Animal management systems, animal management methods, and animal husbandry methods

Abstract

This invention provides an animal management system, animal management methods, and animal husbandry methods that make it less likely to overlook diseases. [Solution] The animal management system 10 includes a sensor 14 attached to the animal 12, a communication device 16 for communicating the data obtained by the sensor 14, and a computer 18 for determining the state of the animal 12 from the obtained data. The computer 18 functions according to a program 28 as a preprocessing means 30 for preprocessing the data 42 obtained by the sensor 14, a calculation means 32 for calculating the duration of the non-stationary state (movement state) of the animal 12, a frequency acquisition means 34 for obtaining the frequency of the duration of the non-stationary state, a feature acquisition means 36 for obtaining a feature f from the frequency, and a determination means 38 for determining the state of the animal 12 from the feature f.

Description

Technical Field

[0001] The present invention relates to an animal management system for managing the state of animals, an animal management method, and an animal breeding method for breeding animals by the animal management method.

Background Art

[0002] Conventionally, a system has been developed for attaching sensors to livestock to determine the physical condition of the livestock. For example, the system of Patent Document 1 below detects and integrates the contact of the moving body of the activity sensor attached to a cow with a terminal. The physical condition is determined based on the integrated value per hour. In particular, by using the integrated value in the time zone including the feeding time, it is determined that the feeding behavior has decreased, and the disease of the cow is determined.

[0003] However, Patent Document 1 has a determination method centered on the feeding behavior. It can only determine diseases that affect the feeding behavior, and there is a risk of overlooking other diseases. Therefore, a method that is less likely to overlook diseases is required.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] An object of the present invention is to provide an animal management system, an animal management method, and an animal breeding method that are less likely to overlook diseases.

Means for Solving the Problems

[0006] The animal management system of the present invention comprises a sensor attached to an animal, means for calculating the duration of the animal's stationary state, the duration of its non-stationary state, or both, using data obtained from the sensor, means for acquiring the frequency of the duration of the stationary state, the duration of the non-stationary state, or both, means for acquiring characteristic quantities of the frequency, and means for determining the animal's state from the characteristic quantities.

[0007] The animal management method of the present invention comprises the steps of: detecting the movement of an animal using a sensor attached to the animal; calculating the duration of the animal's stationary state, the duration of its non-stationary state, or both, using data obtained from the sensor; determining the continuous duration of the stationary state, the non-stationary state, or both; obtaining the frequency of the duration of the stationary state, the duration of the non-stationary state, or both; and determining the animal's state from the frequency. The animal is raised using the animal management method of the present invention. [Effects of the Invention]

[0008] According to the present invention, the state of an animal is determined from the frequency of non-stationary states, stationary states, or both. This is not limited to the foraging behavior described in the prior art, and makes it less likely to miss diseases. [Brief explanation of the drawing]

[0009] [Figure 1] This figure shows the configuration of the animal management system of the present invention. [Figure 2] This diagram shows the structure of a program that runs on a computer. [Figure 3] This is an example of a graph showing the relationship between acceleration and time. [Figure 4] This is an example of a histogram of a healthy animal. [Figure 5] This is an example of a histogram of animals in poor health. [Figure 6] This is an example of a graph showing changes in feature quantities. [Figure 7] This is a flowchart of animal management methods. [Modes for carrying out the invention]

[0010] The animal management system, animal management method, and animal rearing method of the present invention will be explained with reference to the drawings.

[0011] The animal management system 10 of the present invention, shown in Figure 1, comprises a sensor 14 attached to an animal 12, a communication device 16 that communicates the data obtained by the sensor 14, and a computer 18 that determines the state of the animal 12 from the obtained data.

[0012] The animals 12 under management are animals that reside within a certain area. For example, cattle kept in a barn or pasture within a certain area are examples of animals 12.

[0013] Sensor 14 is an accelerometer. This accelerometer includes a capacitive 3-axis accelerometer. Because the animal 12 moves in irregular directions, the 3-axis accelerometer is used to detect the movement of the animal 12. In this description, the data obtained from sensor 14 is the acceleration obtained by combining the accelerations in the three orthogonal axes. The accelerometer is attached to the animal 12 to detect the movement of the animal 12. If the animal 12 is a cow, sensor 14 is attached to a collar 20 attached to the cow's neck to detect the cow's movement. Motion detection by sensor 14 may be performed continuously or at regular time intervals. It is also possible to use a 1-axis or 2-axis accelerometer. In that case, multiple accelerometers are attached to a single animal 12 so that the acceleration in each of the three orthogonal axes can be detected. Multiple accelerometers are used to detect the movement of the animal 12.

[0014] The communication device 16 includes a device that wirelessly transmits and receives data obtained from the sensor 14 and transmits it to the computer 18 via the network 22. The transmitter 24 is placed on the animal 12, and the repeater 26 is placed in or around the area where the animal 12 is located. For example, the transmitter 24 and the repeater 26 may include wireless communication devices such as Bluetooth® or WiFi and a router. The transmitter 24 is attached to the animal 12 together with the sensor 14. The sensor 14 and transmitter 24 attached to the animal 12 are designed to be powered by a power source such as batteries. The data obtained from the sensor 14 is wirelessly transmitted from the transmitter 24 to the repeater 26. The repeater 26 transmits the data to the computer 18, which determines the state of the animal 12, via the network 22. If there are multiple animals 12, a sensor 14 and a transmitter 24 are attached to each animal 12, and the data from each sensor 14 is transmitted to the computer 18 via the repeater 26 and the network 22. The data may include an identification code to distinguish the sensors 14.

[0015] A computer 18 that determines the condition of the animal 12 is connected to a network 22. The network 22 may be a communication network such as the internet, or it may be a closed LAN within a certain area.

[0016] The data detected by sensor 14 is transmitted to computer 18 via network 22 in chronological order. Computer 18 receives the values ​​from sensor 14 in chronological order via network 22 and processes them into chronological data. Alternatively, the values ​​detected by sensor 14 may be processed into chronological data and then transmitted to computer 18 as a single data unit. The values ​​detected by sensor 14 may be transmitted randomly, and computer 18 may then arrange them in chronological order to create chronological data. The program described later may include a function to arrange the data from sensor 14 in chronological order.

[0017] As shown in FIG. 2, the computer 18 functions as preprocessing means 30 for preprocessing the data D1 obtained by the sensor 14 by the program 28, calculation means 32 for calculating the duration width of the non-stationary state (motion state, behavior state) of the animal 12, frequency acquisition means 34 for acquiring the frequency of the duration width of the non-stationary state, feature quantity acquisition means 36 for acquiring the feature quantity f from the frequency, and determination means 38 for determining the state of the animal 12 from the feature quantity f.

[0018] The program 28 is executed by a processor or the like of the computer 18. Further, the data D1 obtained by the sensor 14 may be stored in storage means such as a solid-state drive or a hard disk drive. The data D1 may be stored for each animal 12 including an identification code for each animal 12. The data D1 may be stored so as to be available as a database. The stored data D1 will be read and written as appropriate.

[0019] The preprocessing means 30 performs a certain process before using the data D1. For example, reference data D2 is acquired in a state where the sensor 14 is stationary in advance, and the reference data D2 is stored in the storage means for each sensor 14. If the sensor 14 is an acceleration sensor, the gravitational acceleration of the earth is acquired in a state where the sensor 14 is stationary. The preprocessing means 30 subtracts the reference data D2 from the acquired data D1 and calculates its magnitude. The movement of the animal 12 is extracted from the data D1. When individual differences occur for each sensor 14, false determination due to the individual differences of the sensor 14 can be prevented. The time for acquiring the reference data D2 may be from several seconds to several tens of seconds, but is not limited to this time. The state where the sensor 14 is stationary is not limited to placing the sensor 14 on a stationary object such as a desk. The reference data D2 may be acquired when the animal 12 is stationary.

[0020] The calculation means 32 calculates the duration width of the non - stationary state of the animal 12 from the pre - processed data. The calculation means 32 may calculate using data for a certain period of time, for example, data from several hours to one day. The duration width is a continuous period of time. When a single continuous non - stationary state occurs without passing through a stationary state, the time of that single non - stationary state becomes the duration width. If there is a stationary state in the middle of the non - stationary state, the non - stationary state will be divided before and after the stationary state, and the duration width will also be divided before and after the stationary time.

[0021] A specific example of the duration width will be described. When the pre - processed data (acceleration) is arranged in time series, it becomes, for example, like the graph shown in FIG. 3. The vertical axis of the graph is the pre - processed data detected by the sensor 14, and the horizontal axis is time. The pre - processed data represents the movement (acceleration) of the animal 12. When the pre - processed data is large, it can be seen that the animal 12 is in a non - stationary state, and when it is small, it can be seen that the animal is stationary or almost stationary. Note that the pre - processed data may be a value converted into an arbitrary index indicating the movement of the animal 12.

[0022] When the pre - processed data is greater than or equal to the first threshold value σ1, it is in a non - stationary state, and when it is less than the first threshold value σ1, it is in a stationary state. The first threshold value σ1 is obtained in advance from the stationary state and non - stationary state of the animal 12 and stored in the storage means. The calculation means 32 calculates the duration width of the portion that is greater than or equal to the first threshold value σ1 in the graph of FIG. 3. Since there are three non - stationary states in FIG. 3, the respective duration widths T1, T2, and T3 are calculated.

[0023] The frequency acquisition means 34 summarizes the duration widths into a frequency distribution. The frequency acquisition means 34 may summarize all the duration widths into a histogram. For example, a histogram is created by setting the width of the bins to be constant (such as 0.5 seconds) and aggregating the frequencies of the duration widths in each bin. The histogram will show the frequencies of the duration widths.

[0024] The feature acquisition means 36 acquires feature quantities f from the histogram. For example, the area of ​​the histogram is calculated as the feature quantity f. Figure 4 is an example of a histogram of a healthy cow in a non-resting state, and Figure 5 is an example of a histogram of a cow in a non-resting state. If the feature quantity f is the area of ​​the histogram, then the feature quantity f of a healthy animal 12 will be large, and the feature quantity f of a sick animal 12 will be small.

[0025] The determination means 38 determines the state of the animal 12 from the feature quantity f. If the feature quantity f is equal to or greater than the second threshold σ2, the animal is determined to be healthy; if it is less than the second threshold σ2, the animal is determined to be unwell. The second threshold σ2 is obtained in advance from the states of good and poor health of the animal 12 and stored in the memory means. For example, suppose that the feature quantity f is obtained for each day, as shown in Figure 6. There is a day (d3) when the feature quantity f is less than the second threshold σ2, and there is a high possibility that the animal 12 has developed some kind of illness and is unable to move. The veterinarian can discover the illness of the animal 12 by actually examining the animal 12 that has been determined to be unwell. Figure 6 uses the feature quantity f for one day, but the feature quantity f is not limited to being obtained from a daily histogram. The feature quantity f can be appropriately selected and set from a few hours to several days. The second threshold σ2 may be the same value for all animals 12, or it may be a different value for each animal 12, as will be described later.

[0026] The system may be configured to issue a warning if the feature f is less than the second threshold σ². The warning may be displayed on a monitor connected to the computer 18, or a predetermined sound may be emitted from a speaker. The warning may also be sent via the network 22 to a computer 48 owned by the caretaker, and the computer 48 may display the warning.

[0027] The judgment result 40 may be stored in a storage device. The caretaker's computer 48 may access the data D1, reference data D2, and judgment result 40 stored in the storage device via the network 22 to view them. In addition, when processing data D1 with program 28, the processing result may be stored in the storage device after each processing step.

[0028] The data D1 and judgment results 40 stored in the storage means may be accessed and displayed on a computer 42 owned by the caretaker. The caretaker can remotely check the condition of the animal 12.

[0029] Next, the animal management method will be explained using the flowchart in Figure 7. (1) Data D1 is obtained from the sensor 14 attached to the animal 12 (S1). Data D1 may be obtained at regular intervals or continuously. If there are multiple animals 12, a sensor D1 is attached to each animal 12, and data D1 is obtained from the sensor 14 of each animal 12. The data D1 obtained from the sensor 12 is transmitted to the computer 18 via the network 22. The data D1 is stored in the storage means of the computer 18.

[0030] (2) The preprocessing means 30 preprocesses the data D1 obtained from the sensor 14 (S2). Preprocessing may be performed each time data D1 is stored in the computer 18, or a certain amount of data D1 may be preprocessed all at once. Data is acquired in advance for each sensor 14 while it is stationary, and this data is used as reference data D2. If the sensor 14 is an acceleration sensor, the reference data D2 will be the acceleration due to gravity. The preprocessing means 30 subtracts the reference data D2 from the data D1 obtained as preprocessing and calculates its magnitude. By subtracting the reference data D2 from the data D1 obtained from the sensor 14, the influence of individual differences in the sensor 14 is removed.

[0031] (3) The calculation means 32 calculates the duration of the non-stationary state from the pre-processed data (S3). The duration of the pre-processed data is determined to be greater than or equal to the first threshold σ1. The determined duration of the state is the duration of time during which the animal 12 is in motion.

[0032] (4) The frequency acquisition means 34 acquires the frequency of non-stationary states from the calculated duration (S4). The frequency can be obtained by creating a histogram. The frequency of the number of non-stationary states at regular intervals can be determined.

[0033] (5) The feature acquisition means 36 acquires feature f from the frequency (S5). Feature f is the area of ​​the histogram, etc.

[0034] (6) The determination means 38 determines the state of the animal 12 from the feature quantity f (S6). If the feature quantity f is greater than or equal to the second threshold σ2, there is no problem with the state of the animal 12. If the feature quantity f is less than the second threshold σ2, the animal 12 may have low activity levels and be in poor health. The computer 18 displays a warning on its monitor or sends a warning to the owner's computer 42, which then displays the warning (S7).

[0035] As described above, the present invention makes it possible to easily determine the state of animal 12 simply by determining the frequency of non-stationary states. Since the determination is not biased towards foraging behavior as in the conventional method, it is less likely to miss diseases in animal 12 than before.

[0036] [Embodiment 2] The above embodiment uses the duration of a non-stationary state, but the duration of a stationary state may also be used. For example, in the acceleration shown in Figure 3, the duration of a state less than the first threshold σ1 is calculated and a histogram is generated. The state of animal 12 may be determined from the area of ​​that histogram. If animal 12 is unwell, the area of ​​the histogram will be large, and if animal 12 is in good health, the area of ​​the histogram will be small. A feature f is obtained from the histogram, and if the area is larger than the second threshold σ2, it may be determined that animal 12 is unwell.

[0037] Histograms may be created for both the non-resting and resting states. The area of ​​each histogram may be calculated, and the state of animal 12 may be determined from that area. If the area of ​​each histogram exceeds a threshold, it may be determined that the animal is in poor health.

[0038] [Embodiment 3] The method for determining the state of animal 12 is not limited to using the area of ​​the histogram. A curve may be fitted to match the height of the histogram, and the area of ​​that curve may be used. Alternatively, the shape and slope of the curve may be used. The size of a predetermined bin, the ratio of the sizes of multiple bins, the ratio of a predetermined bin to the whole, the area of ​​bins above a threshold, etc., may also be used.

[0039] [Embodiment 4] The data D1 obtained from sensor 14 may be used as is or processed before use. For example, a moving average of the data D1 obtained from sensor 14 may be calculated and that moving average may be used as the data D1 obtained from sensor 14. If sensor 14 acquires data D1 at regular intervals, the moving average may be calculated by taking the average of the data D1 acquired at a certain time and the data D1 before and after it. When the state of animal 12 changes frequently, the changes can be mitigated by using a moving average.

[0040] [Embodiment 5] The second threshold σ2 may be different for each animal 12. A means may be provided to determine the second threshold σ2 for determining the state of the animal 12. The amount of activity often differs depending on the personality, age, sex, etc. of the animal 12. Data D1 may be obtained statistically and its average value may be used as the second threshold σ2, or the second threshold σ2 may be determined by machine learning after accumulating data D1. The accuracy of the determination can be improved by using the optimal second threshold σ2 for each animal 12. In addition, the first threshold σ1 may also be different for each animal 12, similar to the second threshold σ2.

[0041] [Embodiment 6] The data D1 obtained from sensor 14 was subtracted by the reference data D2, but it can also be used without subtraction. When there is no or very little individual variation in sensor 14, using the data D1 without subtracting from the reference data D2 is unlikely to cause errors in the data D1.

[0042] [Embodiment 7] Sensor 14 is not limited to an accelerometer. For example, a camera may be used as a sensor. A certain area may be photographed with the camera, and the activity level of the animal 12 may be obtained by image processing. The above data will be activity levels obtained from image processing instead of acceleration. The above calculation of duration, acquisition of frequency, acquisition of features, and determination are performed using the activity levels.

[0043] Furthermore, the present invention can be implemented in various forms with improvements, modifications, and changes based on the knowledge of those skilled in the art, without departing from its spirit. [Explanation of Symbols]

[0044] 10: Animal Management Systems 12:Animal 14: Sensor 16: Communication equipment 18: Computer 20: Collar 22: Network 24: Transmitter 26: Repeater 28: Program 30: Pre-treatment means 32: Calculation means 34: Frequency acquisition means 36: Feature acquisition methods 38: Judgment means 40: Judgment result 42: Other computers D1: Data D2: Reference data σ1: First threshold σ2: Second threshold

Claims

1. Sensors attached to animals, A means for calculating the duration of the animal's resting state, the duration of its non-resting state, or both, using data obtained from the aforementioned sensor, Means for acquiring the frequency of the duration of the stationary state, the duration of the non-stationary state, or both; Means for obtaining the frequency feature quantities, A means for determining the state of an animal from the aforementioned feature quantities, An animal management system equipped with [unspecified features].

2. The means for obtaining the frequency generates a histogram from the duration of the stationary state, the duration of the non-stationary state, or both. The means for obtaining the frequency features obtains the histogram features, The means for determining the state of the animal uses histogram features to determine the state of the animal. The animal management system according to claim 1.

3. The animal management system according to claim 1 or 2, further comprising a preprocessing means for obtaining reference data by setting the sensor to a stationary state in advance, and subtracting the reference data from the data obtained by the sensor.

4. The animal management system according to claim 1 or 2, wherein the means for performing the calculation is to perform the calculation from data obtained from a sensor after the data has been given a moving average over a certain period of time.

5. The process involves detecting the animal's movement using sensors attached to the animal, A step of using the data obtained from the aforementioned sensor to calculate the duration of the animal's resting state, the duration of its non-resting state, or both. A step of determining the continuous time of the aforementioned stationary state, non-stationary state, or both, A step of obtaining the frequency of the duration width of the static state, the duration width of the non-static state, or both; A step of obtaining the frequency feature quantity, A step of determining the animal's condition from the aforementioned frequency, Animal management methods that include these.

6. The step of obtaining the aforementioned frequency involves generating a histogram from the duration of the stationary state, the duration of the non-stationary state, or both. The step of obtaining the frequency features involves obtaining the histogram features, The means for determining the state of the animal is to determine the state of the animal using histogram features, according to claim 5 of the animal management method.

7. The animal management method according to claim 5 or 6, wherein the sensor is kept stationary in advance to obtain reference data, and the reference data is subtracted from the data obtained by the sensor.

8. The animal management method according to claim 5 or 6, wherein the calculation of the duration of the animal's resting state, the duration of its non-resting state, or both, is performed after the data obtained from the sensor has been converted to a moving average over a certain period of time.

9. An animal rearing method comprising the animal management method described in claim 5 or 6.

Images