Cow health prediction analysis method

Abstract

The invention relates to the technical field of dairy cow breeding, in particular to a dairy cow health prediction analysis method, which comprises the following steps: S1, acquiring feeding parameters of dairy cows, the feeding parameters comprising feeding amount, feeding time and feeding times; s2, according to the feeding parameters of the dairy cow, obtaining feeding characteristics of the feeding parameters, calculating to obtain a change trend of the feeding characteristics, and according to the change trend of the feeding characteristics, obtaining a preliminary health level of the dairy cow; s3, controlling a camera to obtain behavior detection data of the dairy cow, and determining a behavior decision level of the dairy cow according to the behavior detection data of the dairy cow; the behavior detection data comprises walking times, walking interval time and walking paths during feeding of the dairy cow; s4, according to the behavior decision level and the preliminary health level, determining a feeding decision level of the dairy cow during feeding; comprehensive evaluation of dairy cow health evaluation is realized, and the accuracy and efficiency of dairy cow health prediction are improved.

Description

Technical Field

[0001] The invention relates to the technical field of dairy cow breeding, and in particular to a dairy cow health prediction and analysis method. Background Art

[0002] In modern dairy farming, efficiency, health, scale and animal welfare have become the goals pursued by farming companies. These indicators are not isolated from each other. Only when dairy cows are comfortable and healthy can they better create greater economic benefits for farming companies.

[0003] However, for large-scale dairy farming, it is a big problem to monitor the health of individual cows. Workers need to observe the cows' movements in front of the monitoring screen in real time to judge their health status. However, this method has many defects. For example, it cannot guarantee that every cow can be monitored in real time, there may be omissions, and it is very time-consuming and labor-intensive, with very high labor costs. In addition, relying solely on manual observation to judge the health status is very subjective and may not be accurate enough.

[0004] For example, Chinese patent publication No. CN111611978A discloses a cow health monitoring system and method based on gait recognition, the system comprising: an image acquisition device for acquiring walking images of cows; a storage device for storing the ID of each cow and the morphological characteristics and standard gait characteristics of each cow; a controller for extracting the morphological characteristics and measured gait characteristics of the monitored cows from the walking images of the cows acquired by the image acquisition device, and comparing the measured gait characteristics of the monitored cows with their corresponding standard gait characteristics, inputting the comparison results into a trained health status prediction model, and obtaining the health status prediction results of the cows output by the health status prediction model; an alarm device for sending an alarm signal when the health status prediction results reach a preset alarm condition.

[0005] However, the existing technology only recognizes the gait of cows when predicting the health of cows, and cannot adjust the recognition of cows according to the different conditions of the cows themselves. At the same time, it ignores the impact of the cow's eating behavior on the cow's health, resulting in inaccurate recognition of the cow's health and failure to promptly discover potential problems with the cow. Summary of the invention

[0006] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for predicting and analyzing the health of dairy cows, comprising: S1, obtaining the feeding parameters of the dairy cows, the feeding parameters including the feeding amount, feeding time, and feeding frequency.

[0007] S2, obtaining eating characteristics of the eating parameters according to the eating parameters of the dairy cows, and calculating the changing trend of the eating characteristics, and obtaining the preliminary health grade of the dairy cow according to the changing trend of the eating characteristics.

[0008] S3, controlling the camera to obtain the cow's behavior detection data, and determining the cow's behavior decision level according to the cow's behavior detection data; the behavior detection data includes the cow's walking times, walking interval time, and walking path when eating.

[0009] S4, determining the eating decision level of the cow when eating based on the behavior decision level and the preliminary health level.

[0010] S5, obtaining the environmental data of the cows according to the temperature and humidity sensors, the environmental data including the temperature and humidity when the cows are eating; and calculating the environmental decision level of the cows according to the environmental data.

[0011] S6, using the milking equipment, determining the milk production of the cow corresponding to the feeding parameters, and obtaining a milk production analysis grade of the milk production.

[0012] S7, obtaining the final health grade of the dairy cow based on the feeding decision grade, the environmental decision grade and the milk production analysis grade.

[0013] The beneficial effects of the present invention are as follows: the present invention realizes a comprehensive assessment of the health status of dairy cows by integrating multi-dimensional data such as eating parameters, behavior detection data, and environmental data; accurately extracts key features and analyzes the changing trends of these features by means of changing trend analysis of eating characteristics and calculation of behavior coefficients; improves the ability to predict the health status of dairy cows based on time series analysis of changing trends of eating characteristics and behavior coefficients; evaluates the specific impact of environmental factors on the health of dairy cows by analyzing the combined values ​​of temperature and humidity and their impact on eating parameters, and provides a basis for optimizing the feeding environment; provides a more scientific health assessment method by comprehensively evaluating the eating decision level, environmental decision level, and milk production analysis level, and reasonably allocating the weights of each level, which helps ranch managers optimize management strategies and improve the health level and production efficiency of dairy cows. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0015] Figure 1 It is a flow chart of a method for predicting and analyzing dairy cow health.

[0016] Figure 2 The present invention is a flowchart of step S2 of a method for predicting and analyzing the health of dairy cows.

[0017] Figure 3 The present invention is a flowchart of step S3 of a method for predicting and analyzing the health of dairy cows.

[0018] Figure 4 The present invention is a flowchart of step S5 of a method for predicting and analyzing the health of dairy cows. DETAILED DESCRIPTION

[0019] The embodiments of the present invention are described in detail below. The embodiments described below are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. If no specific techniques or conditions are specified in the embodiments, the techniques or conditions described in the literature in the art or the product specifications are used.

[0020] See also Figure 1 A method for predicting and analyzing the health of dairy cows includes: S1, obtaining the feeding parameters of dairy cows, the feeding parameters including feeding amount, feeding time, and feeding frequency.

[0021] S2, obtaining eating characteristics of the eating parameters according to the eating parameters of the dairy cows, and calculating the changing trend of the eating characteristics, and obtaining the preliminary health grade of the dairy cow according to the changing trend of the eating characteristics.

[0022] S3, controlling the camera to obtain the cow's behavior detection data, and determining the cow's behavior decision level according to the cow's behavior detection data; the behavior detection data includes the cow's walking times, walking interval time, and walking path when eating.

[0023] S4, determining the eating decision level of the cow when eating based on the behavior decision level and the preliminary health level.

[0024] S5, obtaining the environmental data of the cows according to the temperature and humidity sensors, the environmental data including the temperature and humidity when the cows are eating; and calculating the environmental decision level of the cows according to the environmental data.

[0025] S6, using the milking equipment, determining the milk production of the cow corresponding to the feeding parameters, and obtaining a milk production analysis grade of the milk production.

[0026] S7, obtaining the final health grade of the dairy cow based on the feeding decision grade, the environmental decision grade and the milk production analysis grade.

[0027] At this time, the preliminary health level set in step S2 is divided into four levels, which are represented by A, B, C, and D respectively. The contents represented by these four levels can be as described below.

[0028] For example, health level A: the amount of food eaten is close to or reaches the recommended daily intake standard; the distribution of eating time is reasonable, and the time of each eating is moderate; the number of eating times meets the recommended eating frequency, and the number of eating times per day is stable; the cows show good appetite, normal eating behavior, good health, and stable milk production.

[0029] Health level B: The amount of food eaten is slightly lower than the recommended daily intake standard, but not less than 80%; the distribution of eating time is basically reasonable, with occasional eating time being too short or too long; the number of times eaten is basically in line with the recommended frequency of eating, but occasionally the number of times eaten is unstable; the cow has a good appetite, but there may be slight nutritional deficiencies or slightly low digestion and absorption efficiency, and the overall health status is acceptable.

[0030] Health level C: The amount of food consumed is significantly lower than the recommended daily intake standard, less than 80%, but not less than 60%; the distribution of eating time is uneven, and eating time is frequently too short or too long; the number of eating times is unstable, and the number of eating times is often too little or too much; the cow may have poor appetite or decreased digestive system function, and its health status needs to be paid attention to, and consideration should be given to adjusting the feed formula or increasing the inspection frequency.

[0031] Health level D: The amount of food eaten is far below the recommended daily intake standard, less than 60%; the eating time is extremely irregular, and the eating time is often too short or almost no eating; the eating frequency is seriously insufficient or excessive, showing obvious loss of appetite; the cow shows severe loss of appetite and may be suffering from a disease, which requires immediate veterinary examination and appropriate treatment measures.

[0032] The amount of food eaten can be known by the amount of feed placed every day; the eating time can be obtained by the time when the cow's head enters and leaves the feeding trough through the camera; the number of times the cow's head eats can be determined by the number of times the cow's head enters the feeding trough. When the time of entering the feeding trough is less than 10s, the number of times is not accumulated, and only the corresponding time is recorded for subsequent analysis.

[0033] like Figure 2 As shown, the method of determining the preliminary health level of the cow in step S2 is as follows: S21, obtaining an eating time series in which eating parameters change over time, and constructing a time series according to the corresponding values ​​of the eating parameters in the form of time points to obtain an eating time series.

[0034] S22, analyzing the eating time series to obtain eating characteristics of the eating time series; the eating characteristics are represented by a moving average of the eating amount, a difference value of the eating frequency, and a distribution probability of the eating time.

[0035] S23, performing trend analysis on the eating characteristics to determine the changing trend of the eating characteristics, calculating the trend amount of the changing trend of the eating characteristics according to the changing trend of the eating characteristics, and obtaining a preliminary health grade according to the trend amount of the changing trend of the eating characteristics.

[0036] The moving average of the feeding amount in the feeding characteristics at this time is expressed as the average of the feeding amount in the three consecutive days. The difference in the number of feedings is the difference between the number of feedings on the current day and the previous day. The distribution probability of the feeding time represents the probability of the time period in which the cow eats. This probability is expressed as the ratio of the number of feedings in the time period corresponding to the feeding time to the total number of feedings.

[0037] The changing trend of the eating characteristics is expressed by calculating the change in the eating characteristics according to the progress of time; and setting the eating characteristics to the corresponding preliminary health level. The trend amount of the changing trend here will select different values ​​according to different eating characteristics.

[0038] The trend quantity of the change trend of the eating characteristics is expressed as follows: the trend slope of the eating amount is calculated based on the moving average value of the eating amount, the average difference value of the eating times is calculated based on the difference value of the eating times, and the eating probability value in the main time period is calculated based on the distribution probability of the eating time; the main time period is the three time periods with the largest distribution probability of the eating time in the historical data, and these three time periods will be rounded to ensure that the main time period at this time is at least expressed as an hour in length; the eating probability value is the sum of the corresponding probabilities of these three time periods; the eating probability value in this main time period can reflect the stability and regularity of the cow's eating at different times, so that it can be found whether the cow eats according to normal rules when eating; and according to the value of the trend quantity at this time, the trend quantity of the change trend of the eating characteristics is compared with the preset standard value in the database to obtain a preliminary health level.

[0039] For example, health level A: food intake trend slope > 0.3 kg / day; average difference in food intake times < 0.5; probability of eating in the main time period > 0.7.

[0040] Health level B: 0.1kg / day < food intake trend slope ≤ 0.3kg / day; 0.5 ≤ average difference in food intake frequency < 1.0; 0.5 ≤ probability of eating in the main time period < 0.7.

[0041] Health level C: -0.1kg / day < food intake trend slope ≤ 0.1kg / day; 1.0 ≤ average difference in food intake frequency < 1.5; 0.3 ≤ probability of eating in the main time period < 0.5.

[0042] Health level D: Food intake trend slope ≤ -0.1kg / day; average difference in food intake frequency ≥ 1.5; probability of eating in the main time period < 0.3.

[0043] When these values ​​are met, the preliminary health characteristics corresponding to the cow's eating at this time are obtained; at this time, the characteristics related to eating can be obtained to reflect whether the cow has abnormal conditions when eating, and the health level corresponding to the abnormal conditions.

[0044] By monitoring feeding parameters, health problems in dairy cows can be detected early; for example, a sudden decrease in feed intake, fewer feedings, or a shorter feeding time may be a sign of deteriorating health.

[0045] According to the changes in eating parameters, feeding management measures can be adjusted in time; for example, increasing the quality of feed, adjusting feeding time or changing feed formula to meet the nutritional needs of dairy cows; by monitoring eating behavior, it can be ensured that dairy cows get adequate nutrition, thereby increasing their milk production and quality; by timely detecting abnormal eating behavior, preventive measures can be taken to reduce the occurrence of diseases; for example, by improving the feeding environment and strengthening health management, the risk of disease can be reduced.

[0046] Reasonable feeding parameters can help improve the welfare of dairy cows. For example, appropriate feeding frequency and time can make cows feel comfortable and reduce stress reactions.

[0047] In step S3, a camera is set to detect the eating behavior of the cows, and detect the number of times the cows walk, the walking interval and the walking path each time they eat; the number of times the cows walk can evaluate the activity level of the cows. Frequent walking may indicate that the cows are uncomfortable or in pain, while less walking may indicate that their activity ability is reduced. The walking interval can detect whether the cows are tired or uncomfortable. Shorter walking intervals may indicate that the cows are in a state of tension or fatigue, and longer intervals may indicate that their activity ability is limited. The walking path is used to detect the activity range of the cows. The walking path is represented by the distance between the starting point of the cow's walking when eating and the end point corresponding to the feeding trough; when the walking path is abnormal, it means that the cows have some behavioral abnormalities or psychological pressure.

[0048] like Figure 3 As shown, step S3 also includes the following implementation: S31, inputting the number of walking times, walking interval time, and walking path into a walking path set, and the number of walking times, walking interval time, and walking path correspond to the elements in the walking path set one by one.

[0049] S32, sequentially processing the walking frequency, walking interval time, and walking path of the cow when eating, and respectively obtaining a first behavior coefficient of the walking frequency, a second behavior coefficient of the walking interval, and a third behavior coefficient of the walking path.

[0050] S33, obtaining a behavior decision level according to the first behavior coefficient, the second behavior coefficient, and the third behavior coefficient.

[0051] When calculating the first behavior coefficient, the second behavior coefficient, and the third behavior coefficient, the data of the cow at this time is divided according to the form of time points to obtain the corresponding behavior coefficients; at this time, the obtained walking times, walking intervals, and walking paths are input into the walking path set, and the walking times, walking intervals, and walking paths are matched one by one with the elements in the walking path set to determine the walking times, walking intervals, and walking paths corresponding to each element, so as to complete the evaluation of the cow's walking while eating.

[0052] The first behavior coefficient is expressed as shown below. A walking path set corresponding to the walking number is obtained, and the first behavior coefficient is calculated.

[0053] Among them, BC1 represents the first behavior coefficient, NW i Indicates the number of walks corresponding to the i-th element in the walking path set, μ NW represents the average number of walks, σ NW represents the standard deviation of the number of walks, n represents the number of elements in the walking path set, and the value of i ranges from 1 to n.

[0054] The second behavior coefficient is expressed as shown below: a walking path set corresponding to the walking interval time is obtained, and the second behavior coefficient is calculated.

[0055] Among them, BC2 represents the second behavior coefficient, WI i represents the walking interval time corresponding to the i-th element in the walking path set, μ WI represents the average walking interval time, σ WI represents the standard deviation of the walking interval time, n represents the number of elements in the walking path set, and the value range of i is 1 to n.

[0056] At this time, compared with the number of walking times and walking intervals, the walking path is more inclined to record the movement trajectory of the cow on the path, and whether there is a problem of not eating or not eating; then the way to judge the walking path at this time is to judge the ratio of the maximum moving path of the cow when walking and the length of the walking path, so as to determine whether the cow has a problem with the movement range at this time.

[0057] The processing method for the third behavior coefficient is shown below.

[0058] All cows are regarded as a point, and the walking path of the cows eating at this time is compared with the corresponding points of the cows, and the point closest to other cows in the current cow's walking path is found as the mark point, and the shortest path sum between the mark points is calculated. The shortest path sum between the mark points is obtained by calculating the shortest path between each mark point in turn and summing up all the shortest paths; the shortest path sum between the mark points is compared with the shortest path corresponding to the starting point and the end point of the walking path to obtain a first proportional value; the shortest path sum between the mark points is compared with the total length of the walking path to obtain a second proportional value, and the walking path is processed according to the first proportional value and the second proportional value to obtain a third behavior coefficient.

[0059] The third behavior coefficient can be expressed as follows: the occurrence probability of the first proportion value and the occurrence probability of the second proportion value in the walking path set are determined in sequence to obtain the third behavior coefficient.

[0060] Among them, BC3 represents the third behavior coefficient, p 1,i represents the probability of occurrence of the first proportion value corresponding to the i-th element in the walking path set, p 2,i represents the probability of occurrence of the second proportional value corresponding to the i-th element in the walking path set, e represents the exponential constant, n represents the number of elements in the walking path set, and the value range of i is 1 to n.

[0061] The implementation method of step S33 includes normalizing the first behavior coefficient, the second behavior coefficient, and the third behavior coefficient to calculate the behavior decision level.

[0062] Among them, DL represents the behavior decision level, BC1 represents the first behavior coefficient, BC2 represents the second behavior coefficient, BC3 represents the third behavior coefficient, w1 represents the weight of the first behavior coefficient, w2 represents the weight of the second behavior coefficient, w3 represents the weight of the third behavior coefficient, and e represents the exponential constant.

[0063] At this time, the behavioral decision level and the preliminary health level will be set accordingly. The behavioral decision level is also divided into four levels. The four levels will use A, B, C, and D to correspond to the preliminary health level, and by setting a score value for each level, the scores of the two levels can be combined to obtain the eating decision level to judge the cow's desire for eating and eating-related data.

[0064] For example, after the behavior decision level is calculated, the value of the behavior decision level will be used as the main classification condition, and the classification of the behavior decision level at this time will be obtained according to the values ​​of the number of walking times, walking interval time, and walking path.

[0065] Behavioral decision level A: The number of walking times per day is between 25 and 30 times; the walking interval is between 10 and 20 minutes; most of the walking path is within a specific area; the behavioral decision level value is between 0.85 and 1.0.

[0066] Behavioral decision level B: The number of walking times per day is between 20-25 times; the walking interval is between 20-30 minutes; most of the time the walking path is within a specific area, but there is slight irregularity; the behavioral decision level value is between 0.65-0.85.

[0067] Behavioral decision level C: The number of walking times per day is between 15-20 times; the walking interval is between 30-40 minutes; the range of walking paths is limited, and the behavioral decision level is between 0.45-0.65.

[0068] Behavioral decision level D: walking less than 15 times a day; walking intervals exceed 40 minutes; walking paths are extremely irregular, and the behavioral decision level is between 0-0.45.

[0069] The behavioral decision level can assist in detecting abnormalities in cows’ eating and determine whether cows have abnormal walking problems while eating, thereby optimizing feeding management and improving cow production efficiency.

[0070] The behavioral decision level described above can tell the corresponding amount of exercise and activity status of the cow before and after eating, so as to judge whether the cow lacks vitality or has health problems, and judge whether the cow's life rhythm is regular based on this behavioral decision level.

[0071] In one case of the present invention, the behavioral decision level and the preliminary health level will be matched according to the corresponding grading conditions to obtain the eating decision level at this time, such as behavioral decision level A and preliminary health level A are grouped together, behavioral decision level B and preliminary health level B are grouped together, behavioral decision level C and preliminary health level C are grouped together, and behavioral decision level D and preliminary health level D are grouped together to obtain the eating decision level at this time.

[0072] Alternatively, the current eating decision level can be obtained based on the corresponding values ​​of the behavior decision level and the initial health level to evaluate the current eating plan of the dairy cow.

[0073] When calculating the eating decision level in step S4, since the preliminary health level is obtained by comparing the trend amount of the eating characteristics, the trend amount of the eating characteristics needs to be calculated at this time to determine the value of the preliminary health characteristics at this time, and the behavior decision level is combined with the calculated preliminary health level to obtain the output eating decision level.

[0074] Where EDL represents the eating decision level, λ HL The grade factor representing the preliminary health grade, λ DL represents the level factor of the behavioral decision level, DL represents the behavioral decision level, FA represents the trend slope of the food intake, FA′ represents the standard value of the trend slope of the food intake, FF represents the average difference value of the number of eating times, FF′ represents the standard value of the average difference value of the number of eating times, p(F) represents the probability of eating in the main time period, and p(F)′ represents the standard value of the probability of eating in the main time period; at this time, the standard value of the trend slope of the food intake, the average difference value of the number of eating times, and the standard value of the probability of eating in the main time period represent the standard deviation of the corresponding values ​​in the historical data, and this standard deviation is used as the standard value used at this time; λ1 represents the weight of the trend slope of the food intake, λ2 represents the weight of the average difference value of the number of eating times, and λ3 represents the weight of the probability of eating in the main time period; at this time, the weights of the trend slope of the food intake, the average difference value of the number of eating times, and the probability of eating in the main time period are selected in order as 0.375, 0.25, and 0.375.

[0075] The level factors of the behavioral decision level and the preliminary health level are set according to the level they are in. The values ​​of the level factors set for the four levels of A, B, C, and D in the behavioral decision level are 0.45, 0.32, 0.18, and 0.05, respectively; the values ​​of the level factors set for the four levels of A, B, C, and D in the preliminary health level are 0.4, 0.3, 0.2, and 0.1, respectively.

[0076] The calculated value at this time is output as the eating decision level. The eating decision level combines multiple behavioral and health indicators and can comprehensively reflect the performance of dairy cows during the eating process. Through the changes in the eating decision level, abnormalities in the eating behavior of dairy cows can be discovered early, and possible health or environmental problems can be intervened in time.

[0077] The environmental decision level recorded in step S5 is mainly reflected in the effect of cows eating at different temperatures. For example, the number of times cows eat is stable, but when the temperature changes to a fixed range, the number of times cows eat will become unstable. At this time, an effect value will be set to quantify the impact of temperature on cows eating. At the same time, it will also be determined whether the humidity changes will affect the number of times cows eat and the corresponding amount of food eaten. The environmental decision level is set according to the influencing factors obtained in this part, so as to obtain relevant data of cows in different environments, and assist in determining how to adjust the cows' eating situation at this time.

[0078] like Figure 4 As shown, the method for obtaining the environmental decision level in step S5 can be as shown below.

[0079] S51, obtain the combination value of temperature and humidity at the corresponding time point to form a temperature and humidity parameter set; here the combination value represents the value of temperature and humidity at a time point, and regards these two values ​​as a combination, and the value of this combination is the combination value. For example, when eating, the temperature of the external environment is 20°C and the humidity is 40%. At this time, the combination value is (20°C, 40%), and the temperature is represented in front of the combination value, and the humidity is represented in the back; determine whether the food intake, eating time, and number of meals will change when this combination value appears; at this time, the probability form can be used to indicate whether the food intake, eating time, and number of meals will change due to different temperature and humidity values.

[0080] S52, obtain the food intake, eating time, and eating frequency corresponding to each combination value in the temperature and humidity set; calculate the conditional probability of food intake, eating time, and eating frequency with the combination value, and output the comprehensive value of the conditional probability of food intake, eating time, and eating frequency as the value of the environmental decision level.

[0081] The conditional probability of food intake is expressed as follows.

[0082] Among them, P(FI|(T,H)) represents the conditional probability of food intake, which means the probability of food intake FI under the condition that temperature and humidity are (T,H); FI represents food intake, T represents temperature, H represents humidity, P(FI∩(T,H)) represents the probability of temperature and humidity (T,H) and food intake FI occurring at the same time; P(T,H) represents the probability of temperature and humidity (T,H).

[0083] The conditional probability of eating time is expressed as shown below.

[0084] Among them, P(ET|(T,H)) represents the conditional probability of eating time, which means the probability of eating time ET under the condition that temperature and humidity are (T,H); ET represents eating time, T represents temperature, H represents humidity, P(ET∩(T,H)) represents the probability of temperature and humidity being (T,H) and eating time ET occurring at the same time; P(T,H) represents the probability of temperature and humidity being (T,H).

[0085] The conditional probability of the number of times of eating is expressed as follows.

[0086] Among them, P(EF|(T,H)) represents the conditional probability of the number of times of eating, which means the probability of eating number EF under the condition that the temperature and humidity are (T,H); EF represents the number of times of eating, T represents the temperature, H represents the humidity, P(EF∩(T,H)) represents the probability of the temperature and humidity being (T,H) and the number of times of eating EF occurring at the same time; P(T,H) represents the probability of the temperature and humidity being (T,H).

[0087] The environmental decision level is expressed as shown below.

[0088]

[0089] Where SDL represents the environmental decision level, m represents the number of elements in the temperature and humidity set, j ranges from 1 to m, and P(FI|(T,H) j ) represents the conditional probability of food intake corresponding to the jth element in the temperature and humidity set; P(ET|(T,H) j ) represents the conditional probability of eating time corresponding to the jth element in the temperature and humidity set, P(EF|(T,H) j ) represents the conditional probability of the number of times the jth element in the temperature and humidity set corresponds to the number of times the food is eaten, μ FI represents the average value of the conditional probability of food intake, σ FI represents the standard deviation of the conditional probability of food intake; μ ET represents the mean value of the conditional probability of eating time, σ ET represents the standard deviation of the conditional probability of eating time, μ EF represents the average value of the conditional probability of eating times, σ EF represents the standard deviation of the conditional probability of eating times; w FI represents the weight of food intake, w ET represents the weight of eating time, w EF Indicates the weight of the number of times you eat.

[0090] The obtained environmental decision level will indicate the corresponding probabilities of the corresponding amount of food intake, frequency of food intake, and time of food intake when the temperature and humidity are at different values. This will reveal how cows generally eat under the influence of the current environment. Based on this data, we can analyze the pattern of cows' eating behavior and find out what temperature has the greatest impact on cows' eating, thereby causing a decline in the overall health level of cows.

[0091] The milk production analysis level in step S6 is expressed as the ratio of food intake to milk production per kilogram of body weight; milk production per kilogram of body weight is a data value obtained by dividing milk production by the weight of the cow. The milk production analysis level is obtained based on the ratio of food intake to milk production per kilogram of body weight.

[0092] The milk analysis grades are shown below.

[0093] Among them, MG represents the milk production analysis grade, FMR represents the ratio of feed intake to milk production per kilogram of body weight, μ FMR represents the average value of the ratio of food intake to milk production per kilogram of body weight, σ FMR It expresses the standard deviation of the ratio of feed intake to milk production per kilogram of body weight.

[0094] At this time, the milk production of the corresponding cows under different feeding decision levels is determined, and the previously calculated feeding parameter values ​​are associated with the current milk production to determine the setting relationship between milk production and feeding parameters, thereby determining the milk production analysis level corresponding to the milk production at this time to evaluate under what feeding conditions the cows can achieve the expected milk production.

[0095] The final health grade in step S7 can be expressed as a weighted average of the eating decision grade, the environment decision grade and the milk production analysis grade.

[0096] Therefore, the final health level is expressed as shown below.

[0097] Among them, FHL represents the final health level, EDL represents the feeding decision level, SDL represents the environmental decision level, MG represents the milk production analysis level, and w EDL represents the weight of the eating decision level, w SDL represents the weight of the environmental decision level, w MG represents the weight of the milk production analysis grade, and e represents the exponential constant.

[0098] The weights set at this time are selected as 0.4, 0.2, and 0.4 in the order of feeding decision level, environmental decision level, and milk production analysis level.

[0099] This completes the predictive analysis of the health of dairy cows. Based on the final health level obtained, the health status of dairy cows can be monitored in a timely manner to prevent abnormal conditions of dairy cows and corresponding abnormal problems.

[0100] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention and they are still covered by the protection scope of the present invention.

Claims

1. A method for predicting and analyzing the health of dairy cows, characterized in that: include: S1, obtaining the feeding parameters of the dairy cows, including feeding amount, feeding time, and feeding frequency; S2, obtaining eating characteristics of the eating parameters according to the eating parameters of the dairy cows, and calculating the changing trend of the eating characteristics, and obtaining the preliminary health grade of the dairy cow according to the changing trend of the eating characteristics; S3, controlling the camera to obtain the cow's behavior detection data, and determining the cow's behavior decision level based on the cow's behavior detection data; the behavior detection data includes the number of times the cow walks when eating, the walking interval time, and the walking path; S4, determining the eating decision level of the cows when eating based on the behavioral decision level and the preliminary health level; S5, obtaining environmental data of the cows according to the temperature and humidity sensors, the environmental data including the temperature and humidity when the cows are eating; and calculating the environmental decision level of the cows according to the environmental data; S6, using milking equipment, determining the milk production of the cow corresponding to the feeding parameters, and obtaining a milk production analysis grade of the milk production; S7, obtaining the final health grade of the dairy cow based on the feeding decision grade, the environmental decision grade and the milk production analysis grade.

2. A method for predicting and analyzing the health of dairy cows according to claim 1, characterized in that: Step S2 also includes: S21, obtaining an eating time series of eating parameters changing over time; S22, analyzing the eating time series to obtain eating characteristics of the eating time series; the eating characteristics are represented by a moving average of the eating amount, a difference value of the eating frequency, and a distribution probability of the eating time; S23, performing trend analysis on the eating characteristics to determine the changing trend of the eating characteristics, calculating the trend amount of the changing trend of the eating characteristics according to the changing trend of the eating characteristics, and obtaining a preliminary health grade according to the trend amount of the changing trend of the eating characteristics.

3. A method for predicting and analyzing the health of dairy cows according to claim 2, characterized in that: The trend quantity of the changing trend of eating characteristics is expressed as follows: the trend slope of eating amount is calculated based on the moving average of eating amount, the average difference value of eating number is calculated based on the difference value of eating number, and the probability value of eating in the main time period is calculated based on the distribution probability of eating time.

4. A method for predicting and analyzing the health of dairy cows according to claim 1, characterized in that: The step S3 also includes the following implementation: S31, inputting the number of walking times, the walking interval time, and the walking path into a walking path set, wherein the number of walking times, the walking interval time, and the walking path correspond to the elements in the walking path set one by one; S32, sequentially processing the walking times, walking intervals, and walking paths of the dairy cows when they are eating, and obtaining a first behavior coefficient of the walking times, a second behavior coefficient of the walking intervals, and a third behavior coefficient of the walking paths, respectively; S33, obtaining a behavior decision level according to the first behavior coefficient, the second behavior coefficient, and the third behavior coefficient.

5. A method for predicting and analyzing the health of dairy cows according to claim 4, characterized in that: The implementation method of step S33 includes normalizing the first behavior coefficient, the second behavior coefficient, and the third behavior coefficient to calculate the behavior decision level: Among them, DL represents the behavior decision level, BC1 represents the first behavior coefficient, BC2 represents the second behavior coefficient, BC3 represents the third behavior coefficient, w1 represents the weight of the first behavior coefficient, w2 represents the weight of the second behavior coefficient, w3 represents the weight of the third behavior coefficient, and e represents the exponential constant.

6. A method for predicting and analyzing the health of dairy cows according to claim 3, characterized in that: The eating decision hierarchy is represented as follows: Where EDL represents the eating decision level, λ HL The grade factor representing the preliminary health grade, λ DL , λ1 represents the weight of the trend slope of food intake, λ2 represents the weight of the average difference in the number of eating times, and λ3 represents the weight of the probability of eating in the main time period.

7. A method for predicting and analyzing the health of dairy cows according to claim 1, characterized in that: The method for obtaining the environmental decision level in step S5 can be as follows: S51, obtaining the combined values ​​of temperature and humidity at corresponding time points to form a temperature and humidity parameter set; S52, obtain the food intake, food time, and food frequency corresponding to each combination value in the temperature and humidity set; calculate the conditional probability of food intake, food time, and food frequency with the combination value, and output the comprehensive value of the conditional probability of food intake, food time, and food frequency as the value of the environmental decision level.

8. A method for predicting and analyzing the health of dairy cows according to claim 7, characterized in that: The environmental decision level is represented as follows: Where SDL represents the environmental decision level, m represents the number of elements in the temperature and humidity set, j ranges from 1 to m, and P(FI|(T,H) j ) represents the conditional probability of food intake corresponding to the jth element in the temperature and humidity set; P(ET|(T,H) j ) represents the conditional probability of eating time corresponding to the jth element in the temperature and humidity set, P(EF|(T,H) j ) represents the conditional probability of the number of times the jth element in the temperature and humidity set corresponds to the number of times the food is eaten, μ FI represents the average value of the conditional probability of food intake, σ FI represents the standard deviation of the conditional probability of food intake; μ ET represents the mean value of the conditional probability of eating time, σ ET represents the standard deviation of the conditional probability of eating time, μ EF represents the average value of the conditional probability of eating times, σ EF represents the standard deviation of the conditional probability of eating times; w FI represents the weight of food intake, w ET represents the weight of eating time, w EF Indicates the weight of the number of times you eat.

9. A method for predicting and analyzing the health of dairy cows according to claim 1, characterized in that: Milk analysis grades are shown below: Among them, MG represents the milk production analysis grade, FMR represents the ratio of feed intake to milk production per kilogram of body weight, μ FMR represents the average value of the ratio of food intake to milk production per kilogram of body weight, σ FMR It expresses the standard deviation of the ratio of feed intake to milk production per kilogram of body weight.

10. A method for predicting and analyzing the health of dairy cows according to claim 1, characterized in that: The final health rating is expressed as follows: Among them, FHL represents the final health level, EDL represents the feeding decision level, SDL represents the environmental decision level, MG represents the milk production analysis level, and w EDL represents the weight of the eating decision level, w SDL represents the weight of the environmental decision level, w MG represents the weight of the milk production analysis grade, and e represents the exponential constant.

Images