JP2019170360A - Identification device, identification method and program - Google Patents

Abstract

To provide a device, an identification method, and a program for supporting management of food collection behavior of farm animals.SOLUTION: The device for identifying food collection behavior comprises: storage means for storing acceleration data measured by an acceleration sensor attached to a farm animal; acquisition means S101 for acquiring from the storage means, a plurality of pieces of acceleration data in a prescribed period; calculation means S102 for calculating a feature amount of each of the plurality of pieces of acceleration data acquired by the acquisition means; peak identification means S104 for, based on the plurality of feature amounts calculated by the calculation means and a preset parameter, identifying at least one of a maximum peak indicating a local maximum feature amount, and a local minimum peak indicating a minimum feature amount; determination means S105 for determining whether or not the number of the peaks identified by the identification means is equal to or more than a prescribed threshold; and behavior identification means S106 for identifying the behavior of the farm animal in the prescribed period, as a food collection behavior.SELECTED DRAWING: Figure 4

Description

  The present invention relates to a specifying device, a specifying method, and a program.

  For livestock farmers raising dairy cattle and fattening cattle, it is one of the important tasks to perform individual management including cattle health management. As part of such individual management, it is also important to manage cattle foraging behavior. This is because, for example, when the amount of feed the cattle feeds is less than usual, it is necessary to take measures such as changing the type of feed or changing the layout of the feed area. .

Abe Ryo, “Agricultural Fundamentals Seminar, Fundamental of Animal Breeding”, New Edition, Agricultural Fisheries Cultural Association, April 2008, p.109, p.122-124.

  However, for example, in a large ranch where the number of animals to be raised ranges from several hundred to several thousand, it is difficult to confirm the feeding behavior of each cow individually by visual observation or the like.

  One embodiment of the present invention has been made in view of the above points, and an object thereof is to support management of livestock foraging behavior.

  In order to solve the above-described problem, an embodiment of the present invention is a specifying device that specifies feeding behavior of livestock, and includes a storage unit that stores acceleration data measured by an acceleration sensor attached to the livestock, and a predetermined unit. An acquisition means for acquiring a plurality of acceleration data from the storage means, a calculation means for calculating feature quantities of each of the plurality of acceleration data acquired by the acquisition means, and a calculation means calculated by the calculation means Based on a plurality of feature amounts and a preset parameter, at least one of a maximum peak indicating a locally maximum feature amount and a minimum peak indicating a locally minimum feature amount A peak specifying unit that specifies the number of peaks, a determination unit that determines whether or not the number of peaks specified by the peak specifying unit is equal to or greater than a predetermined threshold, and the determination unit If the number of the peaks is determined to the equal to or greater than a predetermined threshold, and having a specifying means action to identify the feeding behavior behavior of the livestock during the predetermined time.

  According to one embodiment of the present invention, management of livestock foraging behavior can be supported.

It is a figure which shows an example of the whole structure of the specific system which concerns on 1st embodiment. It is a figure which shows an example of the measurement data memorize | stored in the measurement data memory | storage part. It is a figure which shows an example of a function structure of the specific process part which concerns on 1st embodiment. It is a flowchart which shows an example of the specific process of foraging action which concerns on 1st embodiment. It is FIG. (1) for demonstrating an example of peak specification. It is a flowchart which shows an example of the peak specific process which concerns on 1st embodiment. It is FIG. (2) for demonstrating an example of peak specification. It is a flowchart which shows an example of the specific process of foraging action which concerns on 2nd embodiment. It is a flowchart which shows an example of the peak specific process which concerns on 2nd embodiment. It is a figure which shows an example of a function structure of the specific process part which concerns on 3rd embodiment. It is a flowchart which shows an example of the specific process of foraging action which concerns on 3rd embodiment. It is a flowchart which shows an example of the tag direction specific process which concerns on 3rd embodiment.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following embodiments, a description will be given of the specifying system 1 that specifies cattle feeding behavior as an example of livestock. Livestock is not limited to cattle.

  Here, the foraging behavior is a state where the cow is performing an operation of eating food (for example, pasture or feed).

[First embodiment]
<Overall configuration>
First, the overall configuration of the specific system 1 according to the present embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of the overall configuration of a specific system 1 according to the first embodiment.

  As shown in FIG. 1, the identification system 1 according to the present embodiment includes a identification device 10 that identifies cattle foraging behavior and one or more tags 20 attached to the cow. In addition, it is preferable that the tag 20 is fixedly attached to the neck portion of the cow.

  The tag 20 is a device attached to the cow. One tag 20 is attached to one cow. The tag 20 includes an acceleration sensor for measuring the acceleration of the cow wearing the tag 20 (X-axis, Y-axis, and Z-axis accelerations).

  The tag 20 transmits measurement data including the acceleration sensor value measured by the acceleration sensor to the specific device 10 every predetermined time (for example, every 2 seconds). The measurement data transmitted to the specific device 10 is accumulated (stored) in a measurement data storage unit 200 described later.

  The identification device 10 is a computer that identifies the foraging behavior of cattle. The identification device 10 includes a identification processing unit 100 and a measurement data storage unit 200.

  The identification processing unit 100 identifies the cattle feeding behavior based on the measurement data stored in the measurement data storage unit 200. The specific processing unit 100 is realized by a process in which one or more programs installed in the specific device 10 are executed by a CPU (Central Processing Unit) or the like.

  The measurement data storage unit 200 stores the measurement data received from the tag 20. The measurement data storage unit 200 can be realized using an auxiliary storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). Note that the measurement data storage unit 200 may be realized using a storage device or the like connected to the specific device 10 via a network.

  The measurement data storage unit 200 stores (stores) a plurality of measurement data at predetermined time intervals (for example, every 2 seconds).

  The configuration of the specific system 1 illustrated in FIG. 1 is an example, and other configurations may be used. For example, the specifying device 10 may be configured by a plurality of computers. In addition, for example, a part of the functions of the specific processing unit 100 may be included in an apparatus (for example, a cloud server) connected to the specific apparatus 10 via a network.

  Moreover, the transmission method of the measurement data from the tag 20 to the specific apparatus 10 is not limited. For example, the tag 20 may transmit measurement data to the specific device 10 via a network such as the Internet, may transmit measurement data to the specific device 10 within a local network, or may be short-range wireless communication For example, the measurement data may be transmitted to the specific device 10.

<Measurement data stored in measurement data storage unit 200>
Here, the measurement data stored in the measurement data storage unit 200 according to the present embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of measurement data stored in the measurement data storage unit 200. In addition, the specific apparatus 10 should just memorize | store (accumulate) the said measurement data in the measurement data memory | storage part 200 by the specific process part 100, when measurement data are received from the tag 20. FIG.

  As shown in FIG. 2, the measurement data storage unit 200 stores one or more measurement data for each tag ID that is an example of identification information for identifying the tag 20. Since one tag 20 is attached to one cow, the tag ID may be identification information for identifying a cow (for example, individual identification information for a cow).

  Each measurement data includes a date and time and an acceleration sensor value. The date and time is, for example, the date and time when the tag 20 transmits measurement data. The date and time may be the date and time when the specific device 10 receives the measurement data.

The acceleration sensor value is an acceleration value measured by the acceleration sensor included in the tag 20. The acceleration sensor value includes an X component indicating an acceleration component in the X-axis direction, a Y component indicating an acceleration component in the Y-axis direction, and a Z component indicating an acceleration component in the Z-axis direction. For example, the measurement data of the date “t ₀ ” includes an X component “a _x0 ”, a Y component “a _y0 ”, and a Z component “a _z0 ” of the acceleration sensor value. Similarly, for example, the measurement data of the date and time “t ₁ ” includes an X component “a _x1 ”, a Y component “a _y1 ”, and a Z component “a _z1 ” of the acceleration sensor value.

  As described above, the measurement data stored in the measurement data storage unit 200 according to the present embodiment stores (stores) measurement data including the date and time and the acceleration sensor value for each tag ID.

<Functional Configuration of Specific Processing Unit 100>
Next, the functional configuration of the specific processing unit 100 according to the present embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of a functional configuration of the specific processing unit 100 according to the first embodiment.

  As illustrated in FIG. 3, the identification processing unit 100 according to the present embodiment includes an acquisition unit 101, a preprocessing unit 102, a norm calculation unit 103, a peak identification unit 104, and a foraging identification unit 105.

  The acquisition unit 101 acquires measurement data from the measurement data storage unit 200. At this time, the acquisition unit 101 acquires measurement data of a predetermined time width (for example, 10 minutes) from the measurement data storage unit 200 for each tag ID, for example.

  The preprocessing unit 102 performs preprocessing on the measurement data acquired by the acquisition unit 101. As the preprocessing, for example, missing interpolation (resampling) processing of measurement data is included. The preprocessing may include, for example, noise removal processing.

  The norm calculation unit 103 calculates the L2 norm of the acceleration sensor value included in the measurement data after the preprocessing by the preprocessing unit 102. Thereby, time-series data of L2 norm is obtained. The norm calculation unit 103 is not limited to the L2 norm, and may generally calculate the Lp norm (p = 1, 2,..., ∞). These Lp norms (p = 1, 2,..., ∞) are examples of feature amounts.

  The peak identifying unit 104 identifies the peak of the L2 norm time-series data obtained by the norm calculating unit 103 using a preset parameter. The peak is the L2 norm that is locally maximum (or the L2 norm that is locally minimum).

  The foraging identification unit 105 determines whether or not the number of peaks identified by the peak identification unit 104 (the number of peaks) is equal to or greater than a predetermined threshold Th, whereby the time of the measurement data acquired by the acquisition unit 101 Whether the behavior of the cow during the width (for example, 10 minutes) is a foraging behavior is specified. At this time, when it is determined that the peak number is equal to or greater than the threshold Th, the behavior of the cow during the time width is specified as the foraging behavior. On the other hand, when it is determined that the peak number is not equal to or greater than the threshold value Th, it is determined that the behavior of the cow during the time span is not foraging behavior.

  Information indicating the behavior of the cow identified by the foraging identification unit 105 (whether it is foraging or not foraging) may be displayed on, for example, the display of the identifying device 10 or the identifying device 10 May be stored in an auxiliary storage device or the like, or may be transmitted to another device (for example, a PC, a smartphone, a tablet terminal, or the like) connected to the specific device 10. Further, when stored in the auxiliary storage device or the like of the specific device 10, for example, it is stored in the auxiliary storage device or the like using another device (for example, a PC, a smartphone, a tablet terminal, or the like) connected to the specific device 10. It may be possible to browse the information (information indicating the behavior of the cow).

<Identification of foraging behavior>
In the following, the process of specifying the cattle foraging behavior will be described with reference to FIG. FIG. 4 is a flowchart illustrating an example of the foraging action specifying process according to the first embodiment. Note that the process shown in FIG. 4 is repeatedly executed, for example, every 10 minutes. However, 10 minutes is an example, and may be repeatedly executed every arbitrary time.

  First, the acquisition unit 101 acquires measurement data of a predetermined time width (for example, 10 minutes) from the measurement data storage unit 200 for each tag ID (step S101). Thus, the acquisition unit 101 acquires measurement data of a predetermined time width (for example, 10 minutes) from the measurement data storage unit 200 for each cow (tag ID). Such a predetermined time width is not limited to 10 minutes, and for example, the user of the specific device 10 can set an arbitrary time width.

  Hereinafter, the description will be continued on the assumption that the measurement data 0, measurement data 1,..., Measurement data (N−1) of the tag ID “Tag 1” for 10 minutes are acquired by the acquisition unit 101.

  Next, the preprocessing unit 102 performs preprocessing on the measurement data acquired by the acquisition unit 101 (step S102). That is, the preprocessing unit 102 performs a missing interpolation (resampling) process on the measurement data. The defect interpolation process is a process of interpolating defects when measurement data cannot be acquired in order to improve the accuracy of collected measurement data. Thereby, the measurement data after pre-processing is renumbered, and measurement data 0, measurement data 1,..., Measurement data (M-1) are obtained. When 10-minute measurement data 0, measurement data 1,..., Measurement data (N-1) is subjected to a missing interpolation (resampling) process at intervals of 2 seconds, M = 300, and measurement data 0 , Measurement data 1,..., Measurement data 299 are obtained.

  Note that the preprocessing unit 102 may perform the defect interpolation process after performing the noise removal process, for example. By performing the noise removal processing, for example, measurement data indicating the behavior of a cow that may become noise (for example, an operation that instantaneously shakes the body or an operation that momentarily shakes the body) can be removed.

  Next, the norm calculation unit 103 calculates the L2 norm of the acceleration sensor value included in the measurement data after the preprocessing by the preprocessing unit 102 (step S103). That is, the norm calculation unit 103 calculates the L2 norm of the acceleration sensor value included in each measurement data i (i = 0,..., M−1).

  Hereafter, the L2 norm of the acceleration sensor value included in the measurement data i is expressed as y [i], and the date and time included in the measurement data i is expressed as x [i]. As a result, L2 norm time-series data {(x [i], y [i])} (i = 0,..., M−1) is obtained.

  Next, the peak identification unit 104 uses the parameters set in advance, the L2 norm time series data {(x [i], y [i])} obtained by the norm calculation unit 103 (i = 0, .., M-1) are identified (step S104). As described above, the peak is the L2 norm that is locally the maximum value or the L2 norm that is locally the minimum value. In the following, the L2 norm having the local maximum value is referred to as “maximum peak”, and the L2 norm having the local minimum value is referred to as “minimal peak”. Details of the processing for specifying the peaks (maximum peak and minimum peak) will be described later.

  Next, the foraging identification unit 105 determines whether or not the number of peaks (peak number) identified by the peak identification unit 104 is equal to or greater than a predetermined threshold Th (step S105). Here, the foraging identification unit 105 may determine whether the peak number of the maximum peak is equal to or greater than the threshold value Th, and determines whether the peak number of the minimum peak is equal to or greater than the threshold value Th. Also good. Or foraging specific part 105 may judge whether the sum total of the peak number of a maximum peak and the peak number of a minimum peak is more than threshold Th. In the present embodiment, as an example, it is assumed that it is determined whether the number of maximum peaks is equal to or greater than a threshold value Th.

  If it is determined in step S105 that the number of maximum peaks is equal to or greater than the threshold Th, the foraging identification unit 105 identifies the behavior of the cow during the time span (ie, 10 minutes) as the foraging behavior ( Step S106).

  On the other hand, in step S105, when it is not determined that the peak number of the maximum peak is equal to or greater than the threshold Th (that is, when it is determined that the peak number of the maximum peak is less than the threshold Th), the foraging identification unit 105 The behavior of the cow during the time span (that is, 10 minutes) is specified as not a foraging behavior (step S107).

  As described above, the identification system 1 according to the present embodiment determines whether or not the peak number of the L2 norm of the acceleration sensor value in the predetermined time width is equal to or greater than the predetermined threshold Th. It is specified whether the behavior of the cow in between is foraging behavior or not foraging behavior. Thereby, the user (for example, livestock farmers etc.) of specific device 10 concerning this embodiment can manage whether a cow performed foraging behavior. For this reason, for example, when there is a cow that does not perform foraging behavior, the cause and the like can be investigated, and various measures such as changing the feed and changing the layout of the feeding area can be taken. It becomes like this.

<Peak identification>
Next, the L2 norm time-series data {(x [i], y [i])} (i = 0,..., M−1) specifying the peak (the processing in step S104 above) Details will be described. In the following Example 1, a case where a peak is specified without considering the prefetch length L will be described, and in Example 2, a case where a peak is specified in consideration of the prefetch length L will be described.

Example 1
First, a case where a peak is specified without considering the prefetch length L will be described. In this case, δ, which is a parameter for determining the difference between successive L2 norms y [i], is used.

  At this time, the peak specifying unit 104 specifies y [i] satisfying the following expression 1 as a maximum peak for i = 1,..., M−2.

y [i] -y [i-1]> δ and y [i] -y [i + 1]> δ (Formula 1)
Moreover, the peak specific | specification part 104 specifies y [i] which satisfy | fills the following formula | equation 2 with respect to i = 1, ..., M-2 as a minimum peak.

y [i-1] -y [i]> δ and y [i + 1] -y [i]> δ (Expression 2)
That is, the value obtained by subtracting the previous L2 norm y [i−1] from y [i] is larger than δ, and the value obtained by subtracting the next L2 norm y [i + 1] from y [i] is δ. If greater than y, i [i] is identified as the maximum peak. On the other hand, the value obtained by subtracting the previous L2 norm y [i−1] from y [i] is smaller than −δ, and the value obtained by subtracting the next L2 norm y [i + 1] from y [i] If y is smaller than −δ, y [i] is identified as a minimum peak.

  Here, as an example, assuming that M = 20 and δ = 55, the peak of a certain L2 norm time-series data {(x [i], y [i])} (i = 0,..., 19) The case specified by Example 1 is shown in FIG.

  As shown in FIG. 5, since y [2] −y [1]> 55 and y [2] −y [3]> 55, y [2] is identified as a maximum peak. Moreover, since y [2] −y [3]> 55 and y [4] −y [3]> 55, y [3] is specified as a minimum peak. Similarly, y [8], y [12], and y [18] are each identified as a maximum peak, and y [13] is identified as a minimum peak. Therefore, in this case, the peak number of the maximum peak is 4, and the peak number of the minimum peak is 2.

  In the above description, the common parameter δ is used in Equation 1 and Equation 2, but for example, a parameter δ having a different value may be used in Equation 1 and Equation 2.

(Example 2)
Next, a case where a peak is specified in consideration of the prefetch length L will be described. By considering the prefetch length L, a peak with a small fluctuation width within the time width x [i + L] −x [i] taking the prefetch length L into account (ie, the difference between y [i] and y [i−1]). Or a peak having a small difference between y [i] and y [i + 1] can be prevented from being treated as a maximum peak or a minimum peak. Thereby, foraging behavior can be specified with higher accuracy than in the first embodiment.

  In the case of the second embodiment, δ that is a parameter for determining a difference between consecutive L2 norms y [i] and L that is a parameter indicating the look-ahead length are used. Hereinafter, the peak identification process in the second embodiment will be described with reference to FIG. FIG. 6 is a flowchart illustrating an example of the peak specifying process according to the first embodiment.

  First, the peak identifying unit 104 initializes various variables (step S201). That is, the peak identification unit 104 initializes a variable i that is an index of each time series data of the L2 norm, a variable max that indicates a temporary maximum peak, and a variable min that indicates a temporary minimum peak to 0, respectively. To do. In addition, after each variable is initialized, the process after the following step S202 is repeatedly performed with respect to i = 0, ..., ML.

  The peak identification unit 104 determines whether y [i]> max is satisfied (step S202).

  When it is determined in step S202 that y [i]> max, the peak specifying unit 104 substitutes y [i] for max and substitutes x [i] for max_pos (step S203). Note that max_pos is a variable to which x [i], which is the date and time of y [i] (provisional maximum peak) that is substituted for max, is substituted.

  On the other hand, when it is not determined in step S202 that y [i]> max, or following step S203, the peak identifying unit 104 determines whether y [i] <min (step S204). .

  When it is determined in step S204 that y [i] <min, the peak identifying unit 104 substitutes y [i] for min and substitutes x [i] for min_pos (step S205). Note that min_pos is a variable in which x [i], which is the date and time of y [i] (provisional minimal peak) that is substituted for min, is substituted.

  On the other hand, if it is not determined in step S204 that y [i] <min, or following step S205, the peak specifying unit 104 determines that y [i] <max−δ and y [i: i + L]. It is determined whether or not the maximum value <max (step S206). Here, y [i: i + L] represents y [i], ..., y [i + L-1]. Therefore, the maximum value of y [i: i + L] is the maximum y among y [i],..., Y [i + L−1].

  When it is determined in step S206 that y [i] <max−δ and the maximum value of y [i: i + L] <max, the peak specifying unit 104 adds the set of max and max_pos to the maximum peak list. (Step S207). The max added to the maximum peak list (that is, y [i] assigned to this max) is the maximum peak. Further, max_pos corresponding to this max (that is, x [i] assigned to this max_pos) is the date and time of the maximum peak.

  Next, the peak specifying unit 104 substitutes ∞ for max and min (step S208).

  When it is not determined in step S206 that y [i] <max−δ and the maximum value of y [i: i + L] <max, the peak identifying unit 104 determines that y [i]> min + δ and y It is determined whether or not the minimum value of [i: i + L]> min (step S209). Here, the minimum value of y [i: i + L] is the minimum y among y [i],..., Y [i + L−1].

  When it is determined in step S209 that y [i]> min + δ and the minimum value of y [i: i + L]> min, the peak identifying unit 104 adds a set of min and min_pos to the minimal peak list ( Step S210). The min added to the minimum peak list (that is, y [i] assigned to this min) is the minimum peak. In addition, min_pos corresponding to this min (that is, x [i] assigned to this min_pos) is the date and time of the minimum peak.

  Next, the peak specifying unit 104 substitutes −∞ for max and min (step S211).

  When it is not determined in step S209 that y [i]> min + δ and the minimum value of y [i: i + L]> min, or following step S209 or step S211, the peak identifying unit 104 determines that i + L ≧ M It is determined whether or not (step S212). M is the number of data items included in the L2 norm time-series data {(x [i], y [i])}.

  When it is not determined in step S212 that i + L ≧ M, the peak identifying unit 104 substitutes i + 1 for i (step S213). That is, the peak specifying unit 104 increments the value of i. And the peak specific | specification part 104 performs the process after step S202 using i after the increment.

  On the other hand, when it is determined in step S212 that i + L ≧ M (that is, i + L = M), the peak specifying unit 104 ends the peak specifying process. At this time, y [i] included in the maximum peak list is a maximum peak. At this time, y [i] included in the minimum peak list is a minimum peak.

  Here, as an example, assuming that M = 20, look-ahead length L = 5, and δ = 55, time series data of a certain L2 norm {(x [i], y [i])} (i = 0,... , 19) is shown in FIG.

  As shown in FIG. 7, y [2] does not satisfy the condition of step S206 (y [i] <max−δ and the maximum value of y [i: i + L] <max). Not. This is because the maximum value at this time is y [2], while the maximum value of y [2: 7] is y [7], and y [2] <y [7]. Similarly, y [12] does not satisfy the condition in step S206, and therefore does not become a maximum peak.

  On the other hand, as shown in FIG. 7, y [8] is identified as a maximum peak because the condition of step S206 is satisfied. Similarly, y [3] and y [13] are identified as minimal peaks.

  When M = 20 and the prefetch length L = 5, it is y [0] to y [15] that the peak identifying unit 104 determines whether or not the peak is present. For this reason, y [18] is not specified as the maximum peak.

  In FIG. 6, the common parameter δ is used in step S206 and step S209, but for example, a parameter δ having a different value may be used in step S206 and step S209.

<Summary of First Embodiment>
As described above, the identification system 1 according to the present embodiment determines whether or not the peak number of the L2 norm of the acceleration sensor value in the predetermined time width is equal to or greater than the predetermined threshold Th. It is specified whether the behavior of the cow in between is foraging behavior or not foraging behavior. Thereby, as above-mentioned, the user (for example, livestock farmer etc.) of the specific apparatus 10 which concerns on this embodiment can manage whether the cow performed foraging behavior, for example, foraging behavior In the case where there is a cow that has not been performed, the cause and the like can be investigated, and various measures such as changing the feed and changing the layout of the feed area can be taken.

  In the first and second embodiments described above, the peak is specified using the parameter δ. However, the value of the parameter δ may be different between the first and second embodiments. In addition, it is preferable that the threshold value Th used in step S105 in FIG. 4 is set to a different value when using the first embodiment and when using the second embodiment. Furthermore, in Example 1 and Example 2, although both the maximum peak and the minimum peak were specified, it is not restricted to this, You may specify only one of a maximum peak or a minimum peak.

  For example, the peak of L2 norm time series data {(x [i], y [i])} (i = 0,..., 299) every 2 seconds in 10 minutes is set to L = 5 and δ = 55. When specified, the threshold Th is preferably about Th = 3 to 10, for example. In particular, when Th = 5, it was verified by experiments that the foraging behavior of cattle can be specified satisfactorily.

  Various methods can be considered as a method for determining the parameter values (that is, the values of L, δ, and Th) that can favorably identify the foraging behavior of cattle. For example, it can be determined by grid search or the like. it can.

[Second Embodiment]
Next, a second embodiment will be described. In the first embodiment, whether the behavior of the cow is a foraging behavior or not a foraging behavior is determined depending on whether or not the identified peak is equal to or greater than the threshold Th. However, in the first embodiment, for example, when an action such as a cow swinging his head or playing with other cows, the action is erroneously specified as a foraging action. was there. Therefore, in the second embodiment, a description will be given of the identification system 1 that can reduce the case where it is erroneously identified as a foraging behavior.

  In the second embodiment, differences from the first embodiment will be mainly described, and the same components as those in the first embodiment will be given the same reference numerals and description thereof will be omitted. .

<Identification of foraging behavior>
In the following, the process for identifying the foraging behavior of the cow will be described with reference to FIG. FIG. 8 is a flowchart illustrating an example of a foraging action specifying process according to the second embodiment. Note that steps S101 to S103 in FIG. 8 are the same as those in FIG.

  Subsequent to step S103, the peak identifying unit 104 uses the preset parameters, and the L2 norm time-series data {(x [i], y [i])} obtained by the norm calculating unit 103 (i = 0,..., M−1) is identified (step S301). Here, in the second embodiment, each local maximum peak is associated with a first slope, a first difference, and a first interval. Similarly, the second peak, the second difference, and the second interval are associated with each minimum peak.

  The first inclination is an inclination of a straight line connecting the minimum peak immediately before the maximum peak and the maximum peak. The first difference is a difference between the maximum peak and the median value of the L2 norm. However, in the calculation of the first difference, for example, the median value of all maximum peaks and minimum peaks during a predetermined time width (for example, 10 minutes) may be used instead of the median value of the L2 norm. . The first interval is the interval (interval in the x direction) between the maximum peak and the maximum peak immediately before the maximum peak.

  Similarly, the second slope is a slope of a straight line connecting the maximum peak immediately before the minimum peak and the minimum peak. The second difference is a difference between the minimum peak and the median value of the L2 norm. However, in the calculation of the second difference, similarly, the median value of all the maximum peaks and the minimum peaks during a predetermined time width (for example, 10 minutes) is used instead of the median value of the L2 norm. May be. The second interval is an interval (interval in the x direction) between the minimal peak and the minimal peak immediately before the minimal peak.

  A peak (maximum peak and minimum peak) associated with the first slope, the first difference, and the first interval (or the second slope, the second difference, and the second interval) is specified. Details of the processing will be described later. In the details of this process, a calculation method of the first inclination, the first difference, the first interval, the second inclination, the second difference, and the second interval will also be described.

  Next, the foraging identification unit 105 excludes peaks satisfying a predetermined condition from the peaks identified by the peak identification unit 104 (step S302). Here, the predetermined conditions include the following (1-1) to (1-3) and (2-1) to (2-3). That is, the foraging identification unit 105 excludes a maximum peak that satisfies any of the conditions (1-1) to (1-3). Similarly, the foraging identification unit 105 excludes a minimum peak that satisfies any one of the conditions (2-1) to (2-3).

(1-1) The first slope associated with the maximum peak is a predetermined threshold Th _{a, 1} or more. (1-2) The first difference associated with the maximum peak is a predetermined threshold Th _{d, 1} or more (1-3) The first interval associated with the maximum peak is a predetermined threshold value Th _{s, 1} or more (2-1) The absolute value of the second slope associated with the minimum peak is predetermined the absolute value of the threshold value _{Th a, 2} or more (2-2) a second differential associated with the minimum peak is associated with a predetermined threshold value _{Th d, 2} or more (2-3) minimum peak The interval of 2 is a predetermined threshold Th _{s, 2} or more. Here, the threshold Th _{a, 1} is, for example, Th _{a, 1} = 70. Further, as the threshold Th _{s, 1} , for example, Th _{s, 1} = 300 can be cited. Further, as the threshold Th _{s, 1} , for example, Th _{s, 1} = 80 (seconds) can be mentioned. Note that the threshold Th _{a, 2} may be, for example, Th _{a, 2} = Th _{a, 1} or the like. Similarly, as the threshold Th _{d, 2} , for example, Th _{d, 2} = Th _{d, 1} may be used. Similarly, the threshold Th _{s, 2} may be, for example, Th _{s, 2} = Th _{s, 1} .

  In addition, said conditions (1-1), (1-2), (2-1), and (2-2) exclude the peak which shows a sudden movement with low possibility of appearing at the time of foraging behavior. Means. On the other hand, since the neck and head usually move with a substantially constant rhythm when eating, the specified peak interval tends to be substantially constant. On the other hand, for example, in an action such as shaking a head or being licked by another cow, the peak interval is often not constant, and the peak interval tends to be larger than when eating. The above conditions (1-3) and (2-3) mean that such a peak whose interval is larger than that at the time of eating is excluded.

  Next, the foraging identification unit 105 determines whether or not the number of peaks after the exclusion in step S104 is equal to or greater than a predetermined threshold Th (step S303). Here, the foraging identification unit 105 may determine whether or not the number of peaks of the maximum peak after exclusion in step S104 is equal to or greater than the threshold Th, or the minimum peak after exclusion in step S104. It may be determined whether the number of peaks is equal to or greater than a threshold value Th. In the present embodiment, as an example, it is assumed that it is determined whether or not the number of maximum peaks after exclusion is equal to or greater than a threshold Th.

  In Step S303, when it is determined that the number of maximum peaks after exclusion is equal to or greater than the threshold Th, the foraging specific unit 105 determines the behavior of the cow during the time span (ie, 10 minutes) as the foraging behavior. Specify (step S304).

  On the other hand, when it is not determined in step S303 that the number of peaks of the maximum peak after exclusion is equal to or greater than the threshold Th (that is, when the number of peaks of the maximum peak after exclusion is determined to be less than the threshold Th), The foraging identification unit 105 identifies the behavior of the cow during the time span (that is, 10 minutes) as not the foraging behavior (step S305).

  As described above, in the identification system 1 according to the present embodiment, the number of peaks excluding the peak satisfying the predetermined condition among the peaks of the L2 norm of the acceleration sensor value in the predetermined time width is equal to or greater than the predetermined threshold Th. By determining whether or not there is, it is specified whether the behavior of the cow during the time span is foraging behavior or not foraging behavior. As a result, for example, when an action such as a cow swinging his head or playing with another cow is performed, the situation in which the action is erroneously identified as a foraging action can be reduced. .

<Peak identification>
In the following, details of the processing in step S301 (peak identification processing) will be described with reference to FIG. FIG. 9 is a flowchart illustrating an example of the peak specifying process according to the second embodiment. Note that step S201 to step S206, step S208 to step S209, and step S211 to step S213 in FIG. 9 are the same as those in FIG. Hereinafter, the variable indicating the first inclination, the variable indicating the first difference, and the variable indicating the first interval are referred to as max_a, max_d, and max_s, respectively. Similarly, a variable indicating the second gradient, a variable indicating the second difference, and a variable indicating the second interval are set to min_a, min_d, and min_s, respectively.

  When it is determined in step S206 that y [i] <max−δ and the maximum value of y [i: i + L] <max, the peak identifying unit 104 sets the pair (max, max_pos, max_a, max_d, max_s). ) Is added to the maximum peak list (step S401). As in the first embodiment, max (that is, y [i] substituted for max) is a maximum peak. max_pos (that is, x [i] substituted for max_pos) is the date and time of the maximum peak.

  Here, in max_a, y [i] and x [i] and the minimum peak with the newest date and time among the minimum peaks included in the minimum peak list (that is, the latest peak is added to the minimum peak list at the present time). And the first slope calculated by the date and time is substituted. That is, if the minimum peak is y [i ′] and the date and time of the minimum peak is x [i ′], the first slope is (y [i] −y [i ′]) / (x [i ] −x [i ′]).

  In addition, in max_d, the first difference calculated by the y [i] and the median value of the upper and lower limits of all y [i] (i = 0,..., M−1). Is substituted. That is, if the median (intermediate value) of the upper and lower limits of y [i] (i = 0,..., M−1) is μ, the first difference is y [i] −μ. Is calculated by As described above, for example, μ may be a median value of all the maximum peaks y [i] and minimum peaks y [i ′] during a predetermined time width (for example, 10 minutes). In this case, after all maximum peaks y [i] and minimum peaks y [i ′] during a predetermined time width are added to the maximum peak list and the minimum peak list, respectively, μ is calculated and max_d is set. What is necessary is just to calculate.

  Furthermore, the first interval calculated by the x [i] and the date and time of the previous maximum peak is substituted for max_s. That is, if the date and time of the previous maximum peak is x [i ″], the first interval is calculated by x [i] −x [i ″].

  Thereby, the maximum peak, the date and time, the first slope, the first difference, and the first interval are stored in the maximum peak list in association with each other.

  When it is determined in step S209 that y [i]> min + δ and the minimum value of y [i: i + L]> min, the peak identifying unit 104 sets the set (min, min_pos, min_a, min_d, min_s). It adds to the minimum peak list (step S402). Similar to the first embodiment, min (that is, y [i] assigned to min) is a minimum peak. min_pos (that is, x [i] substituted for min_pos) is the date and time of the minimum peak.

  Here, in min_a, the y [i] and x [i] and the latest peak with the newest date / time among the maximum peaks included in the maximum peak list (that is, the last added to the maximum peak list at the present time). And the second slope calculated by the date and time are substituted. That is, if the maximum peak is y [i ′] and the date and time of the maximum peak is x [i ′], the second slope is (y [i] −y [i ′]) / (x [i ] −x [i ′]).

  Also, in min_d, the second difference calculated by the y [i] and the median value of the upper and lower limits of all y [i] (i = 0,..., M−1). Is substituted. That is, if the median value (intermediate value) of the upper and lower limits of y [i] (i = 0,..., M−1) is μ, the second difference is y [i] −μ. Is calculated by As described above, for example, μ may be a median value of all the maximum peaks y [i] and minimum peaks y [i ′] during a predetermined time width (for example, 10 minutes). In this case, after all maximum peaks y [i] and minimum peaks y [i ′] during a predetermined time width are added to the maximum peak list and the minimum peak list, μ is calculated and min_d is set to What is necessary is just to calculate.

  Furthermore, the second interval calculated by the x [i] and the date and time of the previous minimal peak is substituted for min_s. That is, if the date and time of the previous minimal peak is x [i ″], the second interval is calculated by x [i] −x [i ″].

  Thereby, the minimum peak, the date, the second slope, the second difference, and the second interval are stored in the minimum peak list in association with each other.

<Summary of Second Embodiment>
As described above, in the identification system 1 according to the present embodiment, the number of peaks excluding the peak satisfying the predetermined condition among the peaks of the L2 norm of the acceleration sensor value in the predetermined time width is equal to or greater than the predetermined threshold Th. By determining whether or not there is, it is specified whether the behavior of the cow during the time span is foraging behavior or not foraging behavior. As a result, as described above, for example, when an action such as a cow swinging his head or playing with another cow, the situation in which the action is erroneously identified as a foraging action Can be reduced. Therefore, according to the specifying system 1 according to the present embodiment, the foraging behavior can be specified with higher accuracy than in the first embodiment.

[Third embodiment]
Next, a third embodiment will be described. In the first embodiment, whether the behavior of the cow is a foraging behavior or not a foraging behavior is determined depending on whether or not the identified peak is equal to or greater than the threshold Th. However, in the first embodiment, for example, a continuous movement that occurs when a cow is in estrus may be erroneously specified as a foraging behavior. Therefore, in the third embodiment, a description will be given of the identification system 1 that can reduce the case where it is erroneously identified as a foraging behavior.

  In the third embodiment, differences from the first embodiment will be mainly described, and the same components as those in the first embodiment will be given the same reference numerals and description thereof will be omitted. .

<Functional Configuration of Specific Processing Unit 100>
First, the functional configuration of the specific processing unit 100 according to the present embodiment will be described with reference to FIG. FIG. 10 is a diagram illustrating an example of a functional configuration of the specific processing unit 100 according to the third embodiment.

  As illustrated in FIG. 10, the identification processing unit 100 according to the present embodiment further includes a tag orientation identification unit 106. The tag orientation identifying unit 106 identifies whether the orientation of the tag 20 attached to the cow is a positive direction or a negative direction. Here, the direction of the tag 20 is a positive direction means that when the tag 20 is attached to the neck portion of the cow without being twisted, a certain axis of a predetermined orthogonal coordinate system fixed to the tag 20 is used. This is when the positive direction and the forward direction of the cow are on the same side. On the other hand, the direction of the tag 20 being a negative direction means that when the tag 20 is attached to the cow's neck without twisting, the negative direction of a specific axis of the orthogonal coordinate system and the forward direction of the cow. Is on the same side.

  In the present embodiment, the tag 20 is assumed to be mounted below the neck portion of the cow, and the specific axis of the tag 20 is assumed to be the Y axis. Further, when the tag 20 is attached to the lower part of the cow's neck without twisting, the Y axis is parallel to the horizontal direction, and the positive direction of the Y axis is the forward direction of the cow. Is an axis orthogonal to the Y axis in the vertical plane, and the positive direction is the direction of gravity. For example, the X axis may be a coordinate axis that is orthogonal to the Y axis and the Z axis and constitutes the left hand system coordinate with the Y axis and the Z axis.

  At this time, the foraging identification unit 105 according to the present embodiment has a predetermined time width (depending on whether the direction of the tag 20 (hereinafter also referred to as “tag direction”) is the positive direction or the negative direction. For example, it is a foraging action based on the magnitude relationship between a predetermined index value calculated from an average value of Y components of acceleration sensor values (hereinafter also referred to as “Y component average”) for 10 minutes and a predetermined threshold value. Specify whether or not.

<Identification of foraging behavior>
In the following, the process for identifying the foraging behavior of the cow will be described with reference to FIG. FIG. 11 is a flowchart illustrating an example of a foraging action specifying process according to the third embodiment. Note that steps S101 to S107 in FIG. 8 are the same as those in FIG.

  If it is determined in step S105 that the number of peaks specified by the peak specifying unit 104 is equal to or greater than the threshold value Th, the specifying processing unit 100 determines whether the direction of the tag 20 attached to the cow is positive or negative. Whether it exists is identified (step S501). Here, as will be described later, in the process of specifying the orientation of the tag 20, the YZ value, which is an example of the index value, is calculated from the Y component average over a predetermined time width. Details of the process for specifying the orientation of the tag 20 will be described later.

  Next, the foraging identification unit 105 determines whether the tag direction identified in step S501 is the positive direction or the negative direction (step S502).

If it is determined in step S502 that the tag direction is the positive direction, the foraging identification unit 105 determines that the YZ value during the time width (for example, 10 minutes) of the measurement data acquired in step S101 is a predetermined threshold value. It determines whether or not L _a or more (step S503). Here, the threshold value _{L A,} for example, by using a _{Y a} and _{Y b} which will be described later, 0 <C <1 becomes constant _C, is represented by _{L A = (Y a -Y b} ) × C. For example, C may be 0.4.

Thus, if the YZ value is determined to be equal to or greater than the threshold value L _A, it is identified and actions cow between the time width of feeding behavior. On the other hand, if the YZ value is determined to not more than the threshold value L _A, behavior of cattle between the time width is determined not to be feeding behavior.

If it is determined in step S502 that the tag orientation is negative, the foraging identification unit 105 determines that the YZ value during the time width (for example, 10 minutes) of the measurement data acquired in step S101 is a predetermined threshold value. equal to or lower than L _B (step S504). Here, the threshold value _{L B,} as well as the threshold value _{L A,} for example, by using a _{Y a} and _{Y b} which will be described later, 0 <C <1 becomes constant _{C, L B = (Y a} -Y b) × Represented by C. For example, C may be 0.4.

Thus, when the YZ value is determined to be equal to or less than the threshold value L _B, it is identified and actions cow between the time width of feeding behavior. On the other hand, if the YZ value is determined not to be the threshold value L _B below, behaviors cattle between the time width is determined not to be feeding behavior.

  As described above, the identification system 1 according to the present embodiment determines whether the index value (YZ value) in a predetermined time width is greater than or less than a predetermined threshold according to the orientation of the tag 20. It is specified whether the behavior of the cow during the time span is foraging behavior or not foraging behavior. This reduces, for example, a situation in which a continuous movement that occurs when a cow is in estrus (ie, a continuous movement when the cow is raised) is erroneously identified as a foraging behavior. be able to.

<Identification of tag orientation>
Hereinafter, the details of the processing in step S501 (processing for specifying the orientation of the tag 20) will be described with reference to FIG. FIG. 12 is a flowchart illustrating an example of a tag orientation specifying process according to the third embodiment.

  First, the acquisition unit 101 acquires the acceleration sensor value of the Y component and the acceleration sensor value of the Z component from the measurement data storage unit 200 for the past 24 hours of the corresponding tag ID (for example, tag ID “Tag1”). (Step S601). Note that the past 24 hours is an example, and the acceleration sensor value of the Y component and the acceleration sensor value of the Z component during an arbitrary time may be acquired.

  Next, the tag orientation identification unit 106 calculates the following index value every 10 minutes (step S602).

-YZ value = Y component average x sign (Z component average)
Y component standard deviation Here, the Y component average is the average of the Y components of the acceleration sensor values over 10 minutes as described above. The Z component average is the average of the Z components of the acceleration sensor value for 10 minutes. The Y component standard deviation is the standard deviation of the Y component of the acceleration sensor value for 10 minutes. sign is a sign function. However, sign (0) returns positive.

  By using the YZ value defined above as an index value, for example, even when the tag 20 attached to the cow is twisted, the index has no influence of the sign inversion of the Y component accompanying the twist. The value can be used. That is, when the tag 20 is twisted, the signs of the Y component and the Z component are both reversed. Therefore, by multiplying the Y component average by the sign of the Z component average, the sign of the Y component accompanying the occurrence of the twist The influence of inversion can be eliminated.

  Note that calculating the index value every 10 minutes is an example, and the index value may be calculated every arbitrary time width.

Next, the tag orientation identification section 106 of the YZ value calculated in step S602 described above, Y component standard deviation to calculate the median value _{Y a} of YZ value exceeds a predetermined threshold value _{S A} (step S603) . Here, the threshold value S _A can be set to an arbitrary value, but for example, S _A = 150 may be set.

Next, the tag orientation identification section 106 of the YZ value calculated in step S602 described above, Y component standard deviation to calculate the median _{Y b} of YZ value is below a predetermined threshold value _{S B} (step S604) . Here, the threshold value S _B can be set to an arbitrary value, but for example, S _B = 40 may be set.

Finally, the tag orientation specifying unit 106 specifies the tag orientation by tag orientation = sign (Y _a −Y _b ) (step S605). Thereby, when the sign (Y _a -Y _b ) is positive, the tag direction is specified as “positive direction”, and when the sign (Y _a -Y _b ) is negative, the tag direction is specified as “negative direction”.

Note that step S604 described above may not be executed. In this case, in step S605 described above, the tag orientation may be specified by tag orientation = sign (Y _a ).

<Summary of third embodiment>
As described above, the identification system 1 according to the present embodiment takes the direction of the tag 20 into consideration, and uses the index value (YZ value) calculated from the Y component average in a predetermined time width, during the time width. Identify whether the cow's behavior is foraging or not. Thus, as described above, for example, a continuous movement that occurs when a cow is in estrus (ie, a continuous movement in a state where the cow is raised) is erroneously identified as a foraging action. Can reduce the situation.

  The present invention is not limited to the specifically disclosed embodiments, and various modifications and changes, combinations of the embodiments, and the like are possible without departing from the scope of the claims.

DESCRIPTION OF SYMBOLS 1 Specific system 10 Specific apparatus 20 Tag 100 Specific processing part 101 Acquisition part 102 Preprocessing part 103 Norm calculation part 104 Peak specific part 105 Foraging specific part 200 Measurement data storage part

Claims (12)

A specific device for identifying foraging behavior of livestock,
Storage means for storing acceleration data measured by an acceleration sensor mounted on the livestock;
Obtaining means for obtaining a plurality of acceleration data during a predetermined time from the storage means;
Calculating means for calculating feature amounts of each of the plurality of acceleration data acquired by the acquiring means;
Based on a plurality of feature amounts calculated by the calculation means and a preset parameter, a maximum peak indicating a locally maximum feature amount, and a minimum peak indicating a locally minimum feature amount, Peak identifying means for identifying at least one of the peaks,
Determining means for determining whether or not the number of peaks specified by the peak specifying means is equal to or greater than a predetermined threshold;
When the determination means determines that the number of peaks is equal to or greater than the predetermined threshold, behavior specifying means for specifying the behavior of the livestock as the foraging behavior during the predetermined time;
A specific apparatus comprising: The peak specifying means includes
Assuming that the plurality of feature quantities that are time-series data are y [i] and the parameter is δ,
For a certain i, a value obtained by subtracting y [i-1] from y [i] is larger than δ, and a value obtained by subtracting y [i + 1] from y [i] is larger than δ. identify y [i] as a maximal peak,
For a certain i, a value obtained by subtracting y [i] from y [i−1] is greater than δ, and a value obtained by subtracting y [i] from y [i + 1] is greater than δ. The specifying device according to claim 1, wherein y [i] is specified as a minimum peak. The peak specifying means includes
The plurality of feature quantities that are time-series data are y [i], the parameters are δ, and L,
For a certain i, the value obtained by subtracting y [i + 1] from y [i] is greater than δ, and the maximum value from y [i] to y [i + L-1] is y [i] or less. If y [i] is identified as a maximal peak,
For a certain i, the value obtained by subtracting y [i] from y [i + 1] is greater than δ, and the minimum value from y [i] to y [i + L-1] is less than or equal to y [i]. 2. The specifying device according to claim 1, wherein y [i] is specified as a minimum peak. The determination means includes
4. The method according to claim 1, wherein a determination is made as to whether or not the number of peaks excluding a peak satisfying a predetermined condition from the peaks specified by the peak specifying unit is equal to or greater than the threshold value. 5. The specific device according to item. The determination means includes
The slope of the straight line from the minimum peak to the maximum peak is not less than a predetermined threshold, the difference between the maximum peak and the median value of the upper limit value and the lower limit value of the feature quantity is not less than a predetermined threshold, or the previous maximum peak and the above Exclude local peaks that meet one of the thresholds with the maximum peak interval,
The absolute value of the slope of the straight line from the maximum peak to the minimum peak is not less than a predetermined threshold, and the absolute value of the difference between the minimum peak and the median value of the upper limit value and the lower limit value of the feature amount is not less than a predetermined threshold, or one 5. The specifying apparatus according to claim 4, wherein a minimum peak satisfying any one of an interval between a previous minimum peak and the minimum peak equal to or greater than a predetermined threshold is excluded. When the determination means determines that the number of peaks is equal to or greater than the predetermined threshold, the direction specifying means for specifying the direction of the acceleration sensor,
The action specifying means includes
The first index value calculated from the average value of the predetermined component of the acceleration data during the predetermined time according to the direction of the acceleration sensor specified by the direction specifying means is greater than or equal to a predetermined value. 6. The identification device according to claim 1, wherein when it is determined that the behavior is not more than the following, the behavior of the livestock during the predetermined time is identified as a foraging behavior. . The orientation specifying means includes
Of the X component value, the Y component value, and the Z component value included in the acceleration data, the second index value and the third index value are determined for each predetermined time width using two predetermined component values. Calculate
The specifying device according to claim 6, wherein a direction of the acceleration sensor is specified based on the second index value and the third index value. The second index value is a value obtained by multiplying the average value of the predetermined first component values in the time width by the sign of the average value of the predetermined second component values in the time width,
The third index value is a standard deviation of the first component value in the time width,
The first index value is an average value of the first component values in a time width corresponding to the predetermined time among the second index values calculated for each time width, and the predetermined index value. The identification device according to claim 7, wherein the second index value is calculated by multiplying a sign of an average value of the second component values in a time width corresponding to a period of . The determination means includes
It is determined whether any of the number of maximum peaks specified by the peak specifying means, the number of minimum peaks, or the total number of maximum peaks and minimum peaks is equal to or greater than the predetermined threshold value. The specifying device according to any one of claims 1 to 8.   The identification device according to claim 1, wherein the feature amount is an L2 norm. An identification device for identifying foraging behavior of livestock, the identification device having storage means for storing acceleration data measured by an acceleration sensor mounted on the domestic animal,
An acquisition procedure for acquiring a plurality of acceleration data during a predetermined time from the storage means;
A calculation procedure for calculating a feature amount of each of the plurality of acceleration data acquired by the acquisition procedure;
Based on a plurality of feature amounts calculated by the calculation procedure and a preset parameter, a maximum peak indicating a locally maximum feature amount, and a minimum peak indicating a locally minimum feature amount; A peak identification procedure for identifying at least one of the peaks,
A determination procedure for determining whether or not the number of peaks identified by the peak identification procedure is greater than or equal to a predetermined threshold;
When it is determined by the determination procedure that the number of peaks is equal to or greater than the predetermined threshold, an action specifying procedure for specifying the behavior of the livestock as the foraging action during the predetermined time;
The specific method characterized by performing.   The program for functioning a computer as each means in the specific apparatus as described in any one of Claims 1 thru | or 10.

Images