual review reason code is output. When the current meal is determined to be an unaccountable window, the corresponding residual feed of the wet feed for that meal is registered as unaccountable residual feed, and the window generation reason code is recorded as suspected spoilage risk, high drinking water dilution risk, or a combination of environmental risks. This process enables historical trough cleaning records, historical residual material status results, and historical anomaly notes to jointly participate in the generation of the current meal window type, and ensures that the window type corresponds to the effective residual material, non-recordable residual material, and trough cleaning loss in the subsequent effective feed intake calculation.

[0090] Example 2:

[0091] Based on Example 1, this example provides a method for comparing the historical feeding rhythm of the same black pig species in the big data-based method for analyzing the precise feeding amount of high-quality black pigs.

[0092] When generating feeding period migration markers, historical feeding rhythms are used to characterize the temporal distribution of feeding occurrence time, feeding duration, feed residue changes, and water consumption status of the same black pig subject in multiple feeding cycles. Historical feeding rhythms of the same black pig subject under the same feeding stage and similar pig house environment conditions are extracted, and the feeding occurrence time, feed residue changes, and water consumption status of the current meal are compared with historical feeding rhythms to obtain the correlation between the current meal's incomplete effective feeding demand and pig house environment data, water consumption status, health markers, or abnormal remarks.

[0093] In one implementation, when the comparison results show that the unfulfilled effective feeding demand of the current meal is concentrated in the high temperature and high humidity period, and there are compensatory feeding records or the drinking status is marked as normal in the subsequent suitable meals, a high temperature temporary feeding mark is generated.

[0094] An abnormal appetite marker is generated when the comparison results show an increase in uneaten food in consecutive meals, accompanied by abnormal drinking status, abnormal health markers, or abnormal remarks.

[0095] In one embodiment of Example 2, when the feeding time migration analysis unit compares the current feeding time, water consumption status, leftover feed status, and pig house environment data with the historical feeding rhythm of the same black pig, it uses the object identifier in the black pig object data as the main comparison index. It then extracts historical feeding feedback data, historical effective feed intake records, historical leftover feed status results, pig house environment data, and health markers for the black pig object from the feeding data warehouse over multiple consecutive feeding cycles, forming an object-level rhythm sample. The object-level rhythm samples are aggregated according to meal identifiers and intraday time periods. The intraday time period includes at least one of morning, noon, evening, night, and early morning, or it can be divided into fixed time segments according to the actual feeding system of the pig farm.

[0096] During the data collection, samples that were registered as unacceptable surplus feed, unreliable surplus feed, or failed manual verification were removed. Subsequent compensatory feeding records corresponding to the high-temperature temporary feeding delay markers were retained, so that the historical feeding rhythm baseline reflects the regular feeding distribution of the same black pig subject under reliable feeding records.

[0097] The historical feeding rhythm baseline is formed by combining the historical feeding time period, feeding duration, feed residue changes, and water consumption. For any given time period within a day, the feeding time migration analysis unit statistically analyzes the proportion of historical effective feed intake to the total historical effective feed intake for that time period, the stable range of feeding duration, the magnitude of feed residue changes, and the proportion of normal water consumption, and uses this as the historical feeding rhythm baseline for the same black pig species. The historical feeding rhythm baseline is generated as follows: ,in, This indicates that the same black pig object was within the same day period. The corresponding historical baseline for foraging rate; This indicates the intraday period currently included in the statistics; Indicates the set of time periods within a day; Represents any intraday time period within the set of intraday time periods; This represents the index of historical samples in an object-level rhythm sample; This indicates that it belongs to the intraday time period. An object-level rhythm sample set; This indicates that it belongs to the intraday time period. An object-level rhythm sample set; Indicates the first The data reliability conversion factor for each historical sample; Indicates the first The historical effective feed intake corresponding to each historical sample; This represents the smoothing constant used to calculate the baseline feed intake ratio. The data reliability conversion factor is configured based on manual confirmation markers, equipment consistency markers, equipment conflict markers, and pending verification markers. Manual confirmation markers and equipment consistency markers correspond to higher values, while equipment conflict markers and pending verification markers correspond to lower values. The smoothing constant is a value greater than 0 to avoid abnormal ratios when historical effective feed intake is extremely low.

[0098] After obtaining the historical feeding rhythm baseline, the feeding time of the current meal is mapped to the corresponding intraday time period, and the non-feeding offset of the current meal relative to the historical feeding rhythm baseline is calculated. The non-feeding offset represents the degree to which the effective feeding demand of the current meal was not met within the normal feeding period, and its calculation method is as follows: ,in, Indicates the time period of the day in which the current meal occurs. Unfeeding offset; This indicates the intraday time period mapped to the current meal; This indicates that the same black pig object was within the same day period. The corresponding historical baseline for foraging rate; This indicates the daily basic effective feed intake requirement for the same black pig species during the current feeding stage. The daily basic effective feed intake requirement is generated based on black pig species data, historical effective feed intake records, and commercial grade targets. When the daily basic effective feed intake requirement for the current feeding stage has not yet been generated, the daily basic effective feed intake requirement that has been manually confirmed in the previous feeding cycle is used. When there is no manually confirmed daily basic effective feed intake requirement, a high-temperature temporary feed intake flag is not generated, and the current meal is output as a flag pending verification. Indicates the time period of the day in which the current meal occurs. The current effective feed intake has already been calculated. The lower limit of the non-feeding offset is 0; when the current effective feed intake reaches or exceeds the demand corresponding to the historical feeding rhythm baseline, the non-feeding offset is 0; when the current effective feed intake is lower than the demand corresponding to the historical feeding rhythm baseline, the non-feeding offset increases as the difference increases.

[0099] For example, the daily baseline effective feed intake requirement is a daily summary of the basic feed intake requirement of the same black pig species during the current feeding stage. The daily baseline effective feed intake requirement is generated based on the age record, weight range record, feeding stage record, historical effective feed intake record, and commercial grade target in the black pig species data, and is stored according to the species identifier and feeding stage record. The current effective feed intake is the sum of effective feed intake calculated from the effective feed intake record for the current meal or the current intraday period. The unfulfilled effective feed intake requirement is the difference between the daily baseline effective feed intake requirement allocated to the current intraday period according to the historical feeding rhythm baseline and the current effective feed intake; when the difference is less than 0, the unfulfilled effective feed intake requirement is recorded as 0.

[0100] To differentiate between environmentally-induced decreased feed intake and decreased feed intake due to health or abnormal appetite, the feed intake migration analysis unit generates environmental decreased feed intake intensity based on current pig house environmental data. Environmental decreased feed intake intensity is determined by temperature, humidity, and ventilation conditions, and is calculated as follows: ,in, Indicates the time period of the day in which the current meal occurs. Environmental mitigation intensity; This indicates the weight of the influence of temperature state on the intensity of environmental stagnation. Indicates the weight of the influence of humidity status on the intensity of environmental stagnation; Indicates the weight of the impact of ventilation status on the environmental mitigation intensity; Indicates the time period of the day in which the current meal occurs. Temperature condition risk conversion value; Indicates the time period of the day in which the current meal occurs. Temperature state; Indicates the time period of the day in which the current meal occurs. Humidity status risk conversion value; Indicates the time period of the day in which the current meal occurs. Humidity status; Indicates the time period of the day in which the current meal occurs. The risk conversion value of ventilation status; Indicates the time period of the day in which the current meal occurs. The ventilation status. The range of each risk conversion value is greater than or equal to 0 and less than or equal to 1. The higher the temperature, the higher the humidity, and the worse the ventilation, the larger the value. The range of each influence weight is greater than or equal to 0 and less than or equal to 1, and the sum of the weights is 1. The influence weights are determined based on the correspondence between historical high temperature and high humidity periods and feed delay records in the same pig house.

[0101] When there is a non-feeding offset in the current meal, the feeding period migration analysis unit reads the feeding feedback data of the suitable intraday period that has been collected and is located after the current meal in the actual time sequence, and calculates the compensation feeding amount in the subsequent suitable intraday period. When the data of the subsequent suitable intraday period has not been collected, the compensation feeding amount is temporarily set to 0 in the current meal judgment, and the meal is retained for the next meal rolling update. Before the data of the subsequent suitable intraday period is collected, the unfulfilled effective feeding demand is not included in the expanded feeding amount of the subsequent suitable meal.

[0102] Compensated feed intake is used to indicate whether the unmet effective feed intake demand of the current meal was made up by the same black pig species during a period of environmental easing. The calculation method is as follows: ,in, Indicates the time period of the day in which the current meal occurs. Subsequent compensatory feed intake; Indicates the suitable time of day following the current meal; This refers to the set of suitable intraday time periods that have been collected after the current meal in actual time sequence. Suitable intraday time periods include at least one of evening, night, and early morning. When early morning is after the current meal, early morning is the early morning time period of the next feeding cycle. Indicates suitable intraday time period The current effective feed intake has already been calculated. This indicates that the same black pig object was within a suitable intraday time period. The corresponding historical feeding ratio baseline. Compensation feeding only counts the portion exceeding the corresponding historical feeding ratio baseline; if the subsequent suitable intraday period does not exceed the historical feeding ratio baseline, the contribution of that period to compensation feeding is 0.

[0103] Specifically, the suitable intraday time period is the period after the current meal, provided that the corresponding pig house environment is not in a high-temperature, high-humidity, and poorly ventilated environmental risk state. Subsequent suitable meals are those that fall within the suitable intraday time period and have a meal identifier. Compensation feed intake records are those records of effective feed intake that have been generated in subsequent suitable meals, and where the current effective feed intake is higher than the required amount allocated to the corresponding historical feed intake ratio baseline. When the data for subsequent suitable meals has not yet been collected, the unfulfilled effective feed intake requirement of the current meal is retained in the uneaten feed processing results and participates in rolling updates after the subsequent suitable meals are calculated.

[0104] The feeding period migration marker is generated based on the non-feeding offset, environmental retardation intensity, subsequent compensatory feed intake, water consumption status, changes in uneaten feed, health markers, and abnormal remarks. The feeding period migration analysis unit outputs a high-temperature retardation feeding marker when the non-feeding offset of the current meal reaches the non-feeding determination threshold, the current meal is in a high-temperature, high-humidity environment with insufficient ventilation, the environmental retardation intensity reaches the environmental retardation threshold, and at least one of the following conditions is met: the compensatory feed intake during the subsequent suitable intraday period reaches the non-feeding offset corresponding to the compensatory feed intake ratio threshold, or the water consumption status is within the normal range.

[0105] The method for determining the "high temperature temporary feeding" marker is as follows: ,in, Indicates the time period of the day in which the current meal occurs. High temperature temporarily halts feeding; and The aforementioned definition will be used; This indicates the threshold for determining if food has not been consumed; Indicates the environmental pausing threshold; Indicates the threshold for compensatory feed intake ratio; Indicates the time period of the day in which the current meal occurs. The state of drinking water; This indicates normal drinking water intake. The threshold for determining no feeding is a value greater than 0 and higher than the change corresponding to the measurement resolution of the effective feeding volume record. The environmental deferral threshold ranges from 0 to 1 and is calibrated based on historical feeding delay records during periods of high temperature and humidity. The compensation feeding ratio threshold ranges from 0 to 1; the higher the value, the stricter the requirements for subsequent compensation feeding. When there are equipment conflict markers for temperature, humidity, or ventilation status without manual confirmation, a high-temperature deferral feeding marker is not output, and the incomplete effective feeding requirement is entered into a conservative feeding process.

[0106] When the comparison results show an increase in uneaten food in consecutive meals, and there are abnormalities in drinking status, health markers, or abnormal remarks, the feeding period migration analysis unit outputs an appetite abnormality marker. An increase in uneaten food in consecutive meals refers to a continuous increase in the current uneaten food weight or uneaten food change status relative to the corresponding historical feeding rhythm baseline over multiple consecutive meals, and the increase reaches the uneaten food abnormality threshold;

[0107] The abnormal threshold for leftover feed is calibrated based on the historical distribution of the variation range of leftover feed for the same black pig during the corresponding time period of the day, and a value greater than 0 is taken. When the range of fluctuation of leftover feed is expanded in the consistently marked samples of the equipment in multiple consecutive feeding cycles, the abnormal threshold for leftover feed is increased. When the number of manually confirmed abnormal appetite records increases, the abnormal threshold for leftover feed is decreased, and the decreased abnormal threshold for leftover feed is higher than the change corresponding to the minimum resolution of the weighing record.

[0108] Continuous leftovers increase the number of meals. Only consecutive meals where the data reliability weight reaches the lower limit of the current meal's data reliability, and the increase in the weight or change in leftover food relative to the corresponding historical feeding rhythm baseline reaches the leftover food abnormality threshold, are counted. If the increase in any meal is below the leftover food abnormality threshold, or if the meal is registered as having unreliable leftover food, fails manual verification, or is not confirmed by equipment conflict, the continuous counting is interrupted and restarted. Abnormal drinking status includes significantly reduced drinking, abnormally increased drinking, or missing drinking records accompanied by equipment conflict markers; abnormal remarks include lethargy, vomiting, diarrhea, refusal to eat, suspected illness, or abnormal records from manual inspection. The method for determining abnormal appetite markers is as follows: ,in, Indicates the time period of the day in which the current meal occurs. Markers of abnormal appetite; This indicates the number of times leftover food will be added up to the current meal. This indicates the threshold for increasing the number of meals with continuously leftover food. Indicates abnormal drinking water status; Indicates the time period of the day in which the current meal occurs. Corresponding health markers; Indicates a health abnormality marker; Indicates the time period of the day in which the current meal occurs. The corresponding exception notes exist. This indicates that an exception note exists.

[0109] The threshold for the number of consecutive uneaten feed meals is a positive integer and is set according to the feeding frequency of the pig farm. The higher the feeding frequency, the larger the threshold can be set for the number of meals; the lower the feeding frequency, the smaller the threshold should be set for the number of meals, so as to identify a continuous decline in feed intake within fewer meals.

[0110] For the same feeding period, the similarity matching weight, similarity admission threshold, judgment threshold for each event type, influence weight of environmental stagnation intensity, non-feeding judgment threshold, environmental stagnation threshold, compensation feeding ratio threshold, leftover feed abnormality threshold, and threshold for increasing the number of consecutive leftover feed feedings are all bound using the same parameter version identifier. The parameter version identifier includes the effective feeding cycle, applicable feeding stage, applicable black pig target range, and effective time. When generating window type, feeding period migration marker, and precise feeding amount analysis results for the current feeding period, the thresholds and weights under the same parameter version identifier are read; when there are intermediate results corresponding to different parameter version identifiers for the same feeding period, the effective time is used to overwrite the intermediate results corresponding to the parameter version identifier of the current feeding period. The parameter version identifier is refreshed after the feeding cycle ends and the historical sample write-back is completed, and remains unchanged within the same feeding cycle.

[0111] When the parameter version identifier is missing or the version switch is not completed, the parameter version identifier that has taken effect in the previous feeding cycle will be used, and the corresponding meal will be marked as pending review.

[0112] When the data reliability of the current meal is below the lower limit of data reliability, a "pending verification" mark is output, and the feeding amount for subsequent meals is not directly increased. When both the "high temperature temporarily withholds feed" mark and the "abnormal appetite" mark meet the output conditions, if there are "abnormal drinking," "abnormal health," or "abnormal remarks" marks for the current meal, the "abnormal appetite" mark is given priority as the output result, and the unfulfilled effective feeding needs are handled with conservative feeding. When the "high temperature temporarily withholds feed" mark meets the output conditions but the "abnormal appetite" mark does not meet the output conditions, if the current meal's drinking status is normal, the health mark is normal, there are no abnormal remarks, and there is a compensatory feed intake during the subsequent suitable daytime period, the "high temperature temporarily withholds feed" mark is used as the output result, and the unfulfilled effective feeding needs are transferred to the subsequent suitable meal.

[0113] Furthermore, the "pending verification" marker is a processing marker generated for the current meal due to insufficient data reliability, insufficient sample size, unconfirmed equipment conflicts, or missing manual inspection records. The "pending verification" marker is associated with the storage object identifier, meal identifier, feed batch identifier, triggering reason, relevant data source, and current calculation time. Meals with the "pending verification" marker are included in historical record storage, but they are not included in threshold recalibration before manual confirmation, nor are they used as confirmation samples for the high-temperature temporary feeding delay marker.

[0114] The high-temperature temporary feeding mark is updated once per feeding cycle, and the abnormal appetite mark is updated on a rolling basis according to the meal. When manual inspection confirms disease, feed trough contamination or equipment malfunction, the feeding period migration analysis unit adopts conservative feeding treatment, temporarily not incorporating the unfulfilled effective feeding needs into the subsequent suitable meal, and writes back the corresponding object identifier, meal identifier, abnormality cause and data source to the feeding data warehouse.

[0115] Example 3:

[0116] Based on Examples 1 and 2, this example provides a high-quality black pig precision feeding analysis system based on big data, such as... Figure 2 As shown, it includes a data access unit, a feeding data warehouse, a wet feed window generation unit, a leftover feed status analysis unit, an effective feed intake calculation unit, a feeding time period migration analysis unit, and a feed amount result generation unit;

[0117] The data access unit is used to receive black pig object data, pig house environment data, wet feed batch data, feeding operation data, feed trough residue data, feeding feedback data and historical records, and form a feeding amount analysis dataset;

[0118] The feeding data warehouse is used to store historical records, historical feeding rhythms, and commodity grade targets. Historical records include historical feeding feedback data, historical effective feed intake records, historical trough cleaning records, historical uneaten feed status results, and historical anomaly notes.

[0119] The wet feed window generation unit is used to generate an effective feeding window for wet feed.

[0120] The residual material status analysis unit is used to generate residual material status results;

[0121] The effective feed intake calculation unit is used to generate effective feed intake records;

[0122] The feeding time migration analysis unit is used to generate feeding time migration markers;

[0123] The feeding amount result generation unit is used to generate accurate feeding amount analysis results for the next feeding cycle.

[0124] The high-quality black pig precision feeding analysis system based on big data also includes a traceability output unit. The traceability output unit is used to associate and output the precision feeding analysis results, adjustment reason codes, data sources, object identifiers, meal identifiers, and feed batch identifiers, and write them back to the feeding data warehouse. The adjustment reason codes include at least one of the following: wet feed window reason code, leftover feed status reason code, feeding period migration reason code, trough cleaning loss reason code, manual review reason code, and health abnormality reason code. The feeding data warehouse will use the written-back data as part of the historical record to participate in the next feeding cycle update.

[0125] When the write-back data contains unaccountable leftover feed, pending verification marks, or recalculation trigger conditions, the feeding result generation unit reads the write-back data and regenerates the recommended feeding amount for each subsequent meal.

[0126] In one implementation, write-back data is used to create a continuous record of feed intake analysis. The write-back data includes precise feed intake analysis results, adjustment reason codes, object identifiers, meal identifiers, feed batch identifiers, data sources, residual feed status results, effective feed intake records, feeding period migration markers, and recalculation trigger conditions. Upon receiving the write-back data, the feeding data warehouse updates the historical records and historical feeding rhythms under the same object identifier. At the start of the next feeding cycle, the feed intake result generation unit reads the updated historical effective feed intake records, historical residual feed status results, and historical feeding rhythms, and combines them with the current meal data to generate new precise feed intake analysis results.

[0127] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0128] Furthermore, those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are meant to be within the scope of this application and form different embodiments. For example, all the embodiments above can be used in any combination. The information disclosed in this background section is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that such information constitutes prior art known to those skilled in the art.

Claims

1. A method for analyzing the precise feeding amount of high-quality black pigs based on big data, characterized in that, include: Acquire data on black pigs, pigsty environment, wet feed batches, feeding operations, feed trough residue, feed intake feedback, and historical records to form a feed intake analysis dataset. Historical records include historical feed intake feedback data, historical effective feed intake records, historical feed trough cleaning records, historical residue status results, and historical anomaly notes. Based on pigsty environmental data, wet feed batch data, and feeding operation data, an effective feed intake window for wet feed is generated, representing the time and state boundaries at which the effective feed intake of wet feed can be counted. The start time of the window is determined based on the feeding time, and the end time of the window is determined based on the feed-to-water ratio, raw material composition, feed trough dwell time, temperature, humidity, ventilation, and historical residual feed status. The current meal is then matched with historical trough cleaning records, historical residual feed status results, and historical anomaly notes to obtain the window type. Based on the data of residual material in the feed trough and the effective feeding window of the wet-mixed feed, the residual material status result is generated, and the effective residual material and the residual material that cannot be counted are distinguished. Based on feeding operation data, leftover feed status results, and feeding feedback data, an effective feed intake record is generated; Based on historical feeding feedback data, current pigsty environment data, and the time period of the current feeding session, determine the feeding time migration and output the feeding time migration marker; Based on data on black pigs, historical effective feed intake records, feed residue status results, and feed intake period migration markers, a precise feed intake analysis result for the next feeding cycle is generated.

2. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 1, characterized in that, The black pig object data corresponds to individual black pigs or black pig pen objects, and includes breed identification, age record, weight range record, feeding stage record, health mark and commercial grade target; The pigsty environmental data includes temperature, humidity, ventilation, and time period markers; The batch data of the wet feed includes feed batch identifier, feed-to-water ratio grade, raw material composition, preparation time, feeding time and dwell time in the feed trough; The feeding operation data includes planned feeding amount, actual feeding amount, meal identification, replenishment record and trough cleaning record; The data on residual material in the trough includes at least one of the following: the state of residual material presence, the appearance of residual material, the weight of residual material, and the state of manual inspection. The feeding feedback data includes at least one of the following: feeding time period, feeding duration, water consumption status, changes in leftover feed, and abnormal remarks.

3. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 2, characterized in that, When generating the feed amount analysis dataset, object identifiers, meal identifiers, feed batch identifiers, and time identifiers are added to the data from automatic feeding equipment, environmental acquisition equipment, feed trough identification equipment, weighing records, drinking records, and manual inspection records. When inconsistent data exists for the same meal, a data credibility tag is generated according to the data source, collection time, cleaning record, and manual confirmation status. The data credibility tag is then associated with and stored with the corresponding data. The data credibility tag includes at least one of the following: equipment consistency tag, equipment conflict tag, manual confirmation tag, and tag pending verification.

4. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 3, characterized in that, When generating the residual material status result, a window status is generated based on whether the residual material is within the effective feeding window of the wet-mixed material. A material status is generated based on at least one of the obtained residual material appearance status, residual material weight status, drinking water status, and cleaning record. The window status and the material status are combined to generate the residual material status result. The status results of the remaining materials include at least one of the following: valid remaining materials, overdue remaining materials, suspected deteriorated remaining materials, drinking water diluted remaining materials, stratified and clumped remaining materials, and unreliable remaining materials. The remaining materials that cannot be counted include at least one of the following: overdue remaining materials, suspected deteriorated remaining materials, drinking water diluted remaining materials, and stratified and clumped remaining materials. The unreliable remaining materials are subject to manual confirmation.

5. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 4, characterized in that, When generating the effective feed intake record, the actual feeding amount is used as the nominal feeding amount benchmark. The effective leftover feed is registered as the subsequent feed intake, the feed that cannot be counted is registered as invalid residue, the material corresponding to the trough cleaning record is registered as trough cleaning loss, and the effective feed intake record is formed according to the meal identification. The effective feed intake record includes nominal feed intake, subsequent feed intake, ineffective residue, trough cleaning loss, effective feed intake, data reliability marker, and calculation time.

6. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 5, characterized in that, When generating the feeding period migration marker, the historical feeding rhythm is used to characterize the temporal distribution of feeding occurrence time, feeding duration, feed residue change and drinking status of the same black pig in multiple feeding cycles; Extract the historical feeding rhythm of the same black pig subject under the same feeding stage and similar pig house environment conditions, and compare the feeding time, leftover feed status and drinking status of the current meal with the historical feeding rhythm to obtain the correlation between the current meal's incomplete effective feeding demand and pig house environment data, drinking status, health markers or abnormal remarks.

7. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 6, characterized in that, The next suitable meal is the meal after the current meal, provided that the corresponding pigsty environmental data does not indicate high temperature and humidity and insufficient ventilation. When the comparison results show that the unfulfilled effective feeding demand of the current meal is concentrated in the high temperature and high humidity period, and there are compensatory feeding records or the drinking status is marked as normal in the subsequent suitable meal, a high temperature temporary feeding mark is generated. An abnormal appetite marker is generated when the comparison results show an increase in uneaten food in consecutive meals, accompanied by abnormal drinking status, abnormal health markers, or abnormal remarks.

8. The method for analyzing the precise feeding amount of high-quality black pigs based on big data according to claim 7, characterized in that, When generating the precise feeding amount analysis results, the basic feeding amount is generated based on the black pig target data, feeding stage, historical effective feed intake records and commercial grade target. The unaccountable feed is deducted based on the feed status results, the effective feed is retained and the trough cleaning loss is recorded. The meal allocation results are adjusted based on the feeding time migration marker. Unfulfilled effective feeding needs corresponding to the high temperature-induced temporary withholding of feeding mark will be included in subsequent suitable meals, while unfulfilled effective feeding needs corresponding to the abnormal appetite mark will be subject to conservative feeding treatment; the results of the precise feeding amount analysis include recommended feeding amount, meal allocation results, uneaten feeding treatment results, trough clearing prompts, supplementary feeding prompts, and recalculation trigger conditions.

9. A big data-based precision feeding analysis system for high-quality black pigs, based on the big data-based precision feeding analysis method for high-quality black pigs as described in any one of claims 1 to 8, characterized in that, include: Data access unit, feeding data warehouse, wet feed window generation unit, leftover feed status analysis unit, effective feed intake calculation unit, feeding time period migration analysis unit, and feed amount result generation unit; The data access unit is used to receive black pig object data, pig house environment data, wet feed batch data, feeding operation data, feed trough residue data, feeding feedback data and historical records, and form a feeding amount analysis dataset; The feeding data warehouse is used to store historical records, historical feeding rhythms, and commodity grade targets. The historical records include historical feeding feedback data, historical effective feed intake records, historical trough cleaning records, historical uneaten feed status results, and historical anomaly notes. The wet feed window generation unit is used to generate an effective feeding window for wet feed. The residual material status analysis unit is used to generate residual material status results; The effective feed intake calculation unit is used to generate effective feed intake records; The feeding time migration analysis unit is used to generate feeding time migration markers; The feeding amount result generation unit is used to generate accurate feeding amount analysis results for the next feeding cycle.

10. The high-quality black pig precision feeding analysis system based on big data according to claim 9, characterized in that, It also includes a traceability output unit, which is used to associate and output the precise feeding amount analysis results, adjustment reason code, data source, object identifier, meal identifier and feed batch identifier, and write them back to the feeding data warehouse; The adjustment reason code includes at least one of the following: wet feed window reason code, leftover feed status reason code, feeding period migration reason code, trough cleaning loss reason code, manual review reason code, and health abnormality reason code. The feeding data warehouse will use the write-back data as part of the historical record to participate in the next feeding cycle update. When the write-back data contains unaccountable leftover feed, pending verification marks, or recalculation trigger conditions, the feeding result generation unit reads the write-back data and regenerates the recommended feeding amount for each subsequent meal.

Images