Intelligent livestock farm production behavior supervision system and method
By establishing a feed surplus prediction model based on odor concentration and environmental data, the subjectivity problem of feed management in traditional animal husbandry has been solved, enabling scientific adjustment of feed supply and reducing resource waste and work disorder.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHANGSHA RUIHE DIGITAL TECH CO LTD
- Filing Date
- 2026-01-23
- Publication Date
- 2026-05-29
Smart Images
Figure CN122115142A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of monitoring production behavior in livestock farms, and in particular to a smart livestock farm production behavior monitoring system and method. Background Technology
[0002] In modern livestock farming, feed costs typically account for a large proportion of total costs and directly affect animal growth performance, health status, and production efficiency. As the direct implementers of daily operations on the farm, the actions of farm workers in storing, preparing, and feeding feed directly impact the effective utilization of feed and the nutritional intake of animals. Inappropriate feed management can lead to feed waste and spoilage, which in turn can cause animal diseases, reduce farming efficiency, and even threaten the quality and safety of livestock products and public health.
[0003] In traditional techniques, the amount and frequency of feed given by feeders are often judged based on past experience, which is highly subjective. Furthermore, it is impossible to immediately detect when there is leftover feed in livestock and to make appropriate reallocation of feed based on the amount of leftover feed, resulting in waste of resources and disorder in the arrangement of livestock feeding work. Summary of the Invention
[0004] To overcome the shortcomings of the existing technology, this application provides a smart livestock farm production behavior monitoring system and method.
[0005] Firstly, this application provides a method for monitoring the production behavior of a smart livestock farm, the method comprising: The target detection area where feed surplus was detected in the current period is obtained, and the historical database of detection in the historical period and the historical supply time period of feed distribution within a day in the historical period are obtained. Based on the historical database, the feed surplus of the target detection area in the first supply time of the historical time period is calculated, so as to detect the feed quality change odor concentration of the target detection area in the second supply time of the historical time period and obtain the odor concentration value. The variation trend of the odor concentration value 1 within the range of environmental data changes and intervals between the second and first supply times in the historical time period is statistically analyzed to obtain the concentration trend feature 1. Based on the concentration trend feature 1, the first feed supply remaining feature set is statistically analyzed. The odor concentration of feed quality change in the target detection area during the third supply time of the historical time period is detected to obtain odor concentration value two. The change trend of the odor concentration value two within the environmental data change value and interval between the third supply time and the second supply time of the historical time period is statistically analyzed to obtain concentration trend feature two. Based on concentration trend feature two, the second feed supply remaining feature set is statistically analyzed. Based on the first and second feed supply surplus feature sets, a feed surplus prediction model is established. An auxiliary testing area adjacent to the target detection area is selected. The feed surplus of the auxiliary testing area in the second supply time of the historical time period is predicted according to the feed surplus prediction model. This is to determine whether the feed surplus prediction model can be used to predict the feed surplus status in the current detection period, and output the verification result. Based on the verification results, the current supply time stage of the target detection area is obtained, and the pre-processed residual amount of feed remaining in the first supply time of the current time stage is detected. The reference predicted residual amount of feed remaining in the second supply time of the current supply time stage is predicted, so as to redistribute the feed supply amount of the second supply time of the current supply time stage, clean up the pre-processed residual amount, and output the feed supply management results.
[0006] Preferably, a target detection area where feed surplus is detected in the current period is obtained, and based on the target detection area, a historical database of detection in historical periods and the historical time period of feed distribution within a day in historical periods are obtained. Extract the historical feed supply amount from the historical database for the first supply time period within a single day in a historical period. The environmental data is obtained by analyzing the environmental data of the target detection area within the first supply time period of the historical time phase. If, during the second supply time of the historical feed supply period, it is detected that there is a remaining amount of the historical feed supply, the remaining amount to be measured is output. Based on the remaining quantity to be measured, the odor concentration value is obtained by detecting the odor concentration of feed quality change in the target detection area during the second supply time of the historical time period.
[0007] Preferably, the environmental data of the detection area of the target detection area in the second supply time period of the historical time period is obtained as environmental data two, and the difference between environmental data one and environmental data two is obtained as the first environmental change value to be measured. The first interval period is obtained by measuring the duration between the second supply period and the first supply period within the historical supply time period. The concentration trend characteristic is obtained by statistically analyzing the variation trend of the odor concentration value within the first interval period and the first environmental variation value. Based on the remaining quantity to be measured, extract the historical feed supply quantity 2, which is the second supply time of the historical time period within a day in the historical period, from the historical database. The historical feed supply quantity 2 is greater than the historical feed supply quantity 1. If, during the feeding of feed in the third feeding time of the historical feeding time period, it is detected that there is a remaining amount of the historical feed supply quantity two, the remaining amount to be measured is output. The remaining quantity to be measured and the concentration trend characteristics are combined to form the first feed supply remaining characteristic set.
[0008] Preferably, based on the remaining quantity to be measured, the odor concentration value is obtained by detecting the odor concentration of feed quality change in the target detection area at the third supply time of the historical time period. The environmental data of the target detection area in the third supply time period of the historical time period is obtained as environmental data three. The difference between environmental data two and environmental data three is obtained as the second environmental change value to be measured. The second interval period is obtained by measuring the duration of the interval between the third and second supply periods within the historical supply time period. The concentration trend characteristic two is obtained by statistically analyzing the variation trend of the odor concentration value two within the second interval period and the first environmental variation value. Based on the remaining quantity to be measured, extract the historical feed supply quantity three from the historical database for the third supply time of the historical time period within a day in the historical period. The historical feed supply quantity three is less than the historical feed supply quantity two. Calculate the remaining feed quantity based on the historical feed supply, and output the remaining feed quantity to be measured. The remaining amount to be measured and the concentration trend feature are combined to form the second feed supply remaining feature set.
[0009] Preferably, a feed surplus prediction model is established based on the first feed supply surplus feature set and the second feed supply surplus feature set. Select an auxiliary testing area adjacent to the target detection area, and measure the auxiliary testing environment variation value of the environmental data changes in the auxiliary testing area during the period between the second supply time and the first supply time in the historical time stage; The auxiliary environmental variation value, odor concentration value, and the first time interval are input into the feed remaining prediction model for testing to obtain the predicted remaining amount; Obtain the actual amount of feed remaining in the auxiliary testing area during the second supply period of the historical time phase; Based on the three remaining quantities to be measured, a preset remaining determination interval is set. If both the predicted remaining quantity and the actual remaining quantity are within the remaining determination interval, the remaining feed prediction model can be used to predict the remaining feed status during the current detection period, and a verification result is output.
[0010] Preferably, based on the verification result, the current time period of feed distribution in the target detection area within a day of the current period is obtained, and the pre-treated residual amount of feed remaining in the target detection area within the first supply time of the current time period is detected. It also detects the interval between the second supply time and the first supply time to which the current supply time period belongs, and obtains the preprocessed interval period; Preprocessed environmental change values for environmental data changes in the target detection area during the period between the second supply time and the first supply time, which is the current supply time phase; Based on the remaining amount of pretreatment, the pretreatment odor concentration value is obtained by predicting the odor concentration of feed quality change in the target detection area during the second supply time of the current supply time period. The pretreatment environmental variation value, pretreatment odor concentration value, and pretreatment remaining amount are input into the feed remaining prediction model for testing. The model predicts the remaining amount of feed in the target detection area during the second supply time of the current supply time period and outputs the reference predicted remaining amount. Based on the reference predicted remaining amount and the feed supply amount of the target detection area in the first supply time of the current supply period, the feed supply amount of the target detection area in the second supply time of the current supply period is redistributed, the pre-processed remaining amount is cleared, and the feed supply management results are output.
[0011] Secondly, a smart livestock farm production behavior monitoring system includes: The feed quality change analysis unit is used to obtain the target detection area where the amount of feed remaining is detected in the current period, and to obtain the historical database of detection in the historical period and the supply time period of feed distribution within a day in the historical period. Based on the historical database, the amount of feed remaining in the target detection area in the first supply time of the historical time period is counted, so as to detect the odor concentration of feed quality change in the target detection area in the second supply time of the historical time period and obtain the odor concentration value. The first unit of supply surplus characteristics is used to statistically analyze the change trend of odor concentration value 1 within the environmental data change value and interval between the second supply time and the first supply time in the historical time period to obtain concentration trend characteristic 1. Based on concentration trend characteristic 1, the first feed supply surplus characteristic set is statistically analyzed. The second unit of the surplus supply feature statistics is used to detect the odor concentration of feed quality change in the target detection area during the third supply time of the historical time period to obtain the odor concentration value two. It also calculates the change trend of the odor concentration value two within the environmental data change value and interval between the third supply time and the second supply time of the historical time period to obtain the concentration trend feature two. Based on the concentration trend feature two, the second feed surplus feature set is calculated. The prediction model building unit is used to build a feed surplus prediction model based on the first feed supply surplus feature set and the second feed supply surplus feature set, select an auxiliary test area adjacent to the target detection area, predict the feed surplus of the auxiliary test area in the second supply time of the historical time period according to the feed surplus prediction model, so as to determine that the feed surplus prediction model can be used to predict the feed surplus status in the current detection period, and output the verification result. The feed supply management unit is used to obtain the current supply time stage of the target detection area based on the verification results, and detect the pre-processed residual amount of feed remaining in the first supply time of the target detection area in the current time stage, so as to predict the reference predicted residual amount of feed remaining in the target detection area in the second supply time of the current supply time stage, so as to redistribute the feed supply in the second supply time of the current supply time stage, clean up the pre-processed residual amount, and output the feed supply management results.
[0012] Compared with the prior art, the present invention has the following characteristics and beneficial effects: By analyzing the historical feed supply and distribution status of the target detection area, along with environmental characteristics and the interval between the next feed supply, the qualitative changes in the remaining feed quantity can be determined. Furthermore, by analyzing the first, second, and third supply times within a single day's feed distribution history, and statistically analyzing the odor concentration trends caused by environmental changes and intervals during the intervals between the first and second supply times, the analysis also examines the odor concentration trends caused by environmental changes and intervals during the intervals between the second and third supply times. The features are statistically analyzed to train a feed surplus prediction model based on the two concentration trends observed under two different conditions. This model is then used to predict the feed surplus in the target detection area during the current period, enabling a more rational reallocation of feed supply for the next feeding period and reducing resource waste. To improve the accuracy of subsequent judgments, an auxiliary detection area adjacent to the target detection area is selected, and the feature information within this area is statistically analyzed to verify whether the feed surplus prediction model can be used to predict the feed surplus situation during the current detection period. This data processing method avoids the subjective judgment in feed supply allocation in traditional techniques and enhances the rigor of the entire data processing process. Attached Figure Description
[0013] Figure 1This is a flowchart illustrating the steps of a smart livestock farm production behavior monitoring method, which is the main feature of this embodiment.
[0014] Figure 2 This is a structural block diagram of a smart livestock farm production behavior monitoring system, which is the main feature of this embodiment. Detailed Implementation
[0015] The present invention will be further described in detail below with reference to the following embodiments.
[0016] Reference Figure 1 A method for monitoring production behavior in intelligent livestock farms, comprising the following steps: S1. Obtain the target detection area where feed surplus is detected in the current period, and obtain the historical database of detection in the historical period and the historical supply time period of feed distribution within a day in the historical period. According to the historical database, calculate the feed surplus of the target detection area in the first supply time of the historical time period, and detect the feed quality change odor concentration of the target detection area in the second supply time of the historical time period to obtain the odor concentration value one.
[0017] S2. Statistically determine the variation trend of odor concentration value 1 within the range of environmental data changes and intervals between the second and first supply times in the historical time period to obtain concentration trend feature 1. Based on concentration trend feature 1, statistically determine the first feed supply surplus feature set.
[0018] S3. Detect the odor concentration of feed quality change in the target detection area during the third supply time of the historical time period to obtain odor concentration value two. Calculate the change trend of odor concentration value two within the environmental data change value and interval between the third supply time and the second supply time of the historical time period to obtain concentration trend feature two. Based on concentration trend feature two, calculate the second feed supply remaining feature set.
[0019] S4. Based on the first and second feed supply surplus feature sets, establish a feed surplus prediction model, select an auxiliary testing area adjacent to the target detection area, and predict the feed surplus of the auxiliary testing area in the second supply time of the historical time period according to the feed surplus prediction model, so as to determine that the feed surplus prediction model can be used to predict the feed surplus status in the current detection period, and output the verification result.
[0020] S5. Based on the verification results, obtain the current supply time stage of the target detection area, and detect the pre-processed residual amount of feed remaining in the first supply time stage of the target detection area in the current time stage, so as to predict the reference predicted residual amount of feed remaining in the second supply time stage of the target detection area in the current supply time stage, so as to redistribute the feed supply amount in the second supply time stage of the current supply time stage, clean up the pre-processed residual amount, and output the feed supply management results.
[0021] Specifically, by analyzing the historical feed supply and distribution status of the target detection area, along with environmental characteristics and the interval between the next feed supply, the qualitative changes in the remaining feed quantity are determined. Furthermore, by analyzing the first, second, and third supply times within a single day's feed distribution history, the changes in feed quality odor concentration during the intervals between the first and second supply times are statistically analyzed, revealing the influence of environmental changes and interval durations on odor concentration trends. Additionally, the changes in feed quality odor concentration during the intervals between the second and third supply times are also analyzed, revealing the influence of environmental changes and interval durations on odor concentration trends. The trend characteristics are statistically analyzed to train a feed surplus prediction model based on the two concentration trend characteristics observed under two different conditions. This model is then used to predict the feed surplus in the target detection area during the current period, enabling a more rational reallocation of feed supply for the next feed supply period and reducing resource waste. To improve the accuracy of subsequent judgments, an auxiliary detection area adjacent to the target detection area is selected, and its characteristic information is statistically analyzed to verify whether the feed surplus prediction model can be used to predict the feed surplus situation during the current detection period. Through the above data processing method, the subjective judgment in feed supply allocation in traditional technologies is avoided, and the rigor of the entire data processing process is enhanced.
[0022] The specific step S1 includes the following sub-steps: The system identifies the target detection area where feed surplus is detected in the current period. Based on the target detection area, it obtains the historical database of detections in historical periods and the historical time period of feed distribution within a single day in those historical periods.
[0023] Extract the historical feed supply amount from the historical database for the first supply time period within a single day in a historical period.
[0024] The environmental data of the target detection area within the first supply time period of the historical time period is obtained as environmental data one.
[0025] If, during the second supply period of the historical feed supply phase, a surplus of historical feed supply quantity 1 is detected, the surplus quantity to be measured 1 is output.
[0026] Based on the remaining quantity to be measured, the odor concentration value is obtained by detecting the odor concentration of feed quality change in the target detection area during the second supply time of the historical time period.
[0027] Specifically, as should be noted, this includes monitoring livestock growth stages, health status, livestock species, feed formulation, and other characteristic information, all under uniform conditions: such as the same growth stage, normal health status, same livestock species, and same feed formulation; historical time period (e.g., feed distribution at three time periods within a day, such as morning, noon, and evening); historical feed supply amount 1 (e.g., the first feed supply is given at time T1 in the morning, if historical feed supply amount 1 is L1); environmental data 1 (if it is H1, including temperature: high temperature environments accelerate fat oxidation and protein denaturation in feed, producing off-flavors and reducing palatability. Conversely, low temperature environments cause feed to become damp, clump, and decompose nutrients, also affecting taste and quality); humidity: when humidity is too high, feed easily absorbs moisture from the air, not only changing the taste but also creating conditions for microbial growth, leading to mold and spoilage; light: excessive light will oxidize fats in feed, producing off-flavors and reducing nutritional value. Illumination can also decompose nutrients such as vitamins in feed, affecting the nutritional balance of the feed, which in turn leads to abnormal feeding and ventilation in livestock; poor ventilation will increase the concentration of harmful gases such as carbon dioxide and ammonia around the feed. These gases will cause chemical reactions in the feed, produce odors, reduce the quality and taste of the feed, and affect the appetite of livestock. The four types of data are: residual amount to be tested (such as a certain amount of feed remaining after the livestock have stopped eating; if a preset residual detection threshold is set (referring to the critical value when the residual amount of feed reaches the threshold for residual amount detection, the resulting qualitative change in odor will significantly affect the feeding status of livestock; this preset residual detection threshold is statistically derived from historical detection data); when the residual amount to be tested is greater than or equal to the residual detection threshold, the qualitative change in the residual amount of feed is analyzed); and odor concentration value 1 (i.e., the gas concentration of residual amount to be tested after the period from noon to morning, which can be automatically and accurately detected by gas sensors and monitoring equipment in existing technology).
[0028] The specific step S2 includes the following sub-steps: The environmental data of the target detection area within the second supply time period of the historical time stage is obtained as environmental data two. The difference between environmental data one and environmental data two is used to obtain the first environmental change value to be measured.
[0029] The first interval period is obtained by measuring the duration between the second supply time and the first supply time in the historical supply period.
[0030] The concentration trend characteristic is obtained by statistically analyzing the variation trend of the odor concentration value within the first interval period and the first environmental variation value.
[0031] Based on the remaining quantity to be measured, extract the historical feed supply quantity 2 from the historical database for the second supply time period of the historical period within a day. The historical feed supply quantity 2 is greater than the historical feed supply quantity 1.
[0032] If, during the third supply time of the historical feed supply period, a surplus of historical feed supply quantity two is detected, the surplus quantity to be measured, quantity two, is output.
[0033] The remaining quantity to be measured and the concentration trend characteristics are combined to form the first feed supply remaining characteristic set.
[0034] Specifically, environmental data 2 (interpreted the same as environmental data 1, e.g., H2), the first environmental variation value to be measured (i.e., the difference between the temperature, humidity, light intensity, and ventilation parameters in H2 and the temperature, humidity, light intensity, and ventilation parameters in H1; if the combined result is h1, then it is the first environmental variation value to be measured), the first interval period (the second supply time of the historical supply period is noon; if the time point is T2, then T2-T1 is the first interval period, e.g., t1), the first concentration trend feature (e.g., using the first interval period and the first environmental variation value to be measured as the x-axis, and the first odor concentration value as the y-axis, statistically analyzing the corresponding gas concentration variation curve path within the t1 time period, i.e., the first concentration trend feature), and the second remaining quantity to be measured (e.g., L2, and the second remaining quantity to be measured is the remaining quantity added on the basis of the first remaining quantity to be measured).
[0035] The specific step S3 includes the following sub-steps: Based on the remaining quantity to be measured, the odor concentration value 2 is obtained by detecting the odor concentration of feed quality change in the target detection area during the third supply time of the historical time period.
[0036] The environmental data of the target detection area within the third supply time period of the historical time stage is obtained as environmental data three. The difference between environmental data two and environmental data three is used to obtain the second environmental change value to be measured.
[0037] The second interval period is obtained by measuring the duration of the interval between the third supply time and the second supply time in the historical time period of statistical supply.
[0038] The trend of odor concentration value 2 within the second time interval and the first environmental variation value is statistically analyzed to obtain concentration trend feature 2.
[0039] Based on the remaining quantity to be measured, extract the historical feed supply quantity three from the historical database for the third supply time period of the historical period within a day. The historical feed supply quantity three is less than the historical feed supply quantity two.
[0040] Statistically analyze the historical feed supply and the remaining feed quantity, and output the remaining quantity to be measured.
[0041] The remaining amount to be measured and the concentration trend characteristics are combined to form the second feed supply remaining characteristic set.
[0042] Specifically, such as odor concentration value two (interpreted the same as odor concentration value one, so it will not be explained further here), environmental data three (interpreted the same as environmental data one, if it is H3), the second environmental variation value to be measured (that is, the temperature, humidity, light intensity, and ventilation parameters in H3 are subtracted from the temperature, humidity, light intensity, and ventilation parameters in H2. If the combined result is h2, it is the second environmental variation value to be measured), the second interval period (the second supply time of the supply history period is noon. If the time point is T3, T3-T2 is the second interval period, if it is t2), concentration trend feature two (if the second interval period and the second environmental variation value to be measured are used as the x-axis, and odor concentration value two is used as the y-axis, the corresponding gas concentration variation curve path in the time period t2 is statistically analyzed, i.e., concentration trend feature two), and the remaining quantity to be measured three (if it is L3, and the remaining quantity to be measured three is the remaining quantity added on the basis of remaining quantity to be measured one and remaining quantity to be measured two).
[0043] The specific step S4 includes the following sub-steps: A feed surplus prediction model is established based on the first feed supply surplus feature set and the second feed supply surplus feature set.
[0044] Select an auxiliary testing area adjacent to the target detection area, and measure the auxiliary testing environmental change value of the environmental data changes in the auxiliary testing area during the period between the second supply time and the first supply time in the historical time period.
[0045] The auxiliary environmental variation value, odor concentration value, and the first time interval are input into the feed surplus prediction model for testing to obtain the predicted surplus amount.
[0046] Obtain the actual amount of feed remaining in the second supply period of the historical time phase in the auxiliary testing area.
[0047] Based on the remaining amount to be measured (I), remaining amount to be measured (II), and remaining amount to be measured (III), a preset remaining judgment interval is set. If both the predicted remaining amount and the actual remaining amount are within the remaining judgment interval, the remaining feed prediction model can be used to predict the remaining feed status during the current detection period, and the verification result is output.
[0048] Specifically, for example, a feed surplus prediction model can be developed by integrating the two gas concentration variation curves (concentration trend feature one and concentration trend feature two) plotted above. If the two gas concentration variation curves show differences in gas concentration values under the same environmental variation values and intervals, and these differences are q1 and q2 respectively, then the average value (q1+q2) / 2 is calculated. If the average is Q1, then the original gas concentration value is replaced with Q1. If the two gas concentration variation curves show differences under different environmental variation values and intervals, then... When the corresponding gas concentration values differ or are the same, the two gas concentration variation curve paths are integrated and supplemented. Finally, a comprehensive gas concentration variation curve path is obtained after integrating the two gas concentration variation curve paths, which is used for subsequent comparative analysis of feature data. The auxiliary testing area (because livestock feeding in adjacent areas may be affected by the quality change in odor caused by the amount of feed left over from livestock in other adjacent areas, leading to a decrease in feed intake and feed surplus) uses the same historical time period feed allocation plan as the target detection area. When the same feed surplus situation occurs, the accuracy of the feed surplus prediction model can be verified by the feed surplus characteristic data in the auxiliary testing area, thereby enhancing the rigor of the data analysis process. The predicted surplus amount (that is, the auxiliary testing environment change value, odor concentration value one, and the first interval period are compared with the same category of data information on the path of a previously statistical comprehensive gas concentration change curve to match the corresponding feed surplus amount, i.e., the predicted surplus amount. It should be noted that the predicted surplus amount is not accurate. This predicted surplus amount is only used for data reference judgment. For example, the feed supply amount in the next time period can be redistributed according to the predicted surplus amount one, the remaining amount two, and the remaining amount three. For example, if there is a maximum value and a minimum value, the minimum value and the maximum value are combined to form the surplus judgment interval value). The verification result (if both the predicted surplus amount and the actual surplus amount are within the surplus judgment interval value, it means that the feed surplus prediction model has high accuracy, i.e., small error, and can be used to predict the feed surplus status in the current testing period).
[0049] The specific step S5 includes the following sub-steps: Based on the successful verification results, the current time period of feed distribution in the target detection area within a day of the current period is obtained, and the amount of pre-treated feed remaining in the target detection area within the first supply time of the current time period is detected.
[0050] The preprocessed interval period is obtained by detecting the interval between the second supply time and the first supply time in the current supply time period.
[0051] Preprocessed environmental change values for environmental data changes in the target detection area during the period between the second supply time and the first supply time, which is the current supply time period.
[0052] Based on the remaining amount of pretreatment, the pretreatment odor concentration value is obtained by predicting the odor concentration of feed quality change in the target detection area during the second supply time of the current supply period.
[0053] The pretreatment environmental variation value, pretreatment odor concentration value, and pretreatment residual amount are input into the feed surplus prediction model for testing. The model predicts the amount of feed surplus in the target detection area during the second supply time of the current supply period and outputs the reference predicted surplus amount.
[0054] Based on the reference predicted remaining amount and the feed supply amount of the target detection area in the first supply time of the current supply period, the feed supply amount of the target detection area in the second supply time of the current supply period is redistributed, the pre-processed remaining amount is cleared, and the feed supply management results are output.
[0055] Specifically, this includes the following parameters: the current supply time phase (explained similarly to the historical supply event phase, but it should be noted that the time point for each feed supply in the current supply time phase can correspond to the same or different time points in the historical supply event phase), the pre-treated remaining amount, the pre-treated interval period, and the pre-treated environmental variation value (all explained similarly to the remaining amount to be measured, the first interval period, and the first environmental variation value to be measured, but not specifically), and the pre-treated odor concentration value (e.g., extracting a portion of the time period from the pre-treated interval period, and this portion being close to the first supply time of the current supply time phase; based on the environmental variation value of the pre-treated remaining amount and the odor concentration variation value of the feed quality change generated within this portion of the time period, to predict the odor concentration variation value of the feed quality change in the remaining time period after removing this portion of the time period in the pre-treated interval period (i.e., prediction using a feed remaining prediction model), which also utilizes existing gas sensors and monitoring equipment). Automated and high-precision detection, by extracting a portion of the pre-treatment interval for analysis, is intended to enable timely measures to be taken before the next supply time. These measures refer to the timely clearing and processing of the pre-treatment surplus of livestock at the previous supply time. The system also includes a reference to the predicted surplus (explained similarly to the predicted surplus above, so it will not be elaborated upon here), and feed supply management results (such as timely transmission of the reference to the predicted surplus to the back end for reference by feed handlers; based on the reference to the predicted surplus, it can be determined that previous pre-treatment surplus still exists and needs to be cleared in a timely manner before supplying feed for the second supply time of the current time period; and the redistributed feed supply is slightly increased based on the feed supply for the first supply time of the current time period. For example, if the original setting for the first supply time of the current time period is G1, and the pre-treatment surplus is L3, then the redistributed feed supply can be G1 + L3 / 2).
[0056] A smart livestock farm production behavior monitoring system, by applying the aforementioned smart livestock farm production behavior monitoring method, includes a feed quality change analysis unit, a supply surplus characteristic statistics unit, a prediction model establishment unit, and a feed supply management unit, as described above. Figure 2The system employs a feed quality change analysis unit to identify the target detection area with remaining feed in the current period, and to access a historical database of feed distribution data for a given day. Based on this database, the system calculates the remaining feed in the target detection area during the first supply time of the historical time period. Then, it detects the feed quality change odor concentration during the second supply time of the historical time period to obtain odor concentration value one. A supply surplus characteristic statistics unit one calculates the trend of odor concentration value one within the environmental data variation and interval between the second and first supply times of the historical time period to obtain concentration trend characteristic one. Based on concentration trend characteristic one, a first feed supply surplus characteristic set is calculated. Finally, a supply surplus characteristic statistics unit two detects the feed quality change odor concentration in the target detection area during the third supply time of the historical time period to obtain odor concentration value two. The system also calculates the environmental data variation and interval between the third and second supply times of the historical time period for odor concentration value two. The trend of changes over a period of time yields the second concentration trend feature. Based on the second concentration trend feature, the second feed supply surplus feature set is statistically determined. The prediction model building unit establishes a supply surplus prediction model based on the first and second feed supply surplus feature sets. An auxiliary testing area adjacent to the target detection area is selected. The supply surplus prediction model predicts the feed surplus amount in the auxiliary testing area during the second supply time period in the historical time phase, thus determining that the supply surplus prediction model can be used to predict the feed surplus status during the current testing period. A verification result is output. The feed supply management unit, based on the verification result, obtains the current supply time phase of the target detection area and detects the pre-processed surplus amount of feed remaining in the target detection area during the first supply time phase of the current time phase. This predicts the reference predicted surplus amount of feed remaining in the target detection area during the second supply time phase of the current supply time phase. The feed supply amount during the second supply time phase of the current supply time phase is then redistributed, and the pre-processed surplus amount is cleared. The feed supply management result is output.
[0057] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.
Claims
1. A method for monitoring the production behavior of a smart livestock farm, characterized in that, Includes the following steps: The target detection area where feed surplus was detected in the current period is obtained, and the historical database of detection in the historical period and the historical supply time period of feed distribution within a day in the historical period are obtained. Based on the historical database, the feed surplus of the target detection area in the first supply time of the historical time period is calculated, so as to detect the feed quality change odor concentration of the target detection area in the second supply time of the historical time period and obtain the odor concentration value. The variation trend of the odor concentration value 1 within the range of environmental data changes and intervals between the second and first supply times in the historical time period is statistically analyzed to obtain the concentration trend feature 1. Based on the concentration trend feature 1, the first feed supply remaining feature set is statistically analyzed. The odor concentration of feed quality change in the target detection area during the third supply time of the historical time period is detected to obtain odor concentration value two. The change trend of the odor concentration value two within the environmental data change value and interval between the third supply time and the second supply time of the historical time period is statistically analyzed to obtain concentration trend feature two. Based on concentration trend feature two, the second feed supply remaining feature set is statistically analyzed. Based on the first and second feed supply surplus feature sets, a feed supply surplus prediction model is established. An auxiliary testing area adjacent to the target detection area is selected. The feed surplus of the auxiliary testing area in the second supply time of the historical time period is predicted according to the feed supply surplus prediction model. This is to determine whether the feed supply surplus prediction model can be used to predict the feed surplus status in the current detection period, and output the verification result. Based on the verification results, the current supply time stage of the target detection area is obtained, and the pre-processed residual amount of feed remaining in the first supply time of the current time stage is detected. The reference predicted residual amount of feed remaining in the second supply time of the current supply time stage is predicted, so as to redistribute the feed supply amount of the second supply time of the current supply time stage, clean up the pre-processed residual amount, and output the feed supply management results.
2. The method for supervising the production behavior of a smart livestock farm according to claim 1, characterized in that, The steps involve: acquiring the target detection area where feed surplus is detected in the current period; acquiring a historical database of feed distribution data for a given day in the historical period; statistically analyzing the feed surplus in the target detection area during the first supply time of the historical period based on the historical database; and detecting the odor concentration of feed quality changes in the target detection area during the second supply time of the historical period to obtain the odor concentration value 1. Obtain the target detection area where feed surplus is detected in the current period, and based on the target detection area, obtain the historical database of detection in the historical period and the historical time period of feed distribution within a day in the historical period. Extract the historical feed supply amount from the historical database for the first supply time period within a single day in a historical period. The environmental data is obtained by analyzing the environmental data of the target detection area within the first supply time period of the historical time phase. If, during the second supply time of the historical feed supply period, it is detected that there is a remaining amount of the historical feed supply, the remaining amount to be measured is output. Based on the remaining quantity to be measured, the odor concentration value is obtained by detecting the odor concentration of feed quality change in the target detection area during the second supply time of the historical time period.
3. The method for supervising the production behavior of a smart livestock farm according to claim 2, characterized in that, The steps for statistically analyzing the variation trend of the odor concentration value 1 within the historical time period from the second supply time to the first supply time, and obtaining the concentration trend feature 1, and then statistically analyzing the first feed supply remaining feature set based on the concentration trend feature 1, are as follows: The environmental data of the target detection area in the second supply time period of the historical time period is obtained as environmental data two. The difference between environmental data one and environmental data two is obtained as the first environmental change value to be measured. The first interval period is obtained by measuring the duration between the second supply period and the first supply period within the historical supply time period. The concentration trend characteristic is obtained by statistically analyzing the variation trend of the odor concentration value within the first interval period and the first environmental variation value. Based on the remaining quantity to be measured, extract the historical feed supply quantity 2 from the historical database, which is the second supply time of the historical time period within a day in the historical period. The historical feed supply quantity 2 is greater than the historical feed supply quantity 1. If, during the feeding of feed in the third feeding time of the historical feeding time period, it is detected that there is a remaining amount of the historical feed supply quantity two, the remaining amount to be measured is output. The remaining quantity to be measured and the concentration trend characteristics are combined to form the first feed supply remaining characteristic set.
4. The method for supervising the production behavior of a smart livestock farm according to claim 3, characterized in that, The steps for detecting the odor concentration of feed quality changes in the target detection area during the third supply time period in the historical time period are as follows: Odor concentration value two is obtained by detecting the odor concentration value two and calculating the trend of environmental data changes and intervals between the third and second supply times in the historical time period. Concentration trend feature two is then obtained. Based on concentration trend feature two, the remaining feature set of the second feed supply is calculated. Based on the remaining quantity to be measured, the odor concentration value is obtained by detecting the odor concentration of feed quality change in the target detection area during the third supply time of the historical time period. The environmental data of the target detection area in the third supply time period of the historical time period is obtained as environmental data three. The difference between environmental data two and environmental data three is obtained as the second environmental change value to be measured. The second interval period is obtained by measuring the duration of the interval between the third and second supply periods within the historical supply time period. The concentration trend characteristic two is obtained by statistically analyzing the variation trend of the odor concentration value two within the second interval period and the first environmental variation value. Based on the remaining quantity to be measured, extract the historical feed supply quantity three from the historical database for the third supply time of the historical time period within a day in the historical period. The historical feed supply quantity three is less than the historical feed supply quantity two. Calculate the remaining feed quantity based on the historical feed supply, and output the remaining feed quantity to be measured. The remaining amount to be measured and the concentration trend feature are combined to form the second feed supply remaining feature set.
5. The method for supervising the production behavior of a smart livestock farm according to claim 4, characterized in that, Based on the first and second feed supply surplus feature sets, a feed surplus prediction model is established. An auxiliary testing area adjacent to the target detection area is selected. The feed surplus prediction model predicts the feed surplus of the auxiliary testing area during the second supply period in the historical time frame, thus determining whether the model can be used to predict the feed surplus status during the current testing period. The specific steps for outputting the verification result are as follows: Based on the first and second feed supply surplus feature sets, a supply surplus prediction model is established. Select an auxiliary testing area adjacent to the target detection area, and measure the auxiliary testing environment variation value of the environmental data changes in the auxiliary testing area during the period between the second supply time and the first supply time in the historical time stage; The auxiliary environmental variation value, odor concentration value, and the first time interval are input into the supply remaining prediction model for testing to obtain the predicted remaining amount; Obtain the actual amount of feed remaining in the auxiliary testing area during the second supply period of the historical time phase; Based on the three remaining quantities to be measured, a preset remaining determination interval is set. If both the predicted remaining quantity and the actual remaining quantity are within the remaining determination interval, the remaining quantity prediction model can be used to predict the remaining feed status during the current testing period, and the verification result is output.
6. The method for supervising the production behavior of a smart livestock farm according to claim 5, characterized in that, Based on the verification results, the following steps are taken: First, obtain the current supply time phase of the target detection area and detect the pre-treated surplus amount of feed remaining in the first supply time phase of the target detection area. Then, predict the reference predicted surplus amount of feed remaining in the second supply time phase of the target detection area. Finally, reallocate the feed supply amount for the second supply time phase of the current supply time phase, clear the pre-treated surplus amount, and output the feed supply management results. Based on the verification results, the current time period of feed distribution in the target detection area within a day of the current period is obtained, and the amount of pre-treated feed remaining in the first supply time of the target detection area in the current time period is detected. It also detects the interval between the second supply time and the first supply time to which the current supply time period belongs, and obtains the preprocessed interval period; Preprocessed environmental change values for environmental data changes in the target detection area during the period between the second supply time and the first supply time, which is the current supply time phase; Based on the remaining amount of pretreatment, the pretreatment odor concentration value is obtained by predicting the odor concentration of feed quality change in the target detection area during the second supply time of the current supply time period. The pretreatment environmental variation value, pretreatment odor concentration value, and pretreatment remaining amount are input into the supply remaining prediction model for testing. The model predicts the remaining amount of feed in the target detection area during the second supply time of the current supply time period and outputs the reference predicted remaining amount. Based on the reference predicted remaining amount and the feed supply amount of the target detection area in the first supply time of the current supply period, the feed supply amount of the target detection area in the second supply time of the current supply period is redistributed, the pre-processed remaining amount is cleared, and the feed supply management results are output.
7. A smart livestock farm production behavior monitoring system, characterized in that, The system is used to implement the intelligent livestock farm production behavior monitoring method according to any one of claims 1-6, including: The feed quality change analysis unit is used to obtain the target detection area where the amount of feed remaining is detected in the current period, and to obtain the historical database of detection in the historical period and the supply time period of feed distribution within a day in the historical period. Based on the historical database, the amount of feed remaining in the target detection area in the first supply time of the historical time period is counted, so as to detect the odor concentration of feed quality change in the target detection area in the second supply time of the historical time period and obtain the odor concentration value. The first unit of supply surplus characteristics is used to statistically analyze the change trend of odor concentration value 1 within the environmental data change value and interval between the second supply time and the first supply time in the historical time period to obtain concentration trend characteristic 1. Based on concentration trend characteristic 1, the first feed supply surplus characteristic set is statistically analyzed. The second unit of the surplus supply feature statistics is used to detect the odor concentration of feed quality change in the target detection area during the third supply time of the historical time period to obtain the odor concentration value two. It also calculates the change trend of the odor concentration value two within the environmental data change value and interval between the third supply time and the second supply time of the historical time period to obtain the concentration trend feature two. Based on the concentration trend feature two, the second feed surplus feature set is calculated. The prediction model building unit is used to establish a feed supply surplus prediction model based on the first feed supply surplus feature set and the second feed supply surplus feature set, select an auxiliary test area adjacent to the target detection area, predict the feed surplus of the auxiliary test area in the second supply time of the historical time period according to the feed supply surplus prediction model, so as to determine that the feed supply surplus prediction model can be used to predict the feed surplus status of the current detection period, and output the verification result. The feed supply management unit is used to obtain the current supply time stage of the target detection area based on the verification results, and detect the pre-processed residual amount of feed remaining in the first supply time of the target detection area in the current time stage, so as to predict the reference predicted residual amount of feed remaining in the target detection area in the second supply time of the current supply time stage, so as to redistribute the feed supply in the second supply time of the current supply time stage, clean up the pre-processed residual amount, and output the feed supply management results.