A method, system, and storage medium for livestock and poultry farming management.

Abstract

This invention discloses a method, system, and storage medium for livestock and poultry farming management, specifically relating to the field of livestock farming management data processing. The method includes reading the immunization record and pen baseline record of the current immunization batch, and writing the pen numbers under the same vaccine batch number and immunization method in the immunization record into an immunization table according to the execution order and execution time. This invention addresses the problems in existing technologies where immunization completion records cannot reflect whether the vaccine effectively reached the target pen area, and where short-term post-immunization stress reactions are difficult to distinguish from suspected missed immunizations and misaligned records in the early stages. It also addresses the issue of the existing technology failing to differentiate between short-term post-immunization stress reactions and suspected missed immunizations or misaligned records in the early stages.

Description

Technical Field

[0001] This invention relates to the field of livestock breeding management data processing technology, and more specifically, to a method, system, and storage medium for livestock and poultry breeding management. Background Technology

[0002] In large-scale poultry farming management, existing technologies mainly focus on immunization program planning, registration of disease prevention materials, and record-keeping of the execution process. Generally, immunization tasks are generated by batch, age, building, pen area, and vaccine type, and the immunization process is confirmed by personnel scanning codes, entering vaccine batch numbers, registering dosage, checking task completion, and reviewing mortality data.

[0003] In scenarios involving caged laying ducks, free-range meat ducks, or separate pen-raising of breeding ducks, immunization via drinking water, spraying, nasal drops, eye drops, or injection needs to be carried out continuously along water lines, personnel routes, or pen sequences. With numerous pens and high flock density within the same building, the effective usage time of diluted vaccines is limited. Water line pressure, spray coverage angle, personnel operation intervals, and overcrowding can all affect the actual state of vaccine delivery to the target pens. In such cases, simply relying on immunization completion records only proves that the immunization task has been registered as completed, not that the target pens have been effectively reached. Short-term decreases in water consumption, reduced activity, or clustering in some pens after immunization may be part of the stress recovery process after reaching the target area. Conversely, pens along the same path that do not show corresponding reactions may have missed immunizations, insufficient administration, or incorrect records. However, existing systems lack big data processing of immunization pathways, pen reactions, and post-immunization observation windows, often requiring investigation only after mortality increases, feed intake decreases, or production indicators fluctuate before the cause can be identified.

[0004] Therefore, the technical problem to be solved by this application is that in the existing poultry farming management, the immunization completion record cannot reflect whether the vaccine has effectively reached the target pen area, and it is difficult to distinguish between short-term stress reaction after immunization and suspected missed immunization or misplaced records in the early stage. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method, system, and storage medium for livestock and poultry farming management. By linking and differentiating the immunization records of the current immunization batch, the pen baseline records, and the post-immunization group records, a state tube is generated according to the reaction sequence. Then, observation, re-immunization verification, and record verification tasks are generated through Bayesian Zonotope trimming and double-layer horizon shrinking, thereby solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for managing poultry farming in animal husbandry, comprising:

[0007] S1. Read the immunization record and column baseline record of the current immunization batch, and write the column numbers of the same vaccine batch number and the same immunization method in the immunization record into an immunization table according to the execution order and execution time;

[0008] S2. After completing immunization in each column of the immunization table, read the post-immunization population record in each column, perform the same field difference between the post-immunization population record and the column baseline record, and sort them according to the collection time to generate a response sequence containing drinking water difference, activity difference, aggregation difference and abnormal population reduction difference.

[0009] S3. Perform joint posterior truncation based on the reaction sequence, generate reach items, omission / exemption pending items, and misplaced items according to the sequential relationship between the drinking water difference, activity difference, aggregation difference, abnormal reduction difference and execution time of each column area, and merge the reach items, misplaced items, and omission / exemption pending items sandwiched by the reach items before and after the execution sequence to generate a state tube.

[0010] S4. Based on the immunization table, perform Bayesian Zonotope pruning on the state tube. Take the same vaccine batch number, the same immunization method, adjacent execution order, and execution time range as batch number band, method band, order band, and time band, respectively. Find the intersection of the state tube with the batch number band, method band, order band, and time band item by item, and delete the state tube segments with empty intersections to generate the pruning tube.

[0011] S5. Based on the cut tube, perform double-layer horizon contraction, determine the front and back boundaries of the missed immunization status tube segments according to the execution priority, and perform inner layer transformation on the segment items according to the collection time. Generate re-immunization verification tasks for the retained missed immunization status tube segments, generate observation tasks for the reached status tube segments, and generate record verification tasks for the misattached status tube segments. Summarize the task type, column number, execution priority, vaccine batch number and immunization method to generate a task table.

[0012] In a preferred embodiment, S1 includes:

[0013] S1-1. Read the immunization record of the current immunization batch, merge multiple records with the same vaccine batch number, immunization method and column number into one column immunization item, and take the first execution time as the execution time of the column immunization item to generate an immunization item table.

[0014] S1-2. Read the immunization item table, sort the immunization items in the column area under the same vaccine batch number and the same immunization method according to the execution time, and sort them in ascending order by column area number when the execution time is the same, and generate a sequence table.

[0015] S1-3. Read the priority table and column reference record, associate the column number in the priority table with the column number in the column reference record, and write the vaccine batch number, immunization method, column number, execution priority, execution time, and column reference record into an immunization table.

[0016] In a preferred embodiment, S2 includes:

[0017] S2-1. Read the immunization table and post-immunization population records. Link the post-immunization population records with the same column number and whose collection time is after the execution time to the execution order and execution time in the immunization table. Subtract the execution time from the collection time and output the post-record table.

[0018] S2-2. Read the post-record table and the column baseline record. Subtract the drinking water value, activity value, aggregation value, and abnormal reduction value in the column baseline record from the drinking water value, activity value, aggregation value, and abnormal reduction value in the post-record table respectively to generate drinking water difference, activity difference, aggregation difference, and abnormal reduction difference. Write the difference value as rising position, falling position, and level position respectively when the difference value is greater than zero, less than zero, and zero, and output the difference position table.

[0019] S2-3. Read the differential position table, write the collection time of the first non-equal position under the same column area number as the reaction start time, write the drinking water difference, activity difference, aggregation difference and abnormal reduction difference under the same collection time as the reaction items in the field order of drinking water, activity, aggregation and abnormal reduction, and arrange the reaction items in the order of time interval to generate the reaction sequence.

[0020] In a preferred embodiment, S3 includes:

[0021] S3-1. Read the reaction sequence and immune table. Find the first non-zero collection time for the drinking water difference, activity difference, aggregation difference and abnormal decrease difference according to the column number. Write the earliest time among the first non-zero collection times as the reaction start time. Combine the first time of drinking water, the first time of activity, the first time of aggregation, the first time of decrease and the reaction start time according to the same column number to generate the first time item.

[0022] S3-2. Read the first time item and perform joint posterior truncation. For columns where the first drinking time exists, the first drinking time is earlier than the first activity time or the first aggregation time, the first reduction time is empty, or the first reduction time is different from the first drinking time, the first activity time, and the first aggregation time, write the column number, execution order, and execution time as the reach item; for columns where the first drinking time and the first activity time are empty, write the column number, execution order, and execution time as the missed immunization pending item; for columns where the reaction start time is earlier than the execution time of this column, and the reaction start time is equal to the execution time of another column in the same immunization table or the collection time after the execution time of another column, write the column number, execution order, and execution time as the misplaced item, and generate the posterior table;

[0023] S3-3. Read the posterior table and the immunization table and continue to perform joint posterior truncation. Define the reached item as the reached state, the misplaced item as the misplaced state, and define the column area with reached items in both the preceding and following columns and with undetermined missed immunization items in this column area as the missed immunization state. Then merge the reached state, missed immunization state and misplaced state to generate a state management system.

[0024] In a preferred embodiment, S4 includes:

[0025] S4-1. Taking a single segment in the status tube as the trimming object, search the immunization table for the column number that has the same vaccine batch number and the same immunization method as the trimming object, and take the current column, the previous column, and the next column according to the execution order; if all three exist, write the vaccine batch number, immunization method, execution order, and execution time of the three as batch number band, method band, order band, and time band; otherwise, write the trimming object as an isolated segment item and generate a three-column band item.

[0026] S4-2. Project the vaccine batch number, immunization method, execution order, and execution time in the cropping object to the batch number band, method band, order band, and time band in the three-column band, respectively; if all four items match, write the cropping object as a same-chain item; otherwise, write the cropping object as a broken-chain item and generate a signature item.

[0027] In a preferred embodiment, S4 further includes:

[0028] S4-3. Perform clamping post-clipping based on signature items and status tubes. When a co-chain item is established and the status type is "reached state", write the tube segment as a reach lock item. When a co-chain item is established, the status type is "missed state", and the execution order is between the previous and subsequent reach states, write the tube segment as a clamping lock item. When a co-chain item is established, the status type is "misaligned state", and the reaction start time is equal to the collection time after the execution time of another column area, write the tube segment as a borrowed time lock item. Otherwise, write the tube segment as a pruning item and generate a lock item.

[0029] S4-4. Execute Hamilton-Jacobi reverse deletion of lock items from the back position to the front position according to the execution order. If each leakage prevention and exemption status segment in the backtracking path has a clamping lock item, write the backtracking path as a pass item; otherwise, write the first leakage prevention and exemption status segment that lacks a clamping lock item and the leakage prevention and exemption status segments before and after it as deletion items.

[0030] S4-5. Perform closed-loop pruning using lock items, deletion items, and three-column items. Write the segments with access lock items, clamp lock items, or time-borrowing lock items that have not been written with deletion items and isolated segments into the pruning tube. Delete the segments with deletion items or isolated segments from the status tube. Write the segments with items to be pruned that have not formed access lock items, clamp lock items, or time-borrowing lock items as items to be verified, and generate the pruning tube.

[0031] In a preferred embodiment, S5 includes:

[0032] S5-1. Taking the unimmunized tube segment in the cut tube as the shrinkage object, find the first touched tube segment in the execution order and write it as the front boundary item, and find the first touched tube segment in the execution order and write it as the back boundary item. If both the front boundary item and the back boundary item exist and their vaccine batch number and immunization method are the same as the shrinkage object, write the shrinkage object as the segment item; otherwise, write it as the break boundary item and generate the outer boundary item.

[0033] S5-2. Link the segment item to the reaction sequence according to the column area number, and take two adjacent reaction items according to the collection time to perform the inner layer horizon transition. If the latter reaction item adds a drinking water difference relative to the former reaction item, and the subsequent adjacent reaction item adds an activity difference or aggregation difference, and no abnormal reduction difference is added at the same collection time, write the segment item as the reach transition item. If the reaction start time is earlier than the execution time of this column area and equal to the collection time after the execution time of another column area, write the segment item as the mis-linked transition item. Otherwise, write the segment item as the omission-free retention item and generate the inner layer state item.

[0034] In a preferred embodiment, S5 further includes:

[0035] S5-3. Perform double-layer horizon contraction on the outer boundary terms and inner state terms, write the boundary terms as segment deletion terms, write the reach transfer terms and mis-hanging transfer terms as resolution terms, write the omission exemption and retention terms as re-exemption terms, and remove the segment deletion terms and resolution terms from the trimming tube to generate the contraction tube.

[0036] S5-4. Combine the shrink tube and the cut tube according to the column area number, execution order, vaccine batch number and immunization method. Write the contact status tube segment as the observation task, write the re-immunization item as the re-immunization verification task, and write the misattached status tube segment as the record verification task. Summarize the task type, column area number, execution order, vaccine batch number and immunization method to generate a task table.

[0037] In a preferred embodiment, a livestock and poultry farming management system includes:

[0038] The batch linking module is used to read the immunization record and column baseline record of the current immunization batch, and write the column number of the same vaccine batch number and the same immunization method in the immunization record into an immunization table according to the execution order and execution time;

[0039] The response sorting module is used to read the post-immunization population records of each column after immunization is completed in each column of the immunization table, perform the same field difference between the post-immunization population records and the column baseline records, sort them according to the collection time, and generate a response sequence containing drinking water difference, activity difference, aggregation difference and abnormal reduction difference.

[0040] The state estimation module performs joint posterior truncation based on the reaction sequence, and generates reach items, omission / exemption pending items, and misplaced items according to the sequential relationship between the drinking water difference, activity difference, aggregation difference, abnormal reduction difference and execution time of each column area. It then merges the reach items, misplaced items, and omission / exemption pending items sandwiched by reach items before and after the execution sequence to generate the state management.

[0041] The constraint pruning module performs Bayesian Zonotope pruning on the state tubes based on the immunization table. It takes the same vaccine batch number, the same immunization method, adjacent execution order, and execution time range as batch number band, method band, order band, and time band, respectively. It finds the intersection of the state tube with the batch number band, method band, order band, and time band item by item, and deletes the state tube segments with empty intersections to generate the pruning tube.

[0042] The task generation module performs double-layer horizon contraction based on the trimmed tube, determines the front and back boundaries of the missed immunization status tube segments according to the execution priority, and performs inner-layer state transformation on the segment items according to the collection time. It generates re-immunization verification tasks for the retained missed immunization status tube segments, generates observation tasks for the reached status tube segments, and generates record verification tasks for the misattached status tube segments. It also summarizes the task type, column number, execution priority, vaccine batch number, and immunization method to generate a task table.

[0043] In a preferred embodiment, a storage medium for livestock and poultry farming management is provided, the storage medium storing a computer program that, when executed by a processor, implements the various steps of the farming management method as described in any one of claims 1-8.

[0044] The technical effects and advantages of this invention are as follows:

[0045] 1. Compare the immunization records with the changes in the population in the post-immunization areas using differential methods. Determine whether a post-immunization response has occurred in each area based on the order of occurrence of drinking water, activity, aggregation, and abnormal population reduction. This will help distinguish between areas that have been reached, suspected missed immunizations, and misplaced records, thus shifting immunization management from a task that has been completed to an early check of whether a response has occurred in the target areas.

[0046] 2. Merge duplicate immunization records under the same vaccine batch number, the same immunization method, and the same column number, and form an execution order according to the execution time and column number. Then link it with the column baseline record to reduce the interference of duplicate scanning and supplementary record entry on the immunization route sorting, so that subsequent differential calculation can directly correspond to the specific column.

[0047] 3. Subtract the pre-immunization baseline value from the corresponding column for the post-immunization drinking water value, activity value, aggregation value, and abnormal reduction value, and arrange them according to the collection time and the time interval after immunization to form a response sequence, so that different columns can be compared according to the post-immunization response process even if the execution time is different.

[0048] 4. Check the status segments within the three columns formed by the same vaccine batch number, the same immunization method, and adjacent columns. Trim status results that do not conform to the batch number, method, sequence, or execution time range to reduce the situation where suspected missed immunization results deviate from the actual immunization path.

[0049] 5. First, check the suspected missed exemption column with the preceding and following reached columns. Then, determine whether it has been converted to reached or recorded incorrectly according to the subsequent response sequence. Only retain the columns that have not yet formed an effective response explanation to generate re-examination verification tasks, so that observation, re-examination verification and record verification fall into the corresponding columns respectively. Attached Figure Description

[0050] Figure 1 This is a flowchart of the method steps of the present invention.

[0051] Figure 2 This is a schematic diagram of the system modules of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Refer to the instruction manual appendix Figure 1-2 The present invention provides a method for managing poultry farming in animal husbandry, comprising:

[0054] S1. Read the immunization record and column baseline record of the current immunization batch, and write the column numbers of the same vaccine batch number and the same immunization method in the immunization record into an immunization table according to the execution order and execution time;

[0055] In this embodiment, step 1 is used to organize the scattered immunization execution data within the current immunization batch into an immunization table that can be directly read for subsequent reaction analysis. During processing, duplicate immunization records generated by repeated scanning, supplementary recording, or multi-terminal synchronization in the same column area are first removed. Then, execution order is formed according to vaccine batch number and immunization method, so that the columns under the same batch of vaccines and the same immunization method have traceable sequential positions. Subsequently, the sorting result is linked with the column baseline record formed before immunization, so that subsequent step S2 can read the data before and after immunization according to the same column area and calculate the difference.

[0056] The implementation process includes the following steps:

[0057] In S1-1, the immunization records for the current immunization batch are generated from the aquaculture management terminal, personnel scanning records, waterline release records, or immunization task sheets. Each immunization record includes at least the immunization batch number, vaccine batch number, immunization method, pen area number, and execution time. After reading all immunization records under the current immunization batch, the system groups them using the vaccine batch number, immunization method, and pen area number as the merge key. When multiple records exist in the same group, they are first sorted in ascending order by execution time. The execution time of the first item in the sorted list is taken as the execution time of the immunization item for that pen area, and the multiple records in that group are merged into one immunization item for that pen area. If immunization records exist for the same pen area under different vaccine batch numbers or different immunization methods, different immunization items for each pen area are generated, and they are not merged. If the execution times are the same, the same time is retained as the execution time of the immunization item for the pen area. Through this process, duplicate scanning, duplicate data entry, and multi-terminal synchronization will not form duplicate pen areas in subsequent sorting, and the output result is an immunization item table.

[0058] In S1-2, the system reads the immunization item table and uses the vaccine batch number and immunization method as sorting and grouping keys to put immunization items in the same column under the same vaccine batch number and immunization method into the same sorting group. Within each sorting group, the items are first sorted in ascending order by execution time. If the execution times are the same, they are sorted in ascending order by column number. The sorted position is written as the execution order. If the column number is a single value, it is sorted by value size. If the column number consists of building number, row number, floor number, and column position number, the building number, row number, floor number, and column position number are compared in sequence until the sorting position is determined. Through this process, the execution order does not depend on the order of personnel input, but is determined by the immunization execution time and column number. The output result is a ranking table.

[0059] In S1-3, the system reads the priority table and the baseline record for each immunization zone. The baseline record for each immunization zone is derived from the population record prior to the immunization execution time of the corresponding immunization zone, and includes at least the baseline values ​​for drinking water, activity, aggregation, and abnormal population reduction. The system uses the immunization zone number as the linking field to link the immunization zone number in the priority table with the immunization zone number in the immunization zone baseline record. After successful linking, the system writes the vaccine batch number, immunization method, immunization zone number, execution priority, execution time, and immunization zone baseline record into the immunization table. If the immunization zone number in the priority table does not have a corresponding baseline record, the immunization item for that immunization zone is not written into the immunization table and is retained as a baseline missing record to prevent S2 from performing differential calculations when baseline values ​​are missing. The resulting immunization table becomes the input for S2 to read the population record after immunization and calculate the differences in drinking water, activity, aggregation, and abnormal population reduction.

[0060] Through the above processing, S1 organizes the immunization execution data from multiple sources, duplicate records, and unordered records into an immunization table organized by vaccine batch number, immunization method, column number, execution order, and execution time. The column baseline record before immunization is embedded in the same data structure, so that the subsequent reaction sequence generation has a direct value basis and also reduces the interference of duplicate immunization records on the execution order and status judgment.

[0061] In practical application: A duck shed uses drinking water immunization to administer batch A vaccine. Due to personnel scanning and system data entry, two records were generated in pen 3 at the management terminal. The system merges the two records with the same vaccine batch number, immunization method, and pen number into one pen immunization item, and takes the earlier execution time as the execution time for that pen. Subsequently, the immunization items of each pen under the same vaccine batch number and immunization method are arranged according to the execution time. If the execution times of pen 4 and pen 5 are the same, their execution order is determined according to the order of pen numbers. Finally, the immunization items of each pen are linked to the last group record before immunization to obtain an immunization table containing the execution order, execution time, and pen baseline record, which is used for subsequent post-immunization response differential calculation.

[0062] S2. After completing immunization in each column of the immunization table, read the post-immunization population record in each column, perform the same field difference between the post-immunization population record and the column baseline record, and sort them according to the collection time to generate a response sequence containing drinking water difference, activity difference, aggregation difference and abnormal population reduction difference.

[0063] In this embodiment, S2 is used to link the execution information in the immunization table with the post-immunization population record, and convert the population changes in different columns after different execution times into the same type of differential record. During processing, the immunization table and the post-immunization population record are first connected by the column number, excluding records before the execution time and at the execution time, and then the time interval is formed by subtracting the execution time from the collection time. Subsequently, the drinking water value, activity value, aggregation value, and abnormal reduction value in the post-immunization population record are subtracted from the corresponding baseline value in the column baseline record to obtain four types of differences. Finally, a response sequence is generated according to the difference direction and collection time to provide a basis for the response sequence for subsequent joint posterior truncation.

[0064] The implementation process includes the following steps:

[0065] In S2-1, the system reads the immunization table and post-immunization population records generated in S1. The post-immunization population records are saved according to the column number and collection time, and include at least the drinking water value, activity value, aggregation value, and abnormal reduction value. The system first uses the column number as a connection field to link the post-immunization population records with the records in the same column of the immunization table. After linking, the system compares the collection time of the post-immunization population records with the execution time in the immunization table. Only records whose collection time is later than the execution time are entered into the post-record table. Records whose collection time is equal to the execution time are recorded as the execution time record and are not entered into the post-record table. For each record entered into the post-record table, the system subtracts the execution time from the collection time to obtain the time interval. The time interval is used to represent the time interval between the completion of immunization of the current column and the completion of immunization of the current population record. Thus, even if different columns have different immunization execution times, they can participate in subsequent sorting and calculation according to their respective post-immunization processes.

[0066] In S2-2, the system reads the post-record table and the column baseline record, and retrieves the baseline value of the same column according to the column number. The difference calculation is performed by subtracting the column baseline field value from the post-immunization field value. That is, the drinking water difference is equal to the drinking water value in the post-record table minus the drinking water baseline value, the activity difference is equal to the activity value in the post-record table minus the activity baseline value, the aggregation difference is equal to the aggregation value in the post-record table minus the aggregation baseline value, and the abnormal reduction difference is equal to the abnormal reduction value in the post-record table minus the abnormal reduction baseline value. When the difference value is greater than zero, the system writes the field as an ascending bit; when the difference value is less than zero, the system writes the field as a descending bit; when the difference value is equal to zero, the system writes the field as a level bit. The abnormal reduction value only counts the reduction in the number of columns caused by death, removal of weak and sick individuals, and removal of abnormally isolated individuals after immunization, and does not include the transfer, removal, and merging of columns executed according to the production plan. Through this processing, the changes of the four types of groups are converted into a difference position table under the same data structure.

[0067] In S2-3, the system reads the difference table and processes the difference records of each column according to the column number. For the same column, the system finds the collection time of the first non-equal position among the drinking water difference, activity difference, aggregation difference, and abnormal reduction difference, and writes this collection time as the reaction start time. The non-equal position is formed by rising or falling positions, indicating that the field has deviated from the column baseline record at the collection time. Subsequently, at the same collection time, the system writes the drinking water difference, activity difference, aggregation difference, and abnormal reduction difference as reaction items in the field order of drinking water, activity, aggregation, and abnormal reduction. For multiple reaction items in the same column, they are arranged in ascending order of time interval, and if the time intervals are the same, they are arranged in the field order. The reaction sequence generated in this way retains the column number, execution order, execution time, collection time, time interval, reaction start time, and four types of differences, providing a data basis for S3 to determine the chronological relationship of changes in drinking water, activity, aggregation, and abnormal reduction.

[0068] Through S2 processing, the post-immunization population records are no longer used as scattered collected data for judgment, but are converted into a reaction sequence with execution order, execution time, time interval, differential direction and reaction start time. This allows the changes in the population in different columns after immunization to be compared according to their own execution process, and provides a direct basis for the generation of subsequent reach items, missed immunization pending items and misplaced items.

[0069] In practical application: In a certain laying duck house, pen 3 completed drinking water immunization at 9:00. Post-immunization population records were generated at 9:20, 9:40, and 10:00. The system only retains records after 9:00 and calculates time intervals of 20 minutes, 40 minutes, and 60 minutes respectively. If the drinking water value at 9:20 is lower than the drinking water baseline value, but the activity value, aggregation value, and abnormal reduction value are still the same as the baseline value, then the drinking water difference is written as a decreasing position, the other fields are written as equal positions, and 9:20 is written as the reaction start time. If the activity difference forms a decreasing position or the aggregation difference forms an increasing position at 9:40, then the system sorts the reaction items at 9:20 and 9:40 according to the time interval to obtain the reaction sequence of that pen, which is used to subsequently determine whether the post-immunization reaction starts with changes in drinking water.

[0070] S3. Perform joint posterior truncation based on the reaction sequence, generate reach items, omission / exemption pending items, and misplaced items according to the sequential relationship between the drinking water difference, activity difference, aggregation difference, abnormal reduction difference and execution time of each column area, and merge the reach items, misplaced items, and omission / exemption pending items sandwiched by the reach items before and after the execution sequence to generate a state tube.

[0071] In this embodiment, S3 is used to convert the differential changes in the reaction sequence into three types of status results: immune reach, suspected missed immunization, and misregistered record. During processing, the collection time of the first change of each field is extracted according to the column number to obtain the first time of drinking water, the first time of activity, the first time of aggregation, the first time of population reduction, and the reaction start time. Then, according to the order of the population reaction after immunization, joint posterior truncation is performed. Columns where the drinking water reaction occurs first and the abnormal population reduction has not formed synchronously are written as reach items. Columns where no drinking water reaction or activity reaction has formed are written as missed immunization pending items. Columns where the reaction occurs earlier than the execution time of this column and coincides with the execution time of another column are written as misregistered items. Finally, the status tube is formed by combining the front and back clamping relationship of the execution order.

[0072] The implementation process includes the following steps:

[0073] In S3-1, the system reads the reaction sequence generated by S2 and the immune table generated by S1, and searches for the time of difference change one by one using the column area number as the processing object. For the same column area, the collection time when the drinking water difference is not equal to zero for the first time is written as the drinking water first time, the collection time when the activity difference is not equal to zero for the first time is written as the activity first time, the collection time when the aggregation difference is not equal to zero for the first time is written as the aggregation first time, and the collection time when the abnormal decline difference is not equal to zero for the first time is written as the decline first time. If a certain field is always zero in the reaction sequence, the first time of that field is recorded as empty. Empty means that the field has not formed a non-zero difference, not that the record is missing. The system then selects the earliest time from the drinking water first time, activity first time, aggregation first time, and decline first time and writes it as the reaction start time. The drinking water first time, activity first time, aggregation first time, decline first time, and reaction start time are merged according to the same column area number to generate the first time item. The execution order and execution time in the immune table are retained along with the column area number for S3-2 to read.

[0074] In S3-2, the system reads the first time item and performs joint posterior truncation. The joint posterior truncation uses the first time of drinking water, the first time of activity, the first time of aggregation, the first time of phasing, the reaction start time, the execution order, and the execution time as the judgment fields. If the first time of drinking water exists and is earlier than the first time of activity or aggregation, and the first time of phasing is empty, or the first time of phasing is not equal to the first time of drinking water, the first time of activity, and the first time of aggregation, it indicates that the column area first formed an immunization response after drinking water, and then the activity or aggregation response occurred, and it did not occur at the same time as the abnormal phasing. The system writes the column number, execution order, and execution time of this column area as the reach item. If the first time of activity or aggregation is empty, the empty field is not included in the earlier comparison. If both the first time of activity and the first time of aggregation are empty, no reach item is formed. If the first time of drinking water is empty and the first time of activity is empty, the system writes the first time of activity, the execution order, and the execution time as the reach item. When the time is empty, it indicates that no drinking or activity response has occurred in the column area. The system will record the column number, execution order, and execution time of the column area as a pending missed immunization item. A pending missed immunization item only indicates that the response will be further truncated by the results of the preceding and following columns, and does not directly form a missed immunization status. If the response start time is earlier than the execution time of this column area, and the response start time is equal to the execution time of another column area in the same immunization table or equal to the collection time after the execution time of another column area, the system will record the column number, execution order, and execution time of the column area as an incorrectly registered item. The other column area is limited to the column area under the same vaccine batch number and the same immunization method. The later collection time is the collection time in the response sequence of this column area that is later than the execution time of the other column area but earlier in time. The response items, pending missed immunization items, and incorrectly registered items are used together to generate a post-test table.

[0075] In S3-3, the system reads the posterior table and the immunization table and continues to perform joint posterior truncation, converting the item results into status results. Items in the posterior table that have been reached are defined as reached status, and items that have been incorrectly linked are defined as incorrectly linked status. For items with pending missed immunizations, the system reads the execution order in the immunization table according to the same vaccine batch number and the same immunization method, searching the preceding and following columns of this column. If both the preceding and following columns have reached items, and this column has items with pending missed immunizations, then this column is defined as a missed immunization status; if neither the preceding nor following column has reached items... If a missed vaccination pending item is generated, it will not be converted to a missed vaccination status. If both incorrectly attached items and reached items are generated in the same column area, the incorrectly attached items will be retained and enter the incorrectly attached status first. The reached items will enter the reached status after the incorrectly attached items are subsequently trimmed and deleted. If both missed vaccination pending items and incorrectly attached items are generated in the same column area, the incorrectly attached items will be retained first, and the missed vaccination pending items will not generate a missed vaccination status. The system will merge the reached status, missed vaccination status, and incorrectly attached status into a status tube. Each segment in the status tube will include at least the column area number, execution order, execution time, vaccine batch number, immunization method, status type, and reaction start time.

[0076] Through the processing of S3, changes in drinking, activity, aggregation, and abnormal reduction in the reaction sequence are converted into state tubes with execution order constraints. The arrival state is derived from the order of the post-immunization reaction, the missed immunization state is derived from the clamping of the missed immunization pending item by the preceding and following arrival items, and the misalignment state is derived from the borrowing time relationship between the reaction start time and the execution time of other columns. This processing avoids relying solely on the change of a single field to make a state judgment, so that the subsequent S4 can trim the state tubes according to the vaccine batch number, immunization method, execution order, and execution time.

[0077] In practical applications: In a certain batch of drinking water immunization, both column 2 and column 4 showed drinking water difference first after immunization, followed by activity difference, and the abnormal decrease difference did not form at the same time as the two. The system wrote them as reach items respectively. Column 3 did not form drinking water first time and activity first time after immunization. The system initially wrote it as a missed immunization pending item. Since column 3 is located between column 2 and column 4, and both column 2 and column 4 have reach items, the system defined column 3 as a missed immunization state. If the reaction start time of column 5 is earlier than the execution time of column 5, and is the same as the collection time after the execution time of column 4, the system wrote column 5 as a mis-attached state. Finally, the reach status, missed immunization status, and mis-attached status are merged to generate a status tube.

[0078] S4. Based on the immunization table, perform Bayesian Zonotope pruning on the state tube. Take the same vaccine batch number, the same immunization method, adjacent execution order, and execution time range as batch number band, method band, order band, and time band, respectively. Find the intersection of the state tube with the batch number band, method band, order band, and time band item by item, and delete the state tube segments with empty intersections to generate the pruning tube.

[0079] In this embodiment, S4 is used to perform Bayesian Zonotope pruning on the state tube generated by S3, so that the arrival state, missed immunization state, and misattached state are subject to same-chain verification in four dimensions: vaccine batch number, immunization method, execution order, and execution time. During processing, a single tube segment in the state tube is first used as the pruning object, and the current column, the previous column, and the next column are taken from the immunization table to form a three-column item. Then, the pruning object is projected onto the batch number band, the method band, the order band, and the time band to obtain same-chain items or broken-chain items. Subsequently, the state type, the previous and previous arrival clamping relationship, and the borrowing time relationship are combined to form a locked item, and the missed immunization state tube segment lacking clamping support is eliminated by Hamilton-Jacobi reverse deletion, and finally the pruned tube is obtained.

[0080] The implementation process includes the following steps:

[0081] In S4-1, the system uses a single segment within the state pipe as the pruning object. Each segment includes at least the column number, execution priority, execution time, vaccine batch number, immunization method, state type, and reaction initiation time. The system retrieves column numbers in the immunization table that share the same vaccine batch number and immunization method as the pruning object, and uses the pruning object's column number as the current column. Under the same vaccine batch number and immunization method, the column with an execution priority one position lower than the current column is designated as the preceding column, and the column with an execution priority one position higher than the current column is designated as the next higher column. The current column is used as the next column; when the current column, this column, and the next column all exist, the system writes the set of vaccine batch numbers of the three as the batch number band, the set of immunization methods of the three as the method band, the set of execution order of the three as the order band, and writes the item with the earliest time to the item with the latest time among the three as the time band, generating a three-column band item; when the current column or the next column does not exist, the pruning object lacks three-column support, the system writes the pruning object as an isolated segment item, and generates a three-column band item containing the isolated segment item;

[0082] In S4-2, the system projects the vaccine batch number, immunization method, execution order, and execution time of the cropped object onto the batch number band, method band, order band, and time band in the three-column band, respectively. When the vaccine batch number belongs to the batch number band, the batch number dimension is hit; when the immunization method belongs to the method band, the method dimension is hit; when the execution order belongs to the order band, the order dimension is hit; when the execution time falls into the time band, the time dimension is hit. When all four dimensions are hit, the system writes the cropped object as a co-linked item; when any dimension is not hit, the system writes the cropped object as a broken-chain item. Orphan items, lacking a preceding or following column area, do not participate in the four-dimensional projection and are directly marked as orphans. The co-linked item, broken-chain item, and orphan item constitute a signature item, which serves as the input for subsequent locking and deletion.

[0083] In S4-3, the system performs clamping post-hoc pruning using signature items and state tubes. For tube segments with a valid signature and a state type of "reached state," the system writes the tube segment as a reach lock item. For tube segments with a valid signature and a state type of "missed state," the system searches the state tube for reach states with execution priority preceding and following the tube segment. If the tube segment's execution priority is between the preceding and following reach states, then the tube segment is written as a clamping lock item. For misaligned tube segments, the system searches the immunization table for another column with the same vaccine batch number and immunization method as the segment. It then takes the collection time in the segment's reaction sequence that is later than the execution time of the other column but earlier, as the subsequent collection time after the execution time of the other column. If the reaction start time equals this subsequent collection time, the segment is written as a time-borrowing lock item. If none of the above three conditions are met, the system writes the segment as a pending-pruning item. Lock items consist of reach lock items, clamp lock items, time-borrowing lock items, and pending-pruning items.

[0084] In S4-4, the system performs Hamilton-Jacobi reverse deletion from the next reachable lock item to the previous reachable lock item according to the execution order. During processing, the next reachable lock item serves as the reverse boundary. The system checks the missing / avoided state segments located between the next reachable lock item and the previous reachable lock item along the decreasing execution order. If each missing / avoided state segment in the backtracking path has a clamping lock item, the backtracking path is written as a pass item. If a missing / avoided state segment lacking a clamping lock item appears in the backtracking path, the first missing / avoided state segment lacking a clamping lock item encountered along the backtracking direction is written as the deletion center, and the previous and next missing / avoided state segments in the backtracking path of the deletion center are written as deletion items. If the deletion center does not have a previous or next missing / avoided state segment, only the existing segments are written as deletion items. This reverse deletion is used to exclude segments that, although initially written as missing / avoided states, lack front and rear reachable clamping supports.

[0085] In S4-5, the system performs closed-loop trimming using locked items, deleted items, and three-column items. Segments with access locks, clamped locks, or time-borrowed locks, but not yet written with deleted items or isolated segments, enter the trimming tube. Segments with deleted items or isolated segments do not enter the trimming tube. Segments with items to be trimmed but not yet having access locks, clamped locks, or time-borrowed locks are written as pending verification items. These pending verification items do not enter the trimming tube but are only saved as trimming records. Deletion here means exclusion in the trimming output and does not change the original record in the status tube. After the system completes processing all segments, it aggregates the segments entering the trimming tube to generate the trimming tube. Each segment in the trimming tube retains the column number, execution order, execution time, vaccine batch number, immunization method, status type, and reaction start time for S5 to perform double-layer horizon contraction.

[0086] Through the processing of S4, the segments in the state tube no longer enter the subsequent task generation solely based on the state results formed by S3. Instead, they first undergo same-chain pruning based on four dimensions: vaccine batch number, immunization method, execution order, and execution time. Then, they undergo post-hoc screening of reach locks, clamp locks, and borrow time locks. Missed immunization state segments also need to undergo Hamilton-Jacobi reverse deletion. This can exclude segments that lack support from the preceding and following columns, are separated from the same vaccine batch number or the same immunization method, or do not meet the borrow time relationship, so that S5 only processes the state results retained after pruning.

[0087] In practical applications: In a certain batch of drinking water immunization, column 3 is written as a missed immunization state by S3. The system takes column 3 as the clipping object and takes columns 2, 3, and 4 from the immunization table to form a three-column entry. If the vaccine batch number, immunization method, execution order, and execution time of column 3 all fall into the three-column entry, it is written as a co-linked entry. If both columns 2 and 4 have reachable states, and the execution order of column 3 is between the two, then column 3 is written as a clamping lock entry. The system then checks the path backward from the reachable lock entry of column 4 to the reachable lock entry of column 2. If column 3 has a clamping lock entry, then column 3 enters the clipping tube. If column 3 lacks a clamping lock entry, it is written as a deletion entry and does not enter the clipping tube.

[0088] S5. Based on the cut tube, perform double-layer horizon contraction, determine the front and back boundaries of the missed immunization status tube segments according to the execution order, and perform inner layer transformation on the segment items according to the collection time. Generate re-immunization verification tasks for the retained missed immunization status tube segments, generate observation tasks for the reached status tube segments, and generate record verification tasks for the misattached status tube segments. Summarize the task type, column number, execution order, vaccine batch number and immunization method to generate a task table.

[0089] In this embodiment, S5 is used to further compress the range of missed immunization status segments based on the trimmed tubes, and convert the status results still retained after trimming into immunization management tasks. During processing, the missed immunization status segments are first taken as the shrinkage target. The reach status segments on both sides of the segment before and after it are searched along the execution sequence in the trimmed tube to determine whether the missed immunization status segment is still within the reach segment under the same vaccine batch number and the same immunization method. Then, the reaction sequence of the segment column is read, and it is determined whether it has been converted into a reach reaction or a misaligned reaction according to the differential addition relationship between adjacent collection times. Then, the segments that have been resolved by reach reaction or misaligned reaction are excluded from the revaccination range, and the segments that have not yet formed a reaction interpretation are retained. Finally, the observation task, revaccination verification task, and record verification task are summarized into a task table.

[0090] The implementation process includes the following steps:

[0091] In S5-1, the system reads the trimming tube generated in S4 and only uses the segments with missed immunization status in the trimming tube as shrinkage targets. For each shrinkage target, the system starts from the execution order of the shrinkage target and searches forward along the execution order in the trimming tube for the first reached status segment, writing the found reached status segment as the preceding bound. Then, it searches backward along the execution order for the first reached status segment, writing the found reached status segment as the following bound. The search process is limited to segments with the same vaccine batch number and the same immunization method as the shrinkage target. If the preceding bound... If both the preceding and following boundary terms exist, and the vaccine batch number and immunization method of both the preceding and following boundary terms are the same as the contraction target, it indicates that the contraction target is still within the front and rear reach range of the same immunization chain, and the system writes the contraction target as a segment term; if the preceding term does not exist, the following boundary term does not exist, or the vaccine batch number or immunization method of either the preceding or following boundary term is different from the contraction target, the contraction target is written as a break term; the segment term and the break term together form the outer boundary term, which is used to limit the subsequent inner layer transition to only process the missed immunization state segments that are still sandwiched by the front and rear reach states;

[0092] In S5-2, the system links the segment items in the outer boundary items with the reaction sequences generated in S2 according to the column area number, and reads two adjacent reaction items in the same column area sequentially according to the acquisition time. The newly added drinking water difference here refers to the subsequent reaction item changing from a level position to an ascending or descending position relative to the previous reaction item, or the subsequent reaction item being the first non-level position in the drinking water field. The newly added activity difference, newly added aggregation difference, and newly added abnormal reduction difference are determined in the same way. If the subsequent reaction item adds a drinking water difference relative to the previous reaction item, and the subsequent reaction items adjacent to the acquisition time add activity differences or aggregation differences, and the acquisition time of the newly added drinking water difference and the acquisition time of the subsequent adjacent reaction items do not add abnormal reduction differences. If the segment in question is within the inner layer of data collection time, it indicates that the original missed immunization status segment has already formed a response at the inner layer data collection time level, and the system will write the segment item as a response transfer item. If the response start time of the segment item is earlier than the execution time of this column, and the response start time is equal to the data collection time after the execution time of another column under the same vaccine batch number and the same immunization method, it indicates that the segment item is more consistent with the time borrowing relationship of recording misalignment, and the system will write the segment item as a misalignment transfer item. The data collection time after the execution time of the other column is the data collection time in the response sequence of this column that is later than the execution time of the other column but earlier in time. If the segment item neither forms a response transfer item nor a misalignment transfer item, the system will write the segment item as a missed immunization retention item. Response transfer items, misalignment transfer items, and missed immunization retention items together generate inner layer status items.

[0093] In S5-3, the system performs double-layer horizon contraction on the outer boundary item and the inner state item. The boundary break item in the outer boundary item indicates that the missed exemption state pipe segment has lost the support of the front and rear reach boundaries, and the system writes the boundary break item as a segment deletion item. The reach transition item in the inner state item indicates that the original missed exemption state pipe segment has been converted to reach response at the acquisition time level, and the misalignment transition item indicates that the original missed exemption state pipe segment has been converted to the record verification direction, and the system writes the reach transition item and the misalignment transition item as resolution items. The missed exemption retention item in the inner state item indicates that the pipe segment has not yet formed a reach response or misalignment response, and the system writes the missed exemption retention item as a re-exemption item. Subsequently, the system removes the segment deletion item and resolution item from the cut pipe output result, retains the re-exemption item and the reach state pipe segment and misalignment state pipe segment that did not participate in the missed exemption contraction, and generates a contracted pipe. The removal in this step means that it does not enter the contracted pipe output and does not change the original record of the cut pipe.

[0094] In S5-4, the system merges the shrink tube and the trimming tube according to the column number, execution order, vaccine batch number, and immunization method to form a data set for task generation. During merging, the revaccination items in the shrink tube are used to generate revaccination verification tasks, the contact status segments retained in the trimming tube are used to generate observation tasks, and the mis-hanging status segments retained in the trimming tube are used to generate record verification tasks. The observation task is used to continue to check the post-immunization response recovery status of the column where the contact status segment is located. The revaccination verification task is used to perform pre-revaccination verification on the column where the revaccination item is located, and the verification content includes the column. The system records the column number, vaccine batch number, immunization method, and execution priority; the verification task is used to check the column number, execution time, vaccine batch number, and immunization method involved in the misaligned management segment; if multiple tasks are generated under the same column number, execution priority, vaccine batch number, and immunization method, the verification task is entered into the task table before the re-immunization verification task, and the re-immunization verification task is entered into the task table before the observation task; the task that enters the task table first is retained, and the remaining tasks are not entered into the task table; the system finally summarizes the task type, column number, execution priority, vaccine batch number, and immunization method to generate a task table;

[0095] Through S5 processing, the segments in the cut tube that were missed in the immunization process are first narrowed down by the front and back reach boundaries at the execution priority level, and then the reaction status is judged at the time of collection. Segments that have already turned into a reach reaction or misattached reaction are no longer included in the scope of revaccination verification, while segments that have not yet formed a reaction explanation are included in the revaccination item. Segments in the reach state and misattached state are included in the observation task and record verification task, respectively. This allows the task table to directly point to the column area, priority, vaccine batch number and immunization method, reducing the scope of revaccination verification being too broad and the omissions in record verification.

[0096] In practical applications: In a certain batch of drinking water immunization, if column 3 is still a missed immunization segment after S4, the system will find the reached immunization segment in column 2 and then the reached immunization segment in column 4 along the execution sequence in the trimming tube. All three segments have the same vaccine batch number and immunization method, so column 3 is written as a segment item. The system continues to read the reaction sequence of column 3. If a new drinking water difference appears in a subsequent collection time, and a new active difference appears in the next collection time, while no new abnormal reduction difference is added, then column 3 is changed from a segment item to a reach transition item and removed from the revaccination range as a resolution item. If column 3 never forms the reaction sequence, it is written as a revaccination item, and a revaccination verification task is generated in the task table.

[0097] Furthermore, a livestock and poultry farming management system includes:

[0098] The batch linking module is used to read the immunization record and column baseline record of the current immunization batch, and write the column number of the same vaccine batch number and the same immunization method in the immunization record into an immunization table according to the execution order and execution time;

[0099] The response sorting module is used to read the post-immunization population records of each column after immunization is completed in each column of the immunization table, perform the same field difference between the post-immunization population records and the column baseline records, sort them according to the collection time, and generate a response sequence containing drinking water difference, activity difference, aggregation difference and abnormal reduction difference.

[0100] The state estimation module performs joint posterior truncation based on the reaction sequence, and generates reach items, omission / exemption pending items, and misplaced items according to the sequential relationship between the drinking water difference, activity difference, aggregation difference, abnormal reduction difference and execution time of each column area. It then merges the reach items, misplaced items, and omission / exemption pending items sandwiched by reach items before and after the execution sequence to generate the state management.

[0101] The constraint pruning module performs Bayesian Zonotope pruning on the state tubes based on the immunization table. It takes the same vaccine batch number, the same immunization method, adjacent execution order, and execution time range as batch number band, method band, order band, and time band, respectively. It finds the intersection of the state tube with the batch number band, method band, order band, and time band item by item, and deletes the state tube segments with empty intersections to generate the pruning tube.

[0102] The task generation module performs double-layer horizon contraction based on the trimmed tube, determines the front and back boundaries of the missed immunization status tube segments according to the execution priority, and performs inner-layer state transformation on the segment items according to the collection time. It generates re-immunization verification tasks for the retained missed immunization status tube segments, generates observation tasks for the reached status tube segments, and generates record verification tasks for the misattached status tube segments. It also summarizes the task type, column number, execution priority, vaccine batch number, and immunization method to generate a task table.

[0103] Furthermore, a storage medium for livestock and poultry breeding management is provided, the storage medium storing a computer program, which, when executed by a processor, implements the various steps of the breeding management method as described above.

[0104] Working Principle: This solution uses the post-immunization population response to verify whether the immunization truly reached the target area, rather than simply checking whether the immunization task was registered as completed. First, it reads the immunization records of the current immunization batch and the baseline records of the target area to create an immunization table containing the vaccine batch number, immunization method, target area number, execution order, and execution time. Then, it performs field-specific differences between the post-immunization population records and the baseline records of the target area to obtain differences in drinking water, activity, aggregation, and abnormal population reduction, forming a response sequence based on the collection time. Subsequently, it uses big data processing to analyze the chronological relationship of responses in each target area, determining whether the target area has achieved a post-immunization response, is suspected of having missed immunization due to a lack of response, or is a misregistration due to an abnormal execution time or target area record. Finally, it uses the same vaccine batch number, same immunization method, adjacent execution order, and execution time range to prune the status results, and finally summarizes the target areas requiring further observation, revaccination verification, and record verification to generate a task table.

[0105] In practical applications, for example, if a duck shed uses drinking water immunization to administer the same batch of vaccine, and pens 2 and 4 show changes in drinking water first after immunization, followed by changes in activity or aggregation, without any simultaneous abnormal herd reduction, the system determines that both have developed post-immunization reactions. If pen 3, located in the middle, shows no changes in drinking water or activity during the same period, the system will not simply assume the entire waterline is abnormal. Instead, considering the preceding and following relationship between pens 2 and 4, pen 3 will be listed as a re-immunization checkpoint. If the reaction time of pen 5 is earlier than the execution time registered for this pen and close to the execution time of another pen, the system will generate a record verification task to check for errors in pen numbering or execution time. This allows for the separate handling of immunization coverage, suspected missed immunizations, and mis-registered records before an increase in mortality or significant fluctuations in production indicators.

[0106] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for managing poultry farming in animal husbandry, characterized in that, include: S1. Read the immunization record and column baseline record of the current immunization batch, and write the column numbers of the same vaccine batch number and the same immunization method in the immunization record into an immunization table according to the execution order and execution time; S2. After completing immunization in each column of the immunization table, read the post-immunization population record in each column, perform the same field difference between the post-immunization population record and the column baseline record, and sort them according to the collection time to generate a response sequence containing drinking water difference, activity difference, aggregation difference and abnormal population reduction difference. S3. Perform joint posterior truncation based on the reaction sequence, generate reach items, omission / exemption pending items, and misplaced items according to the sequential relationship between the drinking water difference, activity difference, aggregation difference, abnormal reduction difference and execution time of each column area, and merge the reach items, misplaced items, and omission / exemption pending items sandwiched by the reach items before and after the execution sequence to generate a state tube. S4. Based on the immunization table, perform Bayesian Zonotope pruning on the state tube. Take the same vaccine batch number, the same immunization method, adjacent execution order, and execution time range as batch number band, method band, order band, and time band, respectively. Find the intersection of the state tube with the batch number band, method band, order band, and time band item by item, and delete the state tube segments with empty intersections to generate the pruning tube. S5. Based on the cut tube, perform double-layer horizon contraction, determine the front and back boundaries of the missed immunization status tube segments according to the execution priority, and perform inner layer transformation on the segment items according to the collection time. Generate re-immunization verification tasks for the retained missed immunization status tube segments, generate observation tasks for the reached status tube segments, and generate record verification tasks for the misattached status tube segments. Summarize the task type, column number, execution priority, vaccine batch number and immunization method to generate a task table.

2. The method for managing poultry farming in animal husbandry according to claim 1, characterized in that: S1 includes: S1-1. Read the immunization record of the current immunization batch, merge multiple records with the same vaccine batch number, immunization method and column number into one column immunization item, and take the first execution time as the execution time of the column immunization item to generate an immunization item table. S1-2. Read the immunization item table, sort the immunization items in the column area under the same vaccine batch number and the same immunization method according to the execution time, and sort them in ascending order by column area number when the execution time is the same, and generate a sequence table. S1-3. Read the priority table and column reference record, associate the column number in the priority table with the column number in the column reference record, and write the vaccine batch number, immunization method, column number, execution priority, execution time, and column reference record into an immunization table.

3. A method for managing poultry farming in animal husbandry according to claim 2, characterized in that: S2 includes: S2-1. Read the immunization table and post-immunization population records. Link the post-immunization population records with the same column number and whose collection time is after the execution time to the execution order and execution time in the immunization table. Subtract the execution time from the collection time and output the post-record table. S2-2. Read the post-record table and the column baseline record. Subtract the drinking water value, activity value, aggregation value, and abnormal reduction value in the column baseline record from the drinking water value, activity value, aggregation value, and abnormal reduction value in the post-record table respectively to generate drinking water difference, activity difference, aggregation difference, and abnormal reduction difference. Write the difference value as rising position, falling position, and level position respectively when the difference value is greater than zero, less than zero, and zero, and output the difference position table. S2-3. Read the differential position table, write the collection time of the first non-equal position under the same column area number as the reaction start time, write the drinking water difference, activity difference, aggregation difference and abnormal reduction difference under the same collection time as the reaction items in the field order of drinking water, activity, aggregation and abnormal reduction, and arrange the reaction items in the order of time interval to generate the reaction sequence.

4. A method for managing poultry farming in animal husbandry according to claim 3, characterized in that: S3 includes: S3-1. Read the reaction sequence and immune table. Find the first non-zero collection time for the drinking water difference, activity difference, aggregation difference and abnormal decrease difference according to the column number. Write the earliest time among the first non-zero collection times as the reaction start time. Combine the first time of drinking water, the first time of activity, the first time of aggregation, the first time of decrease and the reaction start time according to the same column number to generate the first time item. S3-2. Read the first time item and perform joint posterior truncation. For columns where the first drinking time exists, the first drinking time is earlier than the first activity time or the first aggregation time, the first reduction time is empty, or the first reduction time is different from the first drinking time, the first activity time, and the first aggregation time, write the column number, execution order, and execution time as the reach item; for columns where the first drinking time and the first activity time are empty, write the column number, execution order, and execution time as the missed immunization pending item; for columns where the reaction start time is earlier than the execution time of this column, and the reaction start time is equal to the execution time of another column in the same immunization table or the collection time after the execution time of another column, write the column number, execution order, and execution time as the misplaced item, and generate the posterior table; S3-3. Read the posterior table and the immunization table and continue to perform joint posterior truncation. Define the reached item as the reached state, the misplaced item as the misplaced state, and define the column area with reached items in both the preceding and following columns and with undetermined missed immunization items in this column area as the missed immunization state. Then merge the reached state, missed immunization state and misplaced state to generate a state management system.

5. A method for managing poultry farming in animal husbandry according to claim 4, characterized in that: S4 includes: S4-1. Taking a single segment in the status tube as the trimming object, search the immunization table for the column number that has the same vaccine batch number and the same immunization method as the trimming object, and take the current column, the previous column, and the next column according to the execution order; if all three exist, write the vaccine batch number, immunization method, execution order, and execution time of the three as batch number band, method band, order band, and time band; otherwise, write the trimming object as an isolated segment item and generate a three-column band item. S4-2. Project the vaccine batch number, immunization method, execution order, and execution time in the cropping object to the batch number band, method band, order band, and time band in the three-column band, respectively; if all four items match, write the cropping object as a same-chain item; otherwise, write the cropping object as a broken-chain item and generate a signature item.

6. A method for managing poultry farming in animal husbandry according to claim 5, characterized in that: S4 further includes: S4-3. Perform clamping post-clipping based on signature items and status tubes. When a co-chain item is established and the status type is "reached state", write the tube segment as a reach lock item. When a co-chain item is established, the status type is "missed state", and the execution order is between the previous and subsequent reach states, write the tube segment as a clamping lock item. When a co-chain item is established, the status type is "misaligned state", and the reaction start time is equal to the collection time after the execution time of another column area, write the tube segment as a borrowed time lock item. Otherwise, write the tube segment as a pruning item and generate a lock item. S4-4. Execute Hamilton-Jacobi reverse deletion of lock items from the back position to the front position according to the execution order. If each leakage prevention and exemption status segment in the backtracking path has a clamping lock item, write the backtracking path as a pass item; otherwise, write the first leakage prevention and exemption status segment that lacks a clamping lock item and the leakage prevention and exemption status segments before and after it as deletion items. S4-5. Perform closed-loop pruning using lock items, deletion items, and three-column items. Write the segments with access lock items, clamp lock items, or time-borrowing lock items that have not been written with deletion items and isolated segments into the pruning tube. Delete the segments with deletion items or isolated segments from the status tube. Write the segments with items to be pruned that have not formed access lock items, clamp lock items, or time-borrowing lock items as items to be verified, and generate the pruning tube.

7. A method for managing poultry farming in animal husbandry according to claim 6, characterized in that: S5 includes: S5-1. Taking the unimmunized tube segment in the cut tube as the shrinkage object, find the first touched tube segment in the execution order and write it as the front boundary item, and find the first touched tube segment in the execution order and write it as the back boundary item. If both the front boundary item and the back boundary item exist and their vaccine batch number and immunization method are the same as the shrinkage object, write the shrinkage object as the segment item; otherwise, write it as the break boundary item and generate the outer boundary item. S5-2. Link the segment item to the reaction sequence according to the column area number, and take two adjacent reaction items according to the collection time to perform the inner layer horizon transition. If the latter reaction item adds a drinking water difference relative to the former reaction item, and the subsequent adjacent reaction item adds an activity difference or aggregation difference, and no abnormal reduction difference is added at the same collection time, write the segment item as the reach transition item. If the reaction start time is earlier than the execution time of this column area and equal to the collection time after the execution time of another column area, write the segment item as the mis-linked transition item. Otherwise, write the segment item as the omission-free retention item and generate the inner layer state item.

8. A method for managing poultry farming in animal husbandry according to claim 7, characterized in that: The S5 also includes: S5-3. Perform double-layer horizon contraction on the outer boundary terms and inner state terms, write the boundary terms as segment deletion terms, write the reach transfer terms and mis-hanging transfer terms as resolution terms, write the omission exemption and retention terms as re-exemption terms, and remove the segment deletion terms and resolution terms from the trimming tube to generate the contraction tube. S5-4. Combine the shrink tube and the cut tube according to the column area number, execution order, vaccine batch number and immunization method. Write the contact status tube segment as the observation task, write the re-immunization item as the re-immunization verification task, and write the misattached status tube segment as the record verification task. Summarize the task type, column area number, execution order, vaccine batch number and immunization method to generate a task table.

9. A management system for livestock and poultry farming, characterized in that, include: The batch linking module is used to read the immunization record and column baseline record of the current immunization batch, and write the column number of the same vaccine batch number and the same immunization method in the immunization record into an immunization table according to the execution order and execution time; The response sorting module is used to read the post-immunization population records of each column after immunization is completed in each column of the immunization table, perform the same field difference between the post-immunization population records and the column baseline records, sort them according to the collection time, and generate a response sequence containing drinking water difference, activity difference, aggregation difference and abnormal reduction difference. The state estimation module performs joint posterior truncation based on the reaction sequence, and generates reach items, omission / exemption pending items, and misplaced items according to the sequential relationship between the drinking water difference, activity difference, aggregation difference, abnormal reduction difference and execution time of each column area. It then merges the reach items, misplaced items, and omission / exemption pending items sandwiched by reach items before and after the execution sequence to generate the state management. The constraint pruning module performs Bayesian Zonotope pruning on the state tubes based on the immunization table. It takes the same vaccine batch number, the same immunization method, adjacent execution order, and execution time range as batch number band, method band, order band, and time band, respectively. It finds the intersection of the state tube with the batch number band, method band, order band, and time band item by item, and deletes the state tube segments with empty intersections to generate the pruning tube. The task generation module performs double-layer horizon contraction based on the trimmed tube, determines the front and back boundaries of the missed immunization status tube segments according to the execution priority, and performs inner-layer state transformation on the segment items according to the collection time. It generates re-immunization verification tasks for the retained missed immunization status tube segments, generates observation tasks for the reached status tube segments, and generates record verification tasks for the misattached status tube segments. It also summarizes the task type, column number, execution priority, vaccine batch number, and immunization method to generate a task table.

10. A storage medium for livestock and poultry farming management, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the various steps of the aquaculture management method as described in any one of claims 1-8.

Images