KR102537863B1 - Sow management and Environmental control System - Google Patents

Abstract

The present invention relates to a sow management and environmental control system. According to the present invention, a system for predicting and managing sow birth and controlling the environment comprises: an environment control device installed in a breeding farm to control the sow's environment; An environmental sensor unit for detecting the sow's environment, one or more 3D sensors for photographing the sow, and a sow measurement unit including one or more thermal imaging cameras for photographing the sow, and the sow image extracted through the 3D sensor and the thermal image camera, A terminal including a sow management/environmental control application for recognizing unusual behavior of sows using the thermal image, predicting delivery of sows, receiving environmental information sensed from the environment sensor unit, and controlling the environment control device. It is possible to provide a sow management and environmental control system comprising a.

Description

Sow management and environmental control system {Sow management and Environmental control System}

The present invention relates to a sow management and environmental control system, and more particularly, recognizes a peculiar behavior based on a 3D image, predicts the time of delivery of a sow, and adjusts the environment to an optimal condition, thereby automatically and accurately determining the time of delivery of a sow. The present invention relates to a sow management and environmental control system capable of improving productivity of sows by making decisions and responding quickly.

The farrowing of sows is closely related to economic feasibility in terms of the pig farming industry, and is the most labor-intensive stage in terms of pig farming households.

Because the manager predicts the time of delivery by observing the change of the sow based on experience, there is a tendency to predict the time of delivery subjectively. In addition, in order to know the exact time of delivery, the sow must be observed for a long time, which requires considerable labor and time.

In addition, pigs, like humans, calculate the expected delivery date when fertilization is confirmed. Depending on the pig's delivery experience (number of pregnancies), health condition, and farm environment, the delivery time changes from the expected delivery date. has difficulty in predicting

If the time of delivery is not accurately recognized, problems such as dystocia and postpartum stagnation may occur, and if these problems are not promptly addressed, productivity of livestock farms may decrease due to health problems of sows and stillbirth of piglets. there is.

Therefore, it is necessary to develop an automation technology capable of managing sows through continuous monitoring and objective farrowing prediction.

However, the currently developed technology for continuously monitoring and quantifying the sow's behavior is still insufficient for industrial application.

On the other hand, in general, pig breeding environment management is the temperature and humidity of the pig house, and ventilation. This is because when a large number of pigs are reared in a small space, humidity becomes a problem when focusing on temperature, and ventilation becomes insufficient when focusing on warmth. This is because problems may occur. Moreover, the pig farrowing room is a place where newborn piglets (pilets) are raised, and if the temperature, humidity, and ventilation of the place are not kept constant, many piglets may die due to respiratory diseases or the like.

In order to solve the above problems, the present invention predicts the time of delivery of the sow by recognizing the specific behavior based on the 3D image and adjusts the environment to the optimal condition, thereby automatically, accurately and quickly determining the time of delivery of the sow. An object of the present invention is to provide a sow management and environmental control system capable of responding to sows and improving productivity of sows.

In order to solve the above problems, the sow management and environment control system according to an embodiment of the present invention is a system for predicting and managing the farrowing of sows and controlling the environment, which is installed in a breeding farm to control the environment for sows. control device; An environmental sensor unit for detecting the sow's environment, one or more 3D sensors for photographing the sow, and a sow measurement unit including one or more thermal imaging cameras for photographing the sow, and the sow image extracted through the 3D sensor and the thermal image camera, A terminal including a sow management/environmental control application for recognizing unusual behavior of sows using the thermal image, predicting delivery of sows, receiving environmental information sensed from the environment sensor unit, and controlling the environment control device. It is possible to provide a sow management and environmental control system comprising a.

Here, the sow measurement unit may further include a receiving frame in which a space for accommodating sows is provided and the environmental sensor unit, the 3D sensor, and the thermal image camera can be installed.

In addition, the sow management/environment control application includes an image extraction unit for extracting one or more sow images from images captured by the 3D sensor and one or more thermal images from images captured by the thermal imaging camera; a conversion unit that converts the extracted sow image into 3D coordinate data; a pre-processing unit for generating 3-dimensional sow data by pre-processing the converted 3-dimensional coordinate data; Through the learned artificial intelligence model, the environmental control device receives environmental information detected from the delivery prediction unit and the environment sensor unit that recognizes a specific behavior using the 3-dimensional sow data and thermal image and predicts the time of delivery. It may include an environment control unit for controlling.

In addition, the pre-processing unit includes a factor extraction unit that extracts factors from the three-dimensional coordinate data, and the factors include head direction, floor angle, focal length, height, length 1, chest half circumference, middle half circumference, posterior half circumference, It is characterized in that it includes at least one of the average torso half circumference, the sum of the torso half circumferences, the number of point clouds, the volume, and feature points.

(Where length 1 is the distance from the front leg with the head removed to the tip of the buttocks, and length 2 is the distance between the front leg and the hind leg.)

In addition, the farrowing predictor recognizes the unusual behavior of the sow using the three-dimensional sow data and the thermal image through the learned artificial intelligence model, and the unusual behavior recognition unit counts the recognized specific behavior to generate behavior analysis data. And it may include a delivery time determination unit for predicting the delivery time of the sow by analyzing the behavior analysis data.

In addition, a weight derivation unit for deriving an estimated weight from the three-dimensional sow data may be further included, and the delivery prediction unit predicts a delivery time by referring to the derived estimated weight.

In addition, the weight derivation unit selects one or more of the 3D sow data according to a factor to generate a candidate list, determines the number of conditions satisfied in the generated candidate list, and derives the first weight of the 3D sow data. a candidate list unit to do; An algorithm derivation unit for deriving a second weight by applying a multiple regression formula to the 3D sow data and a final derivation unit for deriving an estimated weight using the first weight and the second weight.

In addition, the candidate list unit selects one or more of the three-dimensional sow data as candidates according to a factor and determines the number of cases in which the candidate list is satisfied with the list generator and the generated candidate list. A list processing unit for deriving the first weight of the dimensional sow data may be included.

In addition, the list generator compares at least one of the head direction, floor angle, focal length, length 1, and total body half circumference of the 3D sow data with sow standard data to generate a first candidate list, and generates a first candidate list. A second candidate list is created by comparing one or more of the head direction, total body half circumference, height, length 1, and median half circumference of the data with the sow standard data, and the head direction, length 1, and chest half circumference of the 3-dimensional sow data , It is characterized in that a third candidate list is generated by comparing at least one of the middle and second half circumferences with the sow standard data.

The sow management and environment control system according to an embodiment of the present invention as described above predicts the delivery time of the sow by recognizing the unusual behavior based on the 3D image, and adjusts the environment to the optimal condition, thereby automatically determining the delivery time of the sow. Accurate and prompt judgment is made so that sows can be easily managed.

In addition, it is possible to quickly respond to problems that may occur during childbirth, such as difficult delivery, so that productivity of sows can be improved.

In addition, it can bring the effect of reducing labor force and improving the profit of farmhouses.

In addition, as it can be applied to various livestock such as pigs as well as cattle, its utilization is expected to expand.

1 is a configuration diagram showing a sow management and environment control system according to an embodiment of the present invention.
FIG. 2 is an exemplary diagram illustrating a state in which an environment sensor unit, a 3D sensor, a thermal imaging camera, and a terminal of a sow management and environment control system according to an embodiment of the present invention are installed in a receiving frame.
Figure 3 is a block diagram showing the configuration of the sow management / environmental control application of Figure 1;
4 is an exemplary view showing a factor;
Figure 5 is a block diagram showing the configuration of the weight derivation unit of Figure 3;
Figure 6 is a block diagram showing the configuration of the delivery prediction unit of Figure 3;

Hereinafter, the description of the present invention with reference to the drawings is not limited to specific embodiments, and various transformations may be applied and various embodiments may be applied. In addition, the content described below should be understood to include all conversions, equivalents, or substitutes included in the spirit and scope of the present invention.

In the following description, terms such as first and second are terms used to describe various components, and are not limited in meaning per se, and are used only for the purpose of distinguishing one component from another.

Like reference numbers used throughout this specification indicate like elements.

Singular expressions used in the present invention include plural expressions unless the context clearly dictates otherwise. In addition, terms such as "include", "include" or "have" described below are intended to designate that features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist. should be construed, and understood not to preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the present invention belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and unless explicitly defined in the present application, it should not be interpreted in an ideal or excessively formal meaning. don't

In addition, in the description with reference to the accompanying drawings, the same reference numerals are given to the same components regardless of reference numerals, and overlapping descriptions thereof will be omitted. In describing the present invention, if it is determined that a detailed description of related known technologies may unnecessarily obscure the subject matter of the present invention, the detailed description will be omitted.

In the present invention, the sow management/environment control application is preferably provided in a mobile communication terminal including a smart phone communicating through a wireless communication network, but may also be provided in a computer environment communicating through a wired communication network. Hereinafter, the mobile communication terminal will be mainly described.

Hereinafter, an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 6 attached.

1 is a configuration diagram showing a sow management and environment control system according to an embodiment of the present invention, and FIG. 2 is an environment sensor unit, 3D sensor, and thermal imaging camera of the sow management and environment control system according to an embodiment of the present invention. and an example view showing a terminal installed in an accommodation frame, FIG. 3 is a block diagram showing the configuration of the sow management/environment control application of FIG. 1, FIG. 4 is an example view showing a factor, and FIG. 3 is a block diagram showing the configuration of the weight derivation unit, Figure 6 is a block diagram showing the configuration of the delivery prediction unit of FIG.

The sow management and environmental control system according to an embodiment of the present invention measures the sow's environment to set conditions suitable for growth and predicts the time of delivery through the 3D image and temperature information of the sow, so that the sow can be easily managed. want to do

Referring to FIG. 1 , a sow management and environmental control system (hereinafter referred to as 'system') according to an embodiment of the present invention may include an environment control device 1, a sow measurement unit 2, and a terminal 3. there is.

The environment control device 1 may be installed in the breeding farm of sows to control the environment of sows. In this case, the environment control device 1 may include one or more of a heating device, a ventilation device, and a humidity control device, and may be individually controlled and operated by the terminal 3 . It is not limited thereto and may further include various devices capable of controlling the environment.

The sow measuring unit 2 is to secure sow appearance information, temperature information, surrounding environment information, etc., and may include an environmental sensor unit 20, a 3D sensor 21, and a thermal imaging camera 22. there is.

The environment sensor unit 20 may secure environmental information by detecting an environment around the sows so as to control the environment of the sows under optimal conditions. In addition, the measured environmental information may be transmitted to the terminal 3 .

The environmental sensor unit 20 may include one or more of a temperature sensor, a humidity sensor, and a gas sensor, preferably including all of them, but is not limited thereto, and may further include various sensors such as a wind speed sensor.

In addition, the gas sensor may include an ammonia sensor, a hydrogen sulfide sensor, and the like.

Accordingly, the environmental information may include environmental temperature information, humidity information, gas information, and the like.

In addition, the environment sensor unit 20 can be installed around the sows even in the breeding farm to more accurately detect the surrounding environment of the sows. This is because the environmental values detected may vary depending on the location in the breeding farm, which is a large space.

One or more 3D sensors 21 may be installed to photograph the sows and secure a three-dimensional image of the sows. At this time, the 3D sensor 21 is a sensor that captures a 3D image with a depth sensor, and preferably a color depth sensor, but is not limited thereto. This is because when a color depth sensor is used, the color is expressed differently according to the depth, so the accuracy of sow shape extraction can be improved.

One or more thermal imaging cameras 22 may be installed to photograph the sows and secure thermal images of the sows. The thermal imaging camera 22 is a special device that captures images using heat, tracks and detects the heat, and shows it as an image. Accordingly, when a living creature (object) having a temperature is photographed, the object can be identified more easily by being distinguished from surrounding objects, and the temperature distribution of the object can be seen at a glance.

In the case of the system of the present invention, there is a feature that the temperature near the tailbone is higher than other parts in the case of sows.

In addition, as shown in FIG. 2, the sow measurement unit 2 may further include an accommodation frame 23.

The accommodation frame 23 is provided with a space to accommodate sows, and may be formed as a frame, but is not limited thereto. In addition, an environment sensor unit 20, a 3D sensor 21, and a thermal imaging camera 22 may be installed. Accordingly, it is possible to photograph the sow accommodated in the accommodation frame 23 and accurately measure the surrounding environment.

The terminal 3 may be a mobile terminal of a user such as a farmer or manager, and may be easily applied to a PC tablet in addition to a mobile terminal.

As described above, the terminal 3 is implemented as a mobile terminal, so it is easy to install, move, and conveniently, so that the manager can easily predict the delivery time of the sow.

In addition, the terminal 3 may be installed on the receiving frame 23, but it is implemented to be detachable and can be used in various ways as needed.

In addition, the terminal 3 has a sow management/environmental control application 30 installed thereon, so that it can receive environmental information, captured images, etc. from the environment sensor unit 20, 3D sensor 21, and thermal imaging camera 22. And, it is possible to control the environment control device (1).

The sow management/environmental control application 30 recognizes the specific behavior of the sow using the sow image and the thermal image, predicts the time of delivery through this, and controls the breeding farm to the optimal environmental condition according to the environmental information. Android , iOS-based general application, but may be provided as a general application or web service-based application according to the terminal 3 or wired/wireless service type. As a method of providing, it may be downloaded by accessing the server through the terminal 3 or downloaded and installed through an online application market (eg, Android market, Apple store, online market of a telecommunications company, etc.).

Referring to Figure 3, sow management / environmental control application 30 data storage unit 31, image extraction unit 32, conversion unit 33, pre-processing unit 34, weight extraction unit 35, farrowing A prediction unit 36 and an environment control unit 37 may be included.

The data storage unit 31 may store specific behavior data, farrowing time data according to specific behaviors, sow standard data generated using pre-collected sow data, optimal environmental conditions according to sow conditions, and the like.

Here, the specific behavior data may be data on specific behaviors that may appear before the sow is born. The unusual behavior data may include a unique behavior sow image, 3D sow data, factor information according to specific behavior, and the like.

The delivery time data according to the specific behavior may be data on the delivery time according to the frequency, cycle, staged behavior, and continuous behavior of the specific behavior.

In addition, sow standard data is standardized data by month or weight by analyzing sow data collected in advance by measuring and 3D photographing sows in the farmhouse, and may include 3-dimensional sow data by month or weight, and collected in advance. One sow data may include 3-dimensional sow data and weight.

In addition, the data storage unit 31 may store biometric information of the sow, behavior analysis data, estimated weight, and information about a predicted delivery time input from the user.

The biometric information of the sow is the biometric information of the sow to be managed, and may include one or more of sow classification number, species, number of months, and expected fertilization date, and it is preferable to include all of them.

The image extraction unit 32 may receive an image captured by the 3D sensor 21 and extract one or more sow images from the captured image through capture. In addition, one or more thermal images may be extracted from the captured image by receiving an image captured by the thermal imaging camera 22 .

It is preferable to capture multiple sow images and thermal images through continuous capture, and continuous capture can be performed according to the value of the number of consecutive captures. The value of the number of consecutive captures may be 10 to 12, but is not limited thereto.

In addition, capturing may be performed at regular intervals, for example, at 2-minute intervals to recognize and count unusual behaviors, but is not limited thereto.

In this way, performance degradation can be minimized by extracting sow images and thermal images through capture.

After arranging the sow images, the conversion unit 33 may extract points and convert them into 3D coordinate data. This 3D coordinate data may be stored in the data storage unit 31 .

Here, the 3D coordinate data may be data on the right or left shape (half shape of the sow) based on the sow's isometric axis, in which all the sow's shape is not implemented, but is not limited thereto, and the sow's shape is all implemented. It may be.

The pre-processing unit 34 may pre-process the converted 3-dimensional coordinate data to generate 3-dimensional sow data. To this end, it may include a validation unit, a noise removal unit, and a factor extraction unit.

The validity verification unit judges the validity of the 3D coordinate data. Since a sow image unsuitable for use may be obtained due to obstruction factors (sunlight, movement of sows, etc.), the validity is verified by checking the number of extracted points. can judge

Validity can be judged according to whether the number of extracted points is greater than or equal to a set number of points, and if it is less than a certain number of points, the corresponding 3D coordinate data can be eliminated so that it is not used to recognize unusual behavior or derive estimated weight. there is.

The noise removal unit removes the noise of the 3D coordinate data, and since other objects (land, building frame, etc.) can be removed

The factor extraction unit may generate 3D sow data by extracting factors from 3D coordinate data. In this case, the 3D sow data may include 3D coordinate data and factors.

At this time, the factors are head direction, floor angle, focal length, height, length 1 (length01), chest half circumference (girth_first), middle half circumference (girth_mid), back half circumference (girth_end), torso half circumference average, It may be one or more of the total body half circumference, the number of point clouds, volume, and feature points, but is not limited thereto, and various factors such as curvature or not may be extracted.

As shown in FIG. 4, length 1 is the distance from the front leg with the head removed to the tip of the buttocks, and length 2 is the distance between the front leg and the hind leg.

In addition, the feature point may be a point corresponding to one or more of the tip of the nose, the tip of the lower mouth, the tip of the leg, the knee, the tailbone, and the tip of the ear of the sow, but is not limited thereto, and may include more diverse feature points.

In this way, by extracting the factor corresponding to the feature part, the specific behavior of the sow can be easily recognized, and the weight can also be estimated.

The weight derivation unit 35 may derive an estimated weight from 3D sow data.

Here, when there are two or more 3D sow data, the weight derivation unit 35 repeats the process of deriving the estimated weight according to the number of 3D sow data to obtain the estimated weight according to each 3D sow data, Estimated weight data corresponding to the number can be obtained. That is, the estimated weight is obtained for each 3-dimensional sow data.

The estimated weight data may include the derived estimated weight and 3-dimensional sow data, but is not limited thereto and may further include other information.

Referring to FIG. 5 , the weight derivation unit 35 may include a candidate list unit 350 , an algorithm derivation unit 351 and a final derivation unit 352 .

The candidate list unit 350 selects one or more 3D sow data according to a factor to generate a candidate list, determines the number of conditions satisfied in the generated candidate list, and derives a first weight of the 3D sow data. can

To this end, the candidate list unit 350 may include a list generator and a list processor.

The list generator may generate a candidate list by selecting one or more three-dimensional sow data as candidates according to factors.

Specifically, the list generator compares one or more three-dimensional sow data through factors based on the sow standard data, and selects candidate 1, candidate 2, and candidate 3 as first, second, and second candidates if they belong to the sow standard data. By adding to the list of 3 candidates, the final first, second and third candidate lists can be created.

In this case, when the 3D sow data is selected as a candidate, the weight according to the matched sow standard data may be included as the weight of the 3D sow data.

The list generator may generate a first candidate list by comparing one or more of the head direction, floor angle, focal length, length 1, and total body half circumference of the 3D sow data with the sow standard data.

The list generator may compare the sow standard data, select the corresponding 3-dimensional sow data as candidate 1, and add it to the first candidate list if the 3-dimensional sow data belongs therein. Through such a process, a first candidate list may be generated.

More specifically, the list generator determines whether the 3D sow data is 'same as the direction of the head' of the sow standard data, 'is within the range of the floor angle', 'is within the range of focal length', 'is within the range of length 1', 'is within the range of half the body' Comparison is performed in the order of 'is within the total circumference range', and only in the case of YES (correct), only 3D sow data that is YES in all cases can be added to the first candidate list.

It is not limited to the above order, and may be easily changed and applied.

In addition, the list creation unit may also create a first candidate backup list. The first candidate backup list is added to the first candidate backup list by adding 3D sow data that is NO (no) only in the last order, 'whether or not the total body half circumference is within the range'. You can create a backup list.

When added to the list, the 3D sow data may include the weight according to the corresponding sow standard data as the weight of the 3D sow data.

In addition, the list generator compares at least one of the head direction, total body circumference, height, length 1, and median circumference of the 3D sow data with the sow standard data, and if the 3D sow data falls within it, the corresponding 3D sow data A second candidate list can be created by selecting as candidate 2 and adding it to the second candidate list.

More specifically, the list generator determines whether the 3D sow data is 'same as the direction of the head' of the sow standard data, 'is within the range of the total half of the trunk', 'is within the range of height', 'is within the range of length 1', 'median' Comparison is performed in the order of whether it is within the half-circumference range, and only in the case of YES (correct), only 3D sow data that is YES in all cases can be added to the second candidate list. It is not limited to the above order, and may be easily changed and applied.

In addition, the list creation unit may also create a second candidate backup list. The second candidate backup list is added to the second candidate backup list by adding 3D sow data that is NO (no) only in the last order, 'is within the mid-range circumference'. You can create a list.

In addition, the list generator compares one or more of the head direction, length 1, chest half circumference, middle half circumference, and posterior half circumference of the 3D sow data with the sow standard data, and if the 3D sow data belongs to it, the corresponding 3D A third candidate list can be generated by selecting sow data as candidate 3 and adding it to the third candidate list.

More specifically, the list generation unit determines whether the 3D sow data is 'same as the head direction' of the sow standard data, 'is within the range of length 1', 'is within the range of the chest circumference', 'is within the range of the middle circumference', ' Comparison is performed in the order of 'is within the range of the posterior half circumference', and only in the case of YES (correct), only 3D sow data that is YES in all cases can be added to the third candidate list. It is not limited to the above order, and may be easily changed and applied.

In addition, the list creation unit may also create a third candidate backup list. The third candidate backup list is added to the third candidate backup list by adding 3D sow data that is NO (no) only in the last order, 'whether it is within the post-violation range'. You can create a list.

The list processing unit may determine the number of conditions satisfied in the generated candidate lists (first, second, and third candidate lists), and may derive a first weight of the 3D sow data in some cases.

Satisfying the condition is whether one or more 3-dimensional sow data is present, and as the first, second, and third candidate lists are generated, the number of cases is 3 satisfactions (first to third candidate lists), 2 satisfactions ( First and second candidate lists, second and third candidate lists, first and third candidate lists), one satisfaction (first, second, and third candidate lists), and zero satisfaction may occur.

The algorithm derivation unit 351 may derive the second weight by applying a multiple regression formula to the 3D sow data.

Specifically, the algorithm derivation unit 351 uses the height, length 2, average body half circumference, and the number of point clouds of the 3D sow data among the factors of the 3D sow data to calculate the second through Equation 1 below. weight can be derived.

[Equation 1]

2nd weight = m ₁ + (h ₁ × height) + (2ㅣ ₁ × length 2) + (g ₁ × body half average) + (p ₁ × number of point clouds of 3-dimensional sow data)

Here, m ₁ is the weight constant, h ₁ is the height constant, 2ㅣ ₁ is the length 2 constant, g ₁ is the half-circumference constant, and p ₁ is the point cloud constant of the 3D sow data.

In addition, the algorithm derivation unit 351 may derive the second weight using Equations 2 to 4 below.

More specifically, the algorithm derivation unit 351 may derive the second weight through Equation 2 using the height, length 1, average body half circumference, and the number of point clouds of the 3-dimensional sow data among the factors. there is.

[Equation 2]

2nd weight = m ₂ + (h ₂ × height) + (1ㅣ ₂ × length 1) + (g ₂ × body half-circumference average) + (p ₂ × number of point clouds of 3-dimensional sow data)

Here, m ₂ is the weight constant, h ₂ is the height constant, 1ㅣ ₂ is the length 1 constant, g ₂ is the half-circumference average constant, and p ₂ is the point cloud constant of the 3D sow data.

In addition, the algorithm derivation unit 351 may derive the second weight through Equation 3 using the height, the average half-circumference of the body, and the number of point clouds of the 3-dimensional sow data among the factors.

[Equation 3]

Second weight = m ₃ + (h ₃ × height) + (g ₃ × body half-circumference average) + (p ₃ × number of point clouds of 3-dimensional sow data)

Here, m ₃ is the weight constant, h ₃ is the height constant, g ₃ is the half circumference constant, and p ₃ is the point cloud constant of the 3D sow data.

In addition, the algorithm derivation unit 351 may derive the second weight through Equation 4 using the average waist circumference and the number of point clouds of the 3-dimensional sow data among the factors.

[Equation 4]

Second weight = m ₄ + (g ₄ × body half circumference average) + (p ₄ × number of point clouds of 3-dimensional sow data)

Here, m ₄ is the weight constant, g ₄ is the half-perimeter average constant, and p ₄ is the point cloud constant of the 3D sow data.

The final derivation unit 352 may derive the estimated weight using the first weight and the second weight, and may derive the estimated weight according to an absolute value of a value obtained by subtracting the first weight from the second weight. When the absolute value is less than the predetermined weight value, the second weight is derived as the estimated weight, and when the absolute value is greater than or equal to the predetermined weight value, the estimated weight may be derived as 0 kg.

The final derivation unit 352 derives the estimated weight through the above process and generates the estimated weight data, so that it can be used to predict the time of delivery, and can be provided to the user and stored in the data storage unit 31.

When stored in the data storage unit 31, the estimated weight may be matched with biometric information of the sow input before measuring the weight of the sow to store the estimated weight.

At this time, if 0 kg is derived as the estimated weight, this may be regarded as a 'weight measurement failure', and accordingly, the final derivation unit 352 informs the user of the 'weight measurement failure' so that the user can be recaptured and re-estimated. .

The delivery prediction unit 36 may recognize a specific behavior and predict a delivery time by using the three-dimensional sow data and the thermal image through the learned artificial intelligence model.

Referring to FIG. 6 , the delivery prediction unit 36 may include a specific behavior recognition unit 360 and a delivery time determination unit 361 .

The unusual behavior recognition unit 360 may recognize unusual behaviors of sows using the three-dimensional sow data and the thermal image through the learned artificial intelligence model, and may generate behavior analysis data by counting the recognized unusual behaviors.

The artificial intelligence model used here is a deep learning algorithm, and it can recognize the unusual behavior of sows through 3D sow data and thermal images by learning specific behavior data.
Meanwhile, the deep learning algorithm as the artificial intelligence model may include a deep belief network, an autoencoder, a convolutional neural network (CNN), a recurrent neural network (RNN), a deep Q-network, and the like, which are known technologies.
Since the technology of recognizing a sow's unusual behavior using a deep learning algorithm is a well-known technology, a detailed description thereof will be omitted below.

At this time, the recognized specific behavior of the sow may be standing, lying down, searching, floor scratching, stall biting, eating, sitting, and chewing, but is not limited thereto, and may be various behaviors.

On the other hand, the unusual behavior recognition unit 360 more accurately recognizes the location, direction, hip position, etc. of sows through thermal images, such as 3D sow data, and recognizes the sow's unusual behavior, so that the recognition accuracy can be further improved. there is.

The unusual behavior recognition unit 360 generates behavior analysis data by counting the recognized unusual behaviors by time, and can be continuously updated through sow images and thermal images extracted at regular intervals.

The delivery time determination unit 361 may predict the delivery time of the sow by analyzing the behavioral analysis data.

Specifically, the delivery time determination unit 361 can predict the delivery time of the sow by analyzing the behavior analysis data through the learned artificial intelligence model. can judge

The artificial intelligence model used here is a deep learning algorithm, which may be learned from childbirth point data according to a specific behavior.

In addition, AI models can identify step-by-step behavior through classification and continuous behavior through regression.

In addition, the delivery time determination unit 361 may predict the delivery time with reference to the estimated weight derived in addition to the behavioral analysis data.

The environment control unit 37 may control the environment control device 1 by receiving environmental information detected by the environment sensor unit 20 . Specifically, the environment control unit 37 may generate optimal environment setting information through an artificial intelligence model learned using the received environment information and the predicted delivery time.

According to the environment setting information, the environment control device 1 can be controlled so that the environment of the breeding farm is set to optimal conditions for breeding sows.

Here, the artificial intelligence model may be learning the optimal environmental condition according to the state of the sow.

In addition, the sow management and environment control system according to an embodiment of the present invention may further include a control unit (not shown).

Only one sow measuring unit 2 and terminal 3 may be provided, or each sow may be provided with a sow measuring unit 2 and a terminal 3, respectively. (not shown) can be managed as a whole accordingly, and can individually control the sow measuring unit 2 and the terminal 3 provided for each sow.

In addition, the control unit may receive all information from the terminal 3 and easily manage it.

As described above, the sow management and environment control system according to an embodiment of the present invention predicts the time of delivery of the sow by recognizing the unusual behavior based on the 3D image, and adjusts the environment to the optimal condition, thereby determining the time of delivery of the sow. automatically, accurately and quickly, so that sows can be easily managed.

In addition, it is possible to quickly respond to problems that may occur during childbirth, such as difficult delivery, so that productivity of sows can be improved.

In addition, it can bring the effect of reducing labor force and improving the profit of farmhouses.

In addition, as it can be applied to various livestock such as pigs as well as cattle, its utilization is expected to expand.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims are also included in the scope of the present invention. that fall within the scope of the right.

1: Environmental control device
2: sow measurement unit
20: environmental sensor unit
21: 3D sensor unit
22: thermal imaging camera
23: receiving frame
3: Terminal
30: sow management/environmental control application
31: data storage unit
32: image extraction unit
33: conversion unit
34: pre-processing unit
35: weight extraction unit
350: candidate list unit
351: algorithm derivation unit
352: final extraction unit
36: delivery prediction unit
360: unusual behavior recognition unit
361: delivery time determination unit
37: environmental control unit

Claims (9)

In the system for predicting and managing the farrowing of sows and controlling the environment,
An environmental control device installed in the breeding farm to control the environment of sows;
A sow measuring unit including an environmental sensor unit for detecting the sow's environment, one or more 3D sensors for photographing the sow, and one or more thermal imaging cameras for photographing the sow; and
Using the sow image and thermal image extracted through the 3D sensor and the thermal imaging camera, recognizing the sow's unusual behavior and predicting the sow's delivery, and receiving the environmental information detected from the environmental sensor unit, the environmental control device Including a sow management / environmental control application that controls,
The sow management/environment control application may include an image extraction unit for extracting one or more sow images from images captured by the 3D sensor and one or more thermal images from images captured by the thermal imaging camera; a conversion unit that converts the extracted sow image into 3D coordinate data; a pre-processing unit for generating 3-dimensional sow data by pre-processing the converted 3-dimensional coordinate data; Through the learned artificial intelligence model, the environmental control device receives environmental information detected from the delivery prediction unit and the environment sensor unit for recognizing unusual behaviors using the 3-dimensional sow data and thermal images and predicting the time of delivery. Including an environment control unit for controlling,
The pre-processing unit includes a factor extraction unit for extracting a factor from the three-dimensional coordinate data,
The factors include head direction, floor angle, focal length, height, length 1, chest half circumference, middle half circumference, posterior half circumference, average torso half circumference, total torso half circumference, number of point clouds, volume and A sow management and environmental control system comprising at least one of the characteristic points.
(Where length 1 is the distance from the front leg with the head removed to the tip of the buttocks, and length 2 is the distance between the front leg and the hind leg.)
According to claim 1,
The sow measurement unit,
A sow management and environmental control system further comprising a receiving frame in which a space for accommodating sows is provided and the environment sensor unit, the 3D sensor, and the thermal imaging camera can be installed.
delete delete According to claim 1 or 2,
The delivery prediction unit,
A peculiar behavior recognition unit that recognizes the sow's unusual behavior using the three-dimensional sow data and thermal image through the learned artificial intelligence model, counts the recognized unusual behavior, and generates behavior analysis data; and
A sow management and environmental control system comprising a delivery time determination unit for predicting the delivery time of the sow by analyzing the behavior analysis data.
According to claim 1 or 2,
Further comprising a weight derivation unit for deriving an estimated weight from the three-dimensional sow data,
The delivery prediction unit,
A sow management and environmental control system, characterized in that for predicting the time of delivery by referring to the derived estimated weight.
According to claim 6,
The weight derivation unit,
a candidate list unit generating a candidate list by selecting one or more 3-dimensional sow data according to a factor, and determining the number of conditions satisfied in the generated candidate list to derive a first weight of the 3-dimensional sow data;
An algorithm derivation unit for deriving a second weight by applying a multiple regression formula to the 3-dimensional sow data; and
A sow management and environmental control system comprising a final derivation unit for deriving an estimated weight using the first weight and the second weight.
According to claim 7,
The candidate list part,
A list generator for generating a candidate list by selecting one or more of the three-dimensional sow data as candidates according to a factor; and
A sow management and environment control system including a list processing unit that determines the number of conditions satisfied in the generated candidate list and derives a first weight of the 3-dimensional sow data according to the case.
According to claim 8,
The list generator,
Comparing at least one of the head direction, floor angle, focal length, length 1, and total body half circumference of the 3-dimensional sow data with sow standard data to generate a first candidate list;
A second candidate list is generated by comparing one or more of the head direction, total body half circumference, height, length 1, and median half circumference of the 3-dimensional sow data with sow standard data;
A sow management and environmental control system characterized in that a third candidate list is generated by comparing one or more of the three-dimensional sow data with head direction, length 1, chest half circumference, middle half circumference, and back half circumference with sow standard data.

Images