A low-production bird identification method, device, apparatus and product

Abstract

The embodiment of the application discloses a low-yield bird identification method, device, equipment and product, and relates to the technical field of Internet of Things. The low-yield bird identification method comprises the following steps: acquiring a reading event sequence of a target RFID tag read by an RFID card reader in a current analysis period in an egg-laying area, the target RFID tag being an RFID tag arranged on a bird to be identified; dividing the reading event sequence into a plurality of behavior sessions based on reading time intervals of adjacent reading events in the reading event sequence, generating session features of each behavior session based on reading events in each behavior session in the plurality of behavior sessions; generating a multi-dimensional feature vector of the bird to be identified based on the session features of each behavior session, inputting the multi-dimensional feature vector into a low-yield probability prediction model, predicting a probability that the bird to be identified is a low-yield bird, and determining whether the bird to be identified is a low-yield bird based on the probability.

Description

Technical Field

[0001] This application relates to the field of Internet of Things (IoT) technology, and in particular to a method, apparatus, equipment, and product for identifying low-yielding poultry. Background Technology

[0002] In the large-scale farming of poultry such as chickens, ducks, and geese, it is necessary to cull low-producing birds to reduce farming costs. Currently, low-producing birds are mainly identified by farmers based on their experience. This method relies on the farmers' subjective judgment, resulting in low accuracy in identifying low-producing birds and easily leading to the wrongful culling. Summary of the Invention

[0003] This application provides a method, apparatus, equipment, and product for identifying low-producing poultry, which can improve the accuracy of identifying low-producing poultry and avoid culling poultry.

[0004] In a first aspect, embodiments of this application provide a method for identifying low-producing poultry, the method comprising: Obtain the sequence of reading events of the target RFID tag read by the RFID reader in the egg-laying area during the current analysis period; wherein, the RFID reader in the egg-laying area reads the RFID tags present in the egg-laying area at a set time interval, the target RFID tag is the RFID tag set on the poultry to be identified, the reading events in the sequence of reading events are sorted according to the reading time of the reading event, and the reading event includes: the reading time and the RSSI value of the target RFID tag read; Based on the reading time interval between adjacent reading events in the reading event sequence, the reading event sequence is divided into multiple behavior sessions, and session features of each behavior session are generated based on the reading events within each behavior session. Based on the session features of each behavioral session, a multi-dimensional feature vector of the poultry to be identified is generated. The multi-dimensional feature vector is then input into a low-production probability prediction model to predict the probability that the poultry to be identified is a low-production poultry. The low-production probability prediction model is trained based on the historical multi-dimensional feature vector of the sample poultry and the corresponding historical egg production performance labels. The historical egg production performance labels include: low-production labels and non-low-production labels. Based on the probability that the poultry to be identified is a low-producing poultry, determine whether the poultry to be identified is a low-producing poultry.

[0005] In one embodiment, dividing the read event sequence into multiple behavior sessions based on the read time interval between adjacent read events in the read event sequence includes: Adjacent read events with a read time interval less than the interval threshold in the read event sequence are assigned to the same behavior session, while adjacent read events with a read time interval not less than the interval threshold are assigned to different behavior sessions, thus obtaining the multiple behavior sessions.

[0006] In one embodiment, the session features include: session duration and in-session activity. The session features of each behavioral session are generated based on read events within each of the multiple behavioral sessions, including: For each behavior session, perform the following processing: Calculate the reading time difference between the start reading event and the end reading event within the behavior session, and determine the reading time difference as the session duration of the behavior session; Calculate the variance of the RSSI values ​​of each read event within the behavior session, and determine the variance as the session activity of the behavior session.

[0007] In one embodiment, generating a multi-dimensional feature vector based on the session features of each behavioral session includes: Based on the session characteristics of each behavioral session, features of multiple dimensions are generated. These features include individual features and group relative features. The individual features include: the number of egg-laying behavioral sessions, the month-on-month change rate of the number of egg-laying behavioral sessions, the coefficient of variation of the session duration of egg-laying behavioral sessions, and the Shannon entropy of the probability distribution of the session start time. The egg-laying behavioral session is a behavioral session whose session duration is within the normal egg-laying session duration and whose activity level within the session is less than the activity level threshold. The group relative features include: the deviation of the individual features from the corresponding group baseline features. The features of the multiple dimensions are concatenated to generate the multi-dimensional feature vector.

[0008] In one embodiment, the group baseline features are determined based on the ranking of the corresponding individual features of each bird in the flock to which the bird to be identified belongs during the current analysis period.

[0009] In one embodiment, determining whether a bird to be identified is a low-producing bird based on the probability that the bird is a low-producing bird includes: If the probability is greater than the probability threshold, then the poultry to be identified is determined to be abnormal in the current analysis period; If the poultry to be identified shows abnormalities for N consecutive analysis cycles, then the poultry to be identified is determined to be a low-producing poultry; wherein, N is greater than or equal to 3.

[0010] In one embodiment, the method further includes: If the poultry to be identified is a low-producing poultry, obtain the key features of the poultry to be identified for N analysis cycles. The key features include: the average session duration of the egg-laying behavior session in each analysis cycle, the number of egg-laying behavior sessions, and the deviation of the number of egg-laying behavior sessions from the baseline number of sessions in the flock. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is greater than the normal session duration threshold, and the number of egg-laying behavior sessions in the N analysis cycles is less than the first quantity threshold, then the low-production type of the poultry to be identified is determined to be pathological brooding. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is within the normal session duration range, and the number of egg-laying behavior sessions in the N analysis cycles is less than the second quantity threshold, and the deviation of the number of egg-laying behavior sessions in the N analysis cycles from the baseline number of sessions in the flock is less than the deviation threshold, then the low-production type of the poultry to be identified is determined to be physiological low production.

[0011] Secondly, embodiments of this application provide an egg production monitoring device, the device comprising: The acquisition module is used to acquire the sequence of reading events of the target RFID tag read by the RFID reader in the egg-laying area during the current analysis period; wherein, the RFID reader in the egg-laying area reads the RFID tags present in the egg-laying area at a set time interval, the target RFID tag is the RFID tag set on the poultry to be identified, the reading events in the sequence of reading events are sorted according to the reading time of the reading event, and the reading event includes: the reading time and the RSSI value of the target RFID tag read; The generation module is used to divide the reading event sequence into multiple behavior sessions based on the reading time interval of adjacent reading events in the reading event sequence, generate session features of each behavior session based on the reading events in each behavior session, and generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavior session. The prediction module is used to generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavioral session, and input the multi-dimensional feature vector into the low-production probability prediction model to predict the probability that the poultry to be identified is a low-production poultry; wherein, the low-production probability prediction model is trained based on the historical multi-dimensional feature vector of the sample poultry and the corresponding historical egg production performance labels, the historical egg production performance labels including: low-production label and non-low-production label. The determination module is used to determine whether the poultry to be identified is a low-producing poultry based on the probability that the poultry to be identified is a low-producing poultry.

[0012] Thirdly, embodiments of this application provide an electronic device, including: A memory on which computer programs are stored; A processor is used to execute the computer program to implement the above-described method for identifying low-producing poultry.

[0013] Fourthly, embodiments of this application provide a computer program product, including a computer program that can be executed by a processor to implement the above-described low-producing poultry identification method.

[0014] Fifthly, embodiments of this application provide a readable storage medium having a computer program stored thereon, the computer program being executable by a processor to implement the above-described low-producing poultry identification method.

[0015] In the technical solution provided in this application, the sequence of reading events from the RFID reader in the egg-laying area, which reads the target RFID tags attached to the poultry to be analyzed, can reflect the egg-laying behavior pattern of the poultry to be analyzed. By having the RFID reader in the egg-laying area read the RFID tags present in the egg-laying area at set time intervals, a multi-dimensional feature vector is generated based on the session characteristics of each behavioral session in the sequence of reading events from the RFID reader in the egg-laying area, which reads the target RFID tags attached to the poultry to be analyzed. The multi-dimensional feature vector is then input into a low-laying probability prediction model to predict the probability that the poultry to be analyzed is a low-laying poultry. This approach combines IoT sensing and machine learning, and based on the behavioral pattern characteristics of the poultry to be analyzed, accurately predicts the probability that the poultry to be analyzed is a low-laying poultry, thereby improving the accuracy and efficiency of identifying low-laying poultry. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the implementation environment to which the low-producing poultry identification method provided in this application embodiment is applicable; Figure 2 This is a flowchart illustrating a method for identifying low-producing poultry provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of a low-yield poultry identification device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] This application provides a method, apparatus, equipment, and product for identifying low-producing poultry.

[0019] To enable those skilled in the art to better understand the technical solutions in this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this application.

[0020] The technical terms used in the embodiments of this application are explained as follows: RFID, Radio Frequency Identification.

[0021] RSSI, Received Signal Strength Indication.

[0022] GBDT, Gradient Boosting Decision Tree. Figure 1 This is a schematic diagram illustrating the implementation environment to which the low-producing poultry identification method provided in this application is applicable. For example... Figure 1 As shown, the implementation environment includes: RFID tags, RFID readers in the egg-laying area, and a cloud service platform.

[0023] RFID tags are placed on poultry in farms, such as chickens, ducks, and geese. Each RFID tag is placed on one poultry and is used to uniquely identify one poultry.

[0024] RFID readers are installed in the egg-laying area of ​​the farm to read RFID tags present in the area at set time intervals. When an RFID tag is read, a read event is generated and sent to the cloud service platform. The reading range of the RFID reader is configured based on the area of ​​the egg-laying area and is approximately equal to that area. If the RFID reader reads an RFID tag, it means that the RFID tag exists in the egg-laying area, and the poultry identified by that RFID tag is in the egg-laying area. If the RFID reader does not read an RFID tag, it means that the RFID tag does not exist in the egg-laying area, and the poultry identified by that RFID tag is not in the egg-laying area. A read event is a record of one read of an RFID tag, which can be structured triplet data, including: the time the RFID tag was read, the tag ID of the read RFID tag, and the RSSI value of the read RFID tag.

[0025] The cloud service platform is used to receive and store read events and read event data sent by RFID readers.

[0026] When identifying whether a poultry is a low-producing poultry, the cloud server can obtain the reading event sequence of the RFID tag set on the poultry by the RFID reader in the current analysis period. Based on the reading time interval of adjacent reading events in the reading event sequence, the reading event sequence is divided into multiple behavioral sessions. Based on the reading events in each behavioral session, session features of each behavioral session are generated. Based on the session features of each behavioral session, a multi-dimensional feature vector is generated. The multi-dimensional feature vector is input into the low-producing probability prediction model to predict the probability that the poultry is a low-producing poultry. Based on the probability, it is determined whether the poultry is a low-producing poultry.

[0027] In the above technical solution, the sequence of reading events from RFID tags on the poultry to be analyzed, detected by the RFID reader in the egg-laying area, reflects the egg-laying behavior pattern of the poultry. By placing RFID tags on the poultry and RFID readers in the egg-laying area, the RFID readers in the egg-laying area read the RFID tags present in the area at set time intervals. Based on the session characteristics of each behavioral session in the sequence of reading events from the RFID readers on the poultry to be analyzed, a multi-dimensional feature vector is generated. This multi-dimensional feature vector is then input into a low-laying probability prediction model to predict the probability that the poultry to be analyzed is a low-laying poultry. This approach combines IoT sensing and machine learning, based on the behavioral pattern characteristics of the poultry to be analyzed, to accurately predict the probability that the poultry to be analyzed is a low-laying poultry, thus improving the accuracy and efficiency of identifying low-laying poultry.

[0028] Figure 2 This is a flowchart illustrating a method for identifying low-producing poultry provided in an embodiment of this application. Figure 2 The method shown can be used on a cloud service platform and executed by the cloud service platform. For example... Figure 2 As shown, the method includes the following steps: S202, obtain the reading event sequence of the target RFID tag read by the RFID reader in the egg-laying area during the current analysis period. The target RFID tag is the RFID tag set on the poultry to be identified.

[0029] The current analysis period refers to the analysis period currently being processed, which can be any analysis period. An analysis period can include multiple consecutive egg-laying cycles. An egg-laying cycle is the average period from the last egg-laying period to the next egg-laying period for poultry, which is, for example, 1 day. An analysis period is, for example, one week.

[0030] The target RFID tag is an RFID tag attached to the poultry to be identified, which can uniquely identify the poultry. The RFID reader in the egg-laying area is an RFID reader set up in the egg-laying area, which reads the RFID tags present in the egg-laying area at set time intervals. The reading range of the RFID reader in the egg-laying area is configured based on the range of the egg-laying area. If the RFID reader in the egg-laying area reads the target RFID tag, it means that the target RFID tag exists in the egg-laying area, and the poultry identified by the target RFID tag is the poultry to be identified in the egg-laying area. If the RFID reader in the egg-laying area does not read the RFID tag, it means that the target RFID tag does not exist in the egg-laying area, and the poultry identified by the target RFID tag is the poultry to be analyzed is not in the egg-laying area.

[0031] The sequence of read events of the target RFID tag read by the RFID reader in the laying area during the current analysis period is a time sequence of these events. Each read event in the sequence is ordered by its read time. Each read event is an event record of a single instance of the RFID reader reading a target RFID tag in the laying area. It is structured data, including: the read time and the RSSI value of the target RFID tag. The read time is the time when the RFID reader in the laying area reads the target RFID tag, reflecting whether the poultry identified by the target RFID tag was in the laying area at that time. The RSSI value of the target RFID tag is the RSSI value of the target RFID tag read by the RFID reader in the laying area, reflecting the specific location and status of the poultry identified by the target RFID tag.

[0032] For example, the sequence of reading events when the RFID reader in the egg-laying area reads the target RFID tag during the current analysis cycle can be: Read event 1, read event 2, read event 3, ... read event N.

[0033] Wherein, read event n is the read event record of the nth time the RFID reader in the egg-laying area reads the target RFID tag in the current analysis period. Read event n includes: the time when the RFID reader in the egg-laying area reads the target RFID tag for the nth time, and the RSSI value of the tag read by the RFID reader in the egg-laying area when it reads the target RFID tag for the nth time, n=1, 2, 3, ... N.

[0034] S204, based on the reading time interval between adjacent reading events in the reading event sequence, divide the reading event sequence into multiple behavior sessions, and generate session features for each behavior session based on the reading events within each behavior session.

[0035] Each behavior session represents a continuous behavior of the poultry being analyzed. In one embodiment, adjacent read events with a read interval less than an interval threshold can be assigned to the same behavior session, while adjacent read events with a read interval not less than the interval threshold can be assigned to different behavior sessions, resulting in multiple behavior sessions.

[0036] Session features of a behavioral session are used to quantify the behavioral patterns of that session. In one embodiment, session features include session duration and in-session activity. The session duration of each behavioral session quantifies the duration the analyzed poultry spends in the laying area during that session, and the in-session activity of each behavioral session quantifies the behavioral state of the analyzed poultry in the laying area, such as resting, brooding, or frequent movement. For each behavioral session, the reading time difference between the start and end reading events can be calculated, and this reading time difference is determined as the session duration. The variance of the RSSI values ​​of each reading event within the behavioral session can be calculated, and this variance is determined as the in-session activity of the behavioral session. The start reading event is the first reading event in the behavioral session, and the reading time of this event is the session start time. The end reading event is the last reading event in the behavioral session, and the reading time of this event is the session end time. The formula is illustrated below:

[0037]

[0038] in, For session duration, This is the session end time. A represents the session start time, and A represents the session activity level. This represents the total number of events read within the behavior session. An index for reading events within a behavior session. For the first action in the session The RSSI value of each read event. This is the average RSSI value of all read events within the behavior session.

[0039] S206, Generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavioral session, and input the multi-dimensional feature vector into the low-yield probability prediction model to predict the probability that the poultry to be identified is a low-yield poultry.

[0040] Among them, the low-production probability prediction model is used to predict and output the probability that the poultry to be analyzed is a low-production poultry based on the multi-dimensional feature vector of the poultry to be identified.

[0041] The low-layout probability prediction model can be trained based on the historical multi-dimensional feature vectors of sample poultry and their corresponding egg production performance tags. In one embodiment, the GBDT model can be trained using the historical multi-dimensional feature vectors of sample poultry as input vectors and the corresponding egg production performance tags as output tags to obtain the low-layout probability prediction model. The historical multi-dimensional feature vectors of sample poultry are generated based on the sequence of RFID tag reading events from RFID readers in the egg production area during historical analysis periods. The specific generation method is similar to the method for generating the multi-dimensional feature vectors of poultry to be identified in the current analysis period. Egg production performance tags are used to identify the actual production performance of sample poultry during the historical analysis period. Egg production performance tags include: low-layout tags and non-low-layout tags.

[0042] Multidimensional feature vectors are used to comprehensively characterize factors related to egg production performance. In one embodiment, session features from various behavioral sessions can be concatenated to generate a multidimensional feature vector.

[0043] In one embodiment, features of multiple dimensions can be generated based on the session features of each behavioral session, and the features of multiple dimensions can be concatenated to generate a multi-dimensional feature vector.

[0044] Among them, the features of multiple dimensions are those related to egg production performance, including individual features and population relative features.

[0045] Individual characteristics include: the number of egg-laying behavior sessions, the rate of change of the number of egg-laying behavior sessions, the coefficient of variation of session duration for egg-laying behavior sessions, and the Shannon entropy of the probability distribution of session start time. An egg-laying behavior session is defined as a session whose duration falls within the normal range for egg-laying sessions and whose activity level is below the activity threshold. The number of egg-laying behavior sessions is the number of egg-laying behavior sessions in the current analysis period, which quantifies egg-laying activity. Population relative characteristics include: the deviation of individual characteristics from the corresponding population baseline characteristics, i.e., the deviation of the number of egg-laying behavior sessions from the population baseline number of egg-laying behavior sessions, the deviation of the month-on-month change rate of the number of egg-laying behavior sessions from the population baseline rate of change, the deviation of the coefficient of variation of session duration for egg-laying behavior sessions from the population baseline coefficient of variation of session duration for egg-laying behavior sessions, and the deviation of the Shannon entropy of the probability distribution of session start time from the population baseline Shannon entropy of session start time.

[0046] In practice, based on the session duration and activity level of each behavioral session in the current analysis period, it can be determined whether each behavioral session is an egg-laying behavioral session, and based on whether each behavioral session is an egg-laying behavioral session, the number of egg-laying behavioral sessions in the current analysis period can be determined.

[0047] The month-on-month change rate of the number of egg-laying behavior sessions is the rate of change of the number of egg-laying behavior sessions in the current analysis period compared to the number of egg-laying behavior sessions in the previous analysis period, which can quantify the egg-laying trend. In practice, the rate of change of the number of egg-laying behavior sessions in the current analysis period compared to the previous analysis period can be calculated based on the number of egg-laying behavior sessions in the current analysis period and the number of egg-laying behavior sessions in the previous analysis period, using the following formula:

[0048] in, This represents the rate of change in the number of egg-laying behavior sessions in the current analysis period compared to the number of egg-laying behavior sessions in the previous analysis period. This represents the number of egg-laying behavior sessions in the current analysis period. This represents the number of egg-laying behavior sessions in the previous analysis period.

[0049] The coefficient of variation (COP) of the session duration of laying behavior sessions is the coefficient of variation of the session duration of laying behavior sessions in the current analysis period, which can quantify the degree of fluctuation in laying time. In practice, the mean and standard deviation of the session duration of laying behavior sessions in the current analysis period can be calculated, and the COP can be calculated based on the following formula:

[0050] in, The coefficient of variation for the session duration of the egg-laying behavior session in the current analysis period. This represents the standard deviation of the session duration for all egg-laying behavior sessions in the current analysis period. This represents the average session duration of all egg-laying behavior sessions in the current analysis period.

[0051] The Shannon entropy of the probability distribution of session start time is the Shannon entropy of the probability distribution of session start time for each behavior within the current analysis period. It is a behavioral regularity entropy used to quantify the regularity of the expected entry times of poultry into the laying area during the current analysis period. In implementation, the current analysis period can be divided into K equal-length sub-intervals. For each sub-interval, the number of behavioral sessions whose session start time falls within that sub-interval is counted. Based on this number and the total number of behavioral sessions in the current analysis period, the session start probability for that sub-interval is calculated. Based on the session start probability of each sub-interval, the Shannon entropy of the probability distribution of session start time is calculated using the following formula:

[0052]

[0053] in, The Shannon entropy value is the probability distribution of the session start time for the current analysis period. To calculate the total number of sub-intervals in the current analysis period, This is the index of the sub-interval in the current analysis period. For the first The probability of starting a session in each sub-interval For the current analysis period, the session start time falls on the [number]th [period]. The number of behavioral sessions in each interval This represents the total number of sessions in the current analysis cycle.

[0054] The deviation of an individual characteristic from the corresponding group baseline characteristic is used to quantify the degree of deviation of the individual characteristics of the analyzed poultry from the group baseline characteristics of the flock to which the analyzed poultry belong. In practice, the deviation of an individual characteristic from the corresponding group baseline characteristic can be calculated based on the following formula:

[0055] in, The current analysis is expected to identify the first batch of poultry. Each individual characteristic is compared to the first characteristic of the flock to which the identified poultry belongs. Deviation of the baseline characteristics of a group The larger the absolute value of a negative value, the more significant the lag between an individual's characteristics and the baseline characteristics of the corresponding group. This value can quantify the tail of individuals. The current analysis is expected to identify the first batch of poultry. Individual characteristics. The current analysis aims to identify the flock to which the poultry belong. Individual group baseline characteristics. The current analysis week aims to identify the flock to which each bird belongs. The standard deviation of individual characteristics.

[0056] This can be used to identify the number of egg-laying behavior sessions, the rate of change in the number of egg-laying behavior sessions, the coefficient of variation of session duration, or the probability distribution of session start time for poultry during the current analysis week. Accordingly, It can be used as a baseline value for the number of egg-laying behavior sessions of the flock in the current analysis period, the baseline value for the rate of change of the number of egg-laying behavior sessions, the baseline value for the coefficient of variation of the session duration of egg-laying behavior sessions, or the baseline value for the Shannon entropy of the probability distribution of the session start time.

[0057] In this embodiment, features of the above-mentioned multiple dimensions are generated based on the session features of each behavioral session. The features of the above-mentioned multiple dimensions are concatenated to generate a multi-dimensional feature vector, which can accurately characterize the behavior of the poultry to be identified in the current analysis week from multiple dimensions such as output, trend, pattern, stability, and relative status. This facilitates the subsequent accurate prediction of the probability that the poultry to be identified is a low-producing poultry based on information from these dimensions, thereby improving the prediction accuracy.

[0058] In one embodiment, the group baseline characteristics can be determined based on the ranking of the corresponding individual characteristics of each bird in the flock to which the birds to be identified belong in the current analysis week, which can be the individual characteristics that specify the position in the ranking.

[0059] For example, regarding the number of egg-laying behavior sessions, the number of egg-laying behavior sessions for each bird in the flock to which the birds to be identified belong can be obtained in the current analysis period. The number of egg-laying behavior sessions for each bird is sorted in ascending order, and the number of egg-laying behavior sessions with a specified rank is determined as the baseline value for the number of egg-laying behavior sessions in the current analysis period. The specified rank can be determined based on quantiles, such as the rank at the 20th percentile of the overall sorted sequence.

[0060] In this embodiment, by ranking the individual characteristics of each bird in the flock to which the birds to be identified belong based on the current analysis week, the corresponding group baseline characteristics are determined. This provides a dynamic baseline, adapts to the dynamic changes of the flock, and avoids misidentification due to changes in flock information.

[0061] S208, Based on the probability that the poultry to be identified is a low-producing poultry, determine whether the poultry to be identified is a low-producing poultry.

[0062] In this embodiment, whether the poultry to be identified is a low-producing poultry is used to provide data support for whether to cull the poultry to be identified. It can output information on whether the poultry to be identified is a low-producing poultry to provide data support for whether to cull the poultry to be identified.

[0063] In one embodiment, the probability of identifying a poultry as a low-producing poultry can be compared with a probability threshold. If the probability is greater than the probability threshold, the poultry to be identified is determined to be abnormal in the current analysis period. If the poultry to be identified is abnormal in N consecutive analysis periods, the poultry to be identified is determined to be a low-producing poultry.

[0064] Low-producing poultry are usually culled. To avoid misidentification of low-producing poultry and subsequent wrongful culling, the probability threshold can be configured to a high probability threshold, such as 0.95. N can be configured to a value greater than 3, that is, N is greater than or equal to 3. Thus, the poultry to be identified as low-producing poultry will only be identified as low-producing poultry when the probability of the poultry to be identified as low-producing poultry remains high for more than 3 analysis periods, thereby minimizing the possibility of wrongful culling.

[0065] Furthermore, if the poultry to be identified is a low-producing poultry, the low-producing type of the poultry to be identified can be determined and the low-producing type of the poultry to be identified can be output to provide data support for the culling method of the poultry to be identified. In one embodiment, after step S208, the following steps are also included: If the poultry to be identified are low-producing poultry, obtain the key features of the poultry to be identified for N analysis periods. The key features include: the average session duration of the egg-laying behavior session, the number of egg-laying behavior sessions, and the deviation of the number of egg-laying behavior sessions from the baseline number of sessions in the flock in each analysis period. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is greater than the normal session duration threshold, and the number of egg-laying behavior sessions in the N analysis cycles is less than the first quantity threshold, then the low-production type of the poultry to be identified is determined to be pathological brooding. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is within the normal session duration range, and the number of egg-laying behavior sessions in the N analysis cycles is less than the second quantity threshold, and the deviation of the number of egg-laying behavior sessions in the N analysis cycles from the baseline number of sessions in the flock is less than the deviation threshold, then the low-production type of the poultry to be identified is determined to be physiological low production.

[0066] The key features of the poultry to be identified in each analysis period can be obtained from the multi-dimensional features generated when calculating the multi-dimensional feature vector of the poultry to be identified in the corresponding analysis period.

[0067] In implementation, each analysis cycle can be continuously used as the current analysis cycle. Each bird in the flock is identified as a potential low-producing bird, and the analysis data for each analysis cycle is output. An individual digital profile can be created for each bird, recording its reading event sequence, behavioral session characteristics, multi-dimensional characteristics, probability of being a low-producing bird, low-producing type, and other information for each analysis cycle, providing a full lifecycle digital profile for each bird. Based on the individual digital profiles of each bird in the flock, various statistical indicators of the flock can be calculated and visualized. This provides comprehensive data support for poultry culling, avoiding false culling.

[0068] Based on the same idea as the above embodiments, this application also provides a device for identifying low-producing poultry. Figure 3 This is a schematic diagram of a low-yield poultry identification device provided in an embodiment of this application. Figure 3 As shown, the low-producing poultry identification device 300 includes: The acquisition module 310 is used to acquire a sequence of reading events of target RFID tags read by the RFID reader in the egg-laying area during the current analysis period; wherein, the RFID reader in the egg-laying area reads RFID tags existing in the egg-laying area at set time intervals, the target RFID tag is an RFID tag set on the poultry to be identified, and the reading events in the reading event sequence are sorted according to the reading time of the reading event, and the reading event includes: the reading time and the RSSI value of the target RFID tag read; The generation module 320 is used to divide the reading event sequence into multiple behavioral sessions based on the reading time interval of adjacent reading events in the reading event sequence, generate session features of each behavioral session based on the reading events in each behavioral session, and generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavioral session. The prediction module 330 is used to generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavioral session, and input the multi-dimensional feature vector into the low-production probability prediction model to predict the probability that the poultry to be identified is a low-production poultry; wherein, the low-production probability prediction model is trained based on the historical multi-dimensional feature vector of the sample poultry and the corresponding historical egg production performance labels, and the historical egg production performance labels include: low-production label and non-low-production label. The determining module 340 is used to determine whether the poultry to be identified is a low-producing poultry based on the probability that the poultry to be identified is a low-producing poultry.

[0069] In one embodiment, the generation module 320 is specifically used for: Adjacent read events with a read time interval less than the interval threshold in the read event sequence are assigned to the same behavior session, while adjacent read events with a read time interval not less than the interval threshold are assigned to different behavior sessions, thus obtaining the multiple behavior sessions.

[0070] In one embodiment, the generation module 320 is specifically used for: For each behavior session, perform the following processing: Calculate the reading time difference between the start reading event and the end reading event within the behavior session, and determine the reading time difference as the session duration of the behavior session; Calculate the variance of the RSSI values ​​of each read event within the behavior session, and determine the variance as the session activity of the behavior session.

[0071] In one embodiment, the prediction module 330 is specifically used for: Based on the session characteristics of each behavioral session, features of multiple dimensions are generated. These features include individual features and group relative features. The individual features include: the number of egg-laying behavioral sessions, the month-on-month change rate of the number of egg-laying behavioral sessions, the coefficient of variation of the session duration of egg-laying behavioral sessions, and the Shannon entropy of the probability distribution of the session start time. The egg-laying behavioral session is a behavioral session whose session duration is within the normal egg-laying session duration and whose activity level within the session is less than the activity level threshold. The group relative features include: the deviation of the individual features from the corresponding group baseline features. The features of the multiple dimensions are concatenated to generate the multi-dimensional feature vector.

[0072] In one embodiment, the group baseline features are determined based on the ranking of the corresponding individual features of each bird in the flock to which the bird to be identified belongs during the current analysis period.

[0073] In one embodiment, the determining module 340 is specifically used for: If the probability is greater than the probability threshold, then the poultry to be identified is determined to be abnormal in the current analysis period; If the poultry to be identified shows abnormalities for N consecutive analysis cycles, then the poultry to be identified is determined to be a low-producing poultry; wherein, N is greater than or equal to 3.

[0074] In one embodiment, the determining module 340 is further configured to: If the poultry to be identified is a low-producing poultry, obtain the key features of the poultry to be identified for N analysis cycles. The key features include: the average session duration of the egg-laying behavior session in each analysis cycle, the number of egg-laying behavior sessions, and the deviation of the number of egg-laying behavior sessions from the baseline number of sessions in the flock. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is greater than the normal session duration threshold, and the number of egg-laying behavior sessions in the N analysis cycles is less than the first quantity threshold, then the low-production type of the poultry to be identified is determined to be pathological brooding. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is within the normal session duration range, and the number of egg-laying behavior sessions in the N analysis cycles is less than the second quantity threshold, and the deviation of the number of egg-laying behavior sessions in the N analysis cycles from the baseline number of sessions in the flock is less than the deviation threshold, then the low-production type of the poultry to be identified is determined to be physiological low production.

[0075] The low-producing poultry identification device of this application embodiment can serve as the execution subject of the low-producing poultry identification method provided in the above embodiments, and realize all the functions of the low-producing poultry identification method provided in the above embodiments. Since the principle is the same, it will not be described again here.

[0076] Based on the same ideas as the above embodiments, this application also provides an electronic device. For example... Figure 4 As shown, the electronic device 400 can vary considerably due to differences in configuration or performance, and may include one or more memories 410 and processors 420. The memory 410 stores a computer program, and the processor 420 processes the computer program to implement the low-producing poultry identification method provided in the above embodiments. To avoid repetition, further details are omitted here.

[0077] Furthermore, one or more embodiments of this application also provide a non-transient readable storage medium for storing a computer program. In one specific embodiment, the storage medium can be a USB flash drive, optical disc, hard disk, etc. The computer program stored on the storage medium can be executed by a processor to implement the aforementioned low-producing poultry identification method. To avoid repetition, further details will not be provided here.

[0078] Furthermore, one or more embodiments of this application also provide a computer program product, including a computer program that can be executed by a processor to implement the low-producing poultry identification method provided in the above embodiments. To avoid repetition, further details are omitted here.

[0079] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0080] In the 1990s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to methodology). However, with technological advancements, many methodological improvements today can be considered direct improvements to hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that a methodological improvement cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the operator programming the device. Designers can program a digital system themselves to "integrate" it onto a PLD, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0081] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0082] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0083] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, when implementing one or more embodiments of this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0084] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0085] Embodiments of this application are described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable parallel device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable parallel device, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0086] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable fraud device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0087] These computer program instructions can also be loaded onto a computer or other programmable device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable device for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0088] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0089] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0090] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0091] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the text "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0092] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, one or more embodiments of this application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, one or more embodiments of this application may take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0093] One or more embodiments of this application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. One or more embodiments of this application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can reside in local and remote computer storage media, including storage devices.

[0094] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0095] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

[0096] It should be understood that the training and prediction processes of the AI ​​models involved in the various embodiments of this application all adhere to the principles of legal data sources, compliant data content, and compliant data governance, and comply with the requirements of Article 5 of the Patent Law. Specifically: Data source legitimacy: All datasets used for AI model training were obtained through legal means, covering three categories: publicly authorized data, data authorized by partners, and self-collected compliant data. Publicly authorized data comes from compliant data sources following open-source licenses such as Apache 2.0, with complete copyright attribution and authorization scope clearly marked, and no unauthorized open-source code or data reuse. Data authorized by partners has been subject to formal data usage agreements, clearly defining the scope, duration, and confidentiality obligations, and possessing a complete authorization chain. For self-collected data involving personal information, strict informed consent procedures have been followed, and anonymization processes (including but not limited to field masking, feature anonymization, and differential privacy technology applications) have been implemented to remove personally identifiable information, fully complying with the requirements of relevant laws and regulations such as the "Interim Measures for the Administration of Generative Artificial Intelligence Services" and the "Personal Information Protection Law."

[0097] Data content compliance: The AI ​​model's dataset undergoes multiple screenings and cleaning processes to remove all content that may violate social morality or harm public interests. It contains no obscene, pornographic, violent, discriminatory, or information that endangers national or public safety, nor does it involve the illegal acquisition or use of genetic resources. For data in sensitive fields (such as healthcare and finance), an additional privacy-preserving computation module (including federated learning and secure multi-party computation technologies) ensures that the data is "usable but not visible," avoiding compliance risks during the original data transmission process and ensuring that the data application scenarios and uses comply with public order and good morals and industry regulatory requirements.

[0098] Data governance norms: A complete data traceability system is established during the AI ​​model training process to automatically record the source, collection time, annotation process, cleaning rules, and permission allocation of training data, generating traceable compliance reports to ensure that the data is verifiable throughout its entire lifecycle. The dataset annotation process for AI models is completed by a professional human R&D team, clearly defining the proportion of human creative contributions and avoiding reliance on AI-generated data that has not undergone substantial human modification, thus meeting the examination requirements for "human main contributions" in AI patent applications.

Claims

1. A method for identifying low-producing poultry, characterized in that, The method includes: Obtain the sequence of reading events of the target RFID tag read by the RFID reader in the egg-laying area during the current analysis period; wherein, the RFID reader in the egg-laying area reads the RFID tags present in the egg-laying area at a set time interval, the target RFID tag is the RFID tag set on the poultry to be identified, the reading events in the sequence of reading events are sorted according to the reading time of the reading event, and the reading event includes: the reading time and the RSSI value of the target RFID tag read; Based on the reading time interval between adjacent reading events in the reading event sequence, the reading event sequence is divided into multiple behavioral sessions, and session features of each behavioral session are generated based on the reading events within each behavioral session. Based on the session features of each behavioral session, a multi-dimensional feature vector of the poultry to be identified is generated. The multi-dimensional feature vector is then input into a low-production probability prediction model to predict the probability that the poultry to be identified is a low-production poultry. The low-production probability prediction model is trained based on the historical multi-dimensional feature vector of the sample poultry and the corresponding historical egg production performance labels. The historical egg production performance labels include: low-production labels and non-low-production labels. Based on the probability that the poultry to be identified is a low-producing poultry, determine whether the poultry to be identified is a low-producing poultry.

2. The method according to claim 1, characterized in that, The step of dividing the read event sequence into multiple behavior sessions based on the read time interval between adjacent read events in the read event sequence includes: Adjacent read events with a read time interval less than the interval threshold in the read event sequence are assigned to the same behavior session, while adjacent read events with a read time interval not less than the interval threshold are assigned to different behavior sessions, thus obtaining the multiple behavior sessions.

3. The method according to claim 1, characterized in that, The session features include: session duration and in-session activity. The session features for each behavioral session are generated based on read events within each of the multiple behavioral sessions, including: For each behavior session, perform the following processing: Calculate the reading time difference between the start reading event and the end reading event within the behavior session, and determine the reading time difference as the session duration of the behavior session; Calculate the variance of the RSSI values ​​of each read event within the behavior session, and determine the variance as the session activity of the behavior session.

4. The method according to claim 3, characterized in that, The generation of multi-dimensional feature vectors based on the session features of each behavioral session includes: Based on the session characteristics of each behavioral session, features of multiple dimensions are generated. These features include individual features and group relative features. The individual features include: the number of egg-laying behavioral sessions, the month-on-month change rate of the number of egg-laying behavioral sessions, the coefficient of variation of the session duration of egg-laying behavioral sessions, and the Shannon entropy of the probability distribution of the session start time. The egg-laying behavioral session is a behavioral session whose session duration is within the normal egg-laying session duration and whose activity level within the session is less than the activity level threshold. The group relative features include: the deviation of the individual features from the corresponding group baseline features. The features of the multiple dimensions are concatenated to generate the multi-dimensional feature vector.

5. The method according to claim 4, characterized in that, The group baseline features are determined based on the ranking of the corresponding individual features of each bird in the flock to which the bird to be identified belongs during the current analysis period.

6. The method according to claim 1, characterized in that, The step of determining whether a bird to be identified is a low-producing bird based on the probability that the bird is a low-producing bird includes: If the probability is greater than the probability threshold, then the poultry to be identified is determined to be abnormal in the current analysis period; If the poultry to be identified shows abnormalities for N consecutive analysis cycles, then the poultry to be identified is determined to be a low-producing poultry; wherein, N is greater than or equal to 3.

7. The method according to claim 6, characterized in that, The method further includes: If the poultry to be identified is a low-producing poultry, obtain the key features of the poultry to be identified for N analysis cycles. The key features include: the average session duration of the egg-laying behavior session in each analysis cycle, the number of egg-laying behavior sessions, and the deviation of the number of egg-laying behavior sessions from the baseline number of sessions in the flock. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is greater than the normal session duration threshold, and the number of egg-laying behavior sessions in the N analysis cycles is less than the first quantity threshold, then the low-production type of the poultry to be identified is determined to be pathological brooding. If the average session duration of the egg-laying behavior sessions in the N analysis cycles is within the normal session duration range, and the number of egg-laying behavior sessions in the N analysis cycles is less than the second quantity threshold, and the deviation of the number of egg-laying behavior sessions in the N analysis cycles from the baseline number of sessions in the flock is less than the deviation threshold, then the low-production type of the poultry to be identified is determined to be physiological low production.

8. A device for identifying low-producing poultry, characterized in that, The device includes: The acquisition module is used to acquire the sequence of reading events of the target RFID tag read by the RFID reader in the egg-laying area during the current analysis period; wherein, the RFID reader in the egg-laying area reads the RFID tags present in the egg-laying area at a set time interval, the target RFID tag is the RFID tag set on the poultry to be identified, the reading events in the sequence of reading events are sorted according to the reading time of the reading event, and the reading event includes: the reading time and the RSSI value of the target RFID tag read; The generation module is used to divide the reading event sequence into multiple behavior sessions based on the reading time interval of adjacent reading events in the reading event sequence, generate session features of each behavior session based on the reading events in each behavior session, and generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavior session. The prediction module is used to generate a multi-dimensional feature vector of the poultry to be identified based on the session features of each behavioral session, and input the multi-dimensional feature vector into the low-production probability prediction model to predict the probability that the poultry to be identified is a low-production poultry; wherein, the low-production probability prediction model is trained based on the historical multi-dimensional feature vector of the sample poultry and the corresponding historical egg production performance labels, the historical egg production performance labels including: low-production label and non-low-production label. The determination module is used to determine whether the poultry to be identified is a low-producing poultry based on the probability that the poultry to be identified is a low-producing poultry.

9. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program to implement the low-producing poultry identification method according to any one of claims 1 to 7.

10. A computer program product, characterized in that, Includes a computer program that can be executed by a processor to implement the low-producing poultry identification method according to any one of claims 1 to 7.

Images