WO2023026504A1 - Information processing system, information processing method, and program - Google Patents

Abstract

This invention addresses the problem of providing assistance to realize a series of businesses from livestock breeding to sales without relying on human experience and intuition.　A server 1 has an image analysis AI_Q, a type analysis unit 103, and a process execution unit 104. By analyzing an image of a pigsty B, the image analysis AI_Q recognizes N pig objects OP1 to OPn included in the image of the pigsty B. The type analysis unit 103 generates predetermined types of information (body weight, muscle mass, presence or absence of death/illness, etc.) pertaining to each of the pig objects OP1 to OPn for each of the N pig objects OP1 to OPn. The processing execution unit 104 converts the weight of each of 50 pigs in the pigsty B into the average weight or the like of the 50 pigs, executes a predetermined process (such as arithmetic processing of feed mixture ratio) using the average weight and multiple types of environment information (measured data of temperature and humidity) as input parameters. The problem is solved thereby.

Description

Information processing system, information processing method and program

The present invention relates to an information processing system, an information processing method, and a program.

The livestock industry is constantly faced with the problem of labor shortages, and much of the work is based on the experience and intuition of the people who work there, making it an industry with low reproducibility. It has become.
As a conventional technology, for example, on a website, there is a livestock animal management system that manages changes in the number of livestock animals by inputting activities and comments on livestock animals in chronological order into a terminal (for example, Patent Document 1, etc.).
In the case of the conventional technology, it is necessary for staff such as breeding managers and staff to input each activity and comment, etc. conducted on livestock animals from a terminal. I can say
In addition, large-scale livestock farms raise as many livestock as one city or town, and there is a limit to grasping and managing livestock health conditions, feed mixes, early detection of diseases, etc. with a small number of people. be.

JP 2015-167529 A

In this way, in the case of only conventional technology, it is necessary for the breeding manager and staff to input each activity and comment, etc. regarding the livestock animal from the terminal, and it is a livestock management method in the livestock industry that is already understaffed. cannot be said to be effective.

The present invention was made in view of this situation, and aims to support the realization of a series of businesses from raising livestock to selling them without relying on human experience and intuition.

In order to achieve the above object, an information processing system according to one aspect of the present invention includes:
In an information processing system for executing a predetermined process on N (N is an integer value of 1 or more) livestock managed in a predetermined place as a first unit,
an imaging device that captures the state of the field and outputs the captured image obtained as a result as a field image;
an output device that detects one or more physical quantities related to the field environment and outputs information including the detection result as field environment information;
an information processing device that executes the predetermined process based on at least part of the field image and the field environment information;
including
The information processing device is
By analyzing the field image, N objects representing each of the N livestock contained in the field image are recognized as N livestock objects, and the livestock objects are used as the second unit. image analysis means for generating and outputting a predetermined type of information as second unit analysis information for each of the N livestock objects;
Using the predetermined type of information of the first unit generated based on the N second unit analysis information as the first unit analysis information, one or more of the plurality of types of first unit information and the plurality of types of environment information a predetermined process executing means for executing the predetermined process using the input parameter as an input parameter;
Prepare.
In this way, the N livestock objects representing the N livestock contained in the field image captured by the imaging device are recognized, the livestock objects are analyzed, converted into the second unit, and the predetermined type of information is converted into the second unit. Generated as unit analysis information for each of N livestock objects, and using a predetermined type of information of a first unit generated based on N second unit analysis information as first unit analysis information, a plurality of types of first unit information and Since predetermined processing is executed using one or more of a plurality of types of environmental information as input parameters, it is possible to manage livestock rearing conditions without burdening the person who manages the livestock reared in the field.
As a result, without relying on human experience and intuition, it is possible to support the realization of a series of businesses ranging from livestock breeding to sales.

An information processing method and program corresponding to the information processing apparatus of one aspect of the present invention are also provided as an information processing method and program of one aspect of the present invention.

According to the present invention, it is possible to support the realization of a series of businesses ranging from livestock breeding to sales without relying on human experience and intuition.

It is a figure which shows an example of the management method of the livestock to which the information processing system of this invention is applied. FIG. 2 is a diagram showing an example of identifying an individual pig from a moving image in the livestock management method of FIG. 1; It is a figure which shows the input data and output data with respect to recommendation AI. It is a figure which shows an example of the management data of feed. 1 is a block diagram showing one embodiment of an information processing system of the present invention; FIG. It is a block diagram showing an example of the hardware constitutions of the server concerning the information processing system of the present invention. 7 is a functional block diagram showing the functional configuration of the server of FIG. 6; FIG. 2 is a diagram showing an example of a screen of an administrator terminal in FIG. 1; FIG. FIG. 8 is a flow chart showing the operation of the server of FIGS. 6 and 7; FIG. It is a figure which shows the mode of acquisition of the moving image of the pig before processing and after processing. Fig. 10 shows an embodiment of installing a movable camera in a piggery; FIG. 10 is a diagram showing an embodiment with a smart phone app;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a diagram showing an example of a livestock management method to which an information processing system of the present invention is applied.

As shown in FIG. 1, in the embodiment, livestock to be managed are, for example, pigs P1 to Pn, and these pigs P1 to Pn are managed in units of places such as the pigsty B. A feeding station E is provided in the pigsty B, and the pigs are fed with feed at a predetermined time every day.

In the pigsty B, for example, one camera CA is installed, for example, on the ceiling, wall surface, pillar, or the like, for imaging the conditions of the pigs P1 to Pn. The cameras CA are installed so that all the pigs P1 to Pn in the pigsty B are imaged, that is, there are no blind spots. An image of the pigsty B captured by the camera CA is distributed as a moving image D. That is, the camera CA captures an image of the pigsty B, and outputs the captured image obtained as a result as the moving image D of the pigsty.
A moving image D is a sequence in which a plurality of unit image groups are arranged in chronological order. A unit image includes, for example, a field or a frame. In other words, the image is a broad concept that includes moving images and still images. In this embodiment, the images captured by camera C are distributed as moving images D, but are output as still images captured at predetermined intervals as necessary. may be
In this embodiment, there is one camera CA, but depending on the scale (size) of the pigsty B, a plurality of cameras CA may be arranged. When the pigsty B is partitioned by a plurality of fences, the camera CA is provided for each fence.

The pigsty B is provided with various sensors SE (output devices) for detecting information about the breeding environment inside or outside the pigsty B (hereinafter referred to as "environmental information").
The sensor SE outputs the value of the external environment parameter related to the heat dissipation of the pigs in the pigsty B as the environment information of the pigsty B, for example.
Specifically, for example, the sensor SE is, for example, a temperature sensor, a humidity sensor, or the like, and the sensor SE outputs measurement data such as temperature and humidity as values of external environment parameters. The environmental information shown here is only an example, and may be other information such as weather, weather forecast, air pressure, carbon dioxide concentration, etc., and may be information related to the breeding environment inside the pigsty B. Anything that can be adopted as a value is suitable.

Here, the need for environment information (temperature/humidity) that can be used as external environment parameter values when raising pigs will be described.
First, the need for temperature will be explained.
Pigs increase their feed intake to maintain body temperature. The effect of temperature control is obvious when you look at the reactions of pigs. In cold weather, pigs eat more food to keep themselves warm.
However, no matter how much food is fed and the body is warmed, there is a critical temperature at which heat loss cannot keep up. Exceeding this threshold temperature causes the body temperature to drop, causing the pig to become hypothermic and eventually die.

On the other hand, when it gets too hot, heat production increases and body temperature rises. The pigs then respond by eating less, but there is a limit to this, and high body temperature also eventually leads to death.
Between these temperature extremes there is a temperature zone called the "optimal productivity zone" where productivity is maximized in swine production. The upper limit of this temperature range is called the "upper critical temperature" and the lower limit the "lower critical temperature". The goal of piggery management is to keep pigs in these temperature neutral ranges (optimal productivity range).
The lower critical temperature is, for example, 16° C., and when the temperature of the pigsty B falls below this 16° C., heat radiation increases steadily. Calories are consumed for heat dissipation, so if the base calorie is low, the protein that is supposed to be meat will burn. Therefore, when the temperature of the pigsty B drops by 1° C., it is necessary to increase the amount of feed corresponding to the drop in temperature.

Next, the need for humidity will be explained.
Since pigs sweat, heat dissipation is affected by humidity.
For example, if it rains on a day in midsummer when the temperature in the pigsty B is close to 40° C., the humidity may reach nearly 100%. Also, when heat builds up in pigs, their feeding behavior weakens and their weight gain worsens.
In this way, the temperature and humidity affect the fattening (weight) of the pigs, so it is necessary to change the mix of feed according to the temperature and humidity at that time, while making the environment of pigsty B the optimum productivity range. There is

Next, a method for managing livestock under such a rearing environment will be described.
In step ST1, a moving image D captured by the camera CA is output.
Further, in step ST2, measurement data such as temperature and humidity measured by the sensor SE are output.

In step ST3, the moving image D captured by the camera CA is input to the image analysis AI_Q. Then, the image analysis AI_Q analyzes the moving image D of the pigsty B to recognize and output objects OP1 to OPn indicating the N pigs P1 to Pn included in the moving image D of the pigsty B. Note that n is a natural number.

Here, the image analysis AI_Q has a learning unit, a model obtained as a result of learning, and a recognition unit (inference unit). A model is generated or updated as a result of AI-based machine learning performed by the learning unit on a previously prepared moving image of each individual pig. When a new moving image D is input, this model can label and output objects OP1 to OPn representing the N pigs P1 to Pn included in the moving image. That is, the recognition unit inputs the new moving image D to the model, and outputs the individual pig object with labeling, which is output from the model, to the outside as the recognition result.
Note that the AI model described above is merely an example, and a model with a different output form (for example, a model that performs analysis processing in step ST4 below and outputs the result of the analysis processing) may exist.

Specifically, for example, as shown in FIG. 2, when a moving image D obtained by photographing the behavior of each of the pigs P1 to Pn in the pigsty B is input to the image analysis AI_Q (AI model), The image analysis AI_Q recognizes the contours (body contours) of each pig Pk (k=1 to n) of one or more pigs P1 to Pn included in the moving image D, and generates objects OP1 to OPn indicating the contours of the respective pigs Pk (k=1 to n). , and individual labels, which are individual identifiers, are associated with each other and output.
In each of the objects OP1 to OPn, the solid line (green outline) indicates an active object, and the dashed line (red outline) indicates a stopped object.
That is, by using the image analysis AI_Q in this embodiment, it is possible to identify individual pigs P from the moving image of the pigsty B, and output the identified N objects OP1 to OPn and their individual labels.

At step ST4, one or more types of analysis processing are performed on the N objects OP1 to OPn output from the image analysis AI_Q.
The analysis processing includes, for example, processing for counting the number of pigs P1 to Pn (number), weight gain determination processing for each of the pigs P1 to Pn (body weight), muscle mass, and body length, and processing for each of the pigs P1 to Pn. Behavior pattern prediction processing (behavior pattern), biological condition determination processing for pigs P1 to Pn (death/sickness), intrusion detection processing for objects other than pigs (human detection), and the like. The contents in parentheses correspond to the processing in FIG.
The results of one or more types of analysis processing are output in pigsty B units. Specifically, as the result of the body weight gain determination process (body weight), etc., an average body weight obtained by averaging the body weights of the pigs P1 to Pn is output.

In step ST5, based on the results of one or more types of analysis processing, environmental data (measured data such as temperature and humidity) input from the sensor SE, or prediction data such as weather forecasts in the case of future prediction, Predetermined processing is executed.
Specifically, in the recommendation AI_R, based on the weight gain data (average weight, median, coefficient of variation, etc.) for each piggery B as the analysis processing result, and the temperature and humidity measurement data input from the sensor SE For example, the mixing ratio of various kinds of feed per 100 kg of feed is calculated, and the calculation result is output as recommended information, that is, recommendation information.
Specifically, as shown in FIG. 3, for example, 73% corn, 16% soybean meal, 0.2% premix, 1.41% secondary linker, 0.9% tankal, 5.41% beet valve, Corn lumps 2.58%, salt 0.5%, lysine hydrochloride 0.00%, etc. are output to the manager terminal 2 managed by the manager U as recommendation information.

The data input to the recommendation AI_R and the mixing ratios of various feeds obtained as a result of calculation are collectively stored and managed in a table as shown in FIG. 4 for each month and day.
By automatically inputting and managing the feed data given to pigs in this way, it becomes possible to analyze and learn the correlation between feed mixture ratio and feed amount and body weight gain.

In this embodiment, the recommend AI_R has a learning unit, a model obtained as a result of learning, and a recommendation unit (inference unit).
A model is generated or updated as a result of machine learning by AI using a lot of previously prepared past data in the table shown in FIG. 4 as learning data. This model can output the mixing ratio of various materials shown on the right side of the table shown in FIG. . That is, the recommendation unit uses the weight gain data (average weight, median, coefficient of variation, etc.) for each piggery B, and the temperature and humidity measurement data input from the sensor SE (or the weather forecast, etc. in the case of future prediction). Predicted data of temperature and humidity obtained from ) are input to the model. Then, the recommendation unit outputs the mixing ratio of various feeds output from the model to the outside as recommendation information.

Next, the system configuration of the information processing system shown in FIG. 5 will be described.
FIG. 5 is a block diagram showing one embodiment of the information processing system of the present invention.
The information processing system shown in FIG. 5 includes an imaging device such as a camera CA installed (arranged) to capture an image from above of a pigsty B in which one or more pigs (pigs P1 to Pn in the example of FIG. 1) are housed. , an output device such as a sensor SE installed (arranged) to detect environmental information of the pigsty B, an administrator terminal 2 operated by an operator U such as an administrator or a staff member, and a server 1 are connected to a network. N and these multiple devices are configured to communicate over the network N.
The network N includes wired networks, wireless networks, etc., in addition to the Internet.
The pigsty B is housing means for activating the pigs P1 to Pn within a certain movement range, and may be separated by blocks such as fences. In this case, the pigs P1 to Pn in the pigsty B are managed in units of blocks.

The camera CA is, for example, a digital camera or a network camera that captures moving images, and outputs a moving image D that captures the inside of the pigsty B from above to the server 1 .

The sensor SE is a thermometer, hygrometer, or the like that measures the temperature, humidity, etc. of the pigsty B, and outputs measurement data of the temperature, humidity, etc. in the pigsty B to the server 1 .

In the server 1, the image analysis AI_Q identifies each of the objects OP1 to OPn included as subjects in the moving image D based on the moving image D acquired from the camera CA. Individuals from P1 to Pn are recognized.
The objects OP1 to OPn are areas within the frame image represented by contour lines or the like indicating the outline of the individual pig.
The server 1 further analyzes the behavioral patterns (habits, etc.) of the pigs P1 to Pn identified as individuals, based on the objects OP1 to OPn indicating the plurality of pigs P1 to Pn in the video D, and analyzes the behavioral patterns (habits, etc.) of the pigs P1 to Pn, and etc. to determine the health status of each pig P1-Pn.
In addition, the server 1 can identify elements (causes) that are the basis of actions such as the sociability of the pigs P1 to Pn, the interaction between the pigs P1 to Pn existing in the same action range, the relationship between the pigs P1 to Pn, and the like. to detect
Details of the functional configuration and processing of the server 1 will be described later with reference to FIG.

FIG. 6 is a block diagram showing an example of the hardware configuration of a server in the information processing system of FIG.

The server 1 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input/output interface 15, an output section 16, and an input section 17. , a storage unit 18 , a communication unit 19 and a drive 20 .

The CPU 11 executes various processes according to programs recorded in the ROM 12 or programs loaded from the storage unit 18 to the RAM 13 .
The RAM 13 also stores data necessary for the CPU 11 to execute various processes.

The CPU 11 , ROM 12 and RAM 13 are interconnected via a bus 14 . An input/output interface 15 is also connected to this bus 14 . An output unit 16 , an input unit 17 , a storage unit 18 , a communication unit 19 and a drive 20 are connected to the input/output interface 15 .
The output unit 16 includes a display, a speaker, and the like, and outputs images and sounds.
The input unit 17 is composed of a keyboard, a pig, or the like, and inputs various information according to user's instruction operation.
The storage unit 18 is configured by a hard disk or the like, and stores data of various kinds of information.

The communication unit 19 controls communication with another communication target (for example, the camera CA in FIG. 1) via the network N. FIG.
A removable medium 31 such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory is mounted in the drive 20 as appropriate. A program read from the removable medium 31 by the drive 20 is installed in the storage unit 18 as necessary. The removable medium 31 can also store various data stored in the storage unit 18 in the same manner as the storage unit 18 .

The administrator terminal 2 also basically has the hardware configuration shown in FIG. Therefore, description of the hardware configuration of the administrator terminal 2 is omitted. In addition, when the administrator terminal 2 is composed of a smart phone or a tablet terminal, the display units of the input unit 17 and the output unit 16 are composed of touch panels.

FIG. 7 is a functional block diagram showing an example of the functional configuration of the server of the information processing system of FIG.

The storage unit 18 of the server 1 shown in FIG. 6 stores image analysis AI_Q and recommendation AI_R.

Image analysis AI_Q has a learning unit, a model, and a recognition unit as described above. The model of this embodiment is a model in which, when a new moving image D is input, objects OP1 to OPn indicating N pigs P1 to Pn included in the moving image are labeled and output.
However, this model is merely an example, and a model that executes at least a part of a predetermined process described later and outputs the execution result may be adopted. Further, although the number of model types is one for convenience of explanation, a plurality of types of models may be provided in the image AI_Q.

Specifically, the image analysis AI_Q inputs the input moving image D into the model for each unit image (frame image), and labels the objects OP1 to OPn representing the N pigs P1 to Pn included in the moving image. output.
The form of output is not particularly limited, and may be a form of outputting as information separate from the moving image D, or a frame indicating the objects OP1 to OPn for each unit image (frame image) constituting the moving image D. A form in which lines and labeling are superimposed and output may be used.

As described above, the recommendation AI_R has a learning unit, a model, and a recognition unit, similar to the image analysis AI_Q. In the model, when inputting information on weight gain in units of pigsty B (pighouse unit to be described later), such as the average weight of pigs P1 to Pn in pigsty B, and environmental information on the temperature and humidity of pigsty B, each It outputs the mixing ratio of the feed.

As shown in FIG. 7, in the server 1, when the CPU 11 executes processing, the video acquisition unit 101, the environment information acquisition unit 102, the type analysis unit 103, the processing execution unit 104, the display control unit 105, etc. function.

The moving image acquisition unit 101 acquires a moving image D (video) obtained as a result of capturing the activity of one or more animals such as pigs P1 to Pn in the pigsty B, and inputs it to the image analysis AI_Q.
The output of the image analysis AI_Q (model) obtained as a result of inputting the moving image D (video), for example, information indicating each of the pig objects OP1 to OPn here, is provided to the type analysis unit 103 .

The environment information acquisition unit 102 acquires measurement data output from the sensor SE (prediction information such as a weather forecast when a predetermined process described later is a process of predicting the future) as environment information, 104.

The type analysis unit 103 performs a plurality of types of analysis for each of the pig objects OP1 to OPn as one unit. Such one unit is hereinafter referred to as "one pig unit".
A plurality of types of analysis are not particularly limited. etc. have been adopted.
A predetermined type of analysis result for each pig by the type analysis unit 103 (for example, the weight of the pig Pk, etc.) is hereinafter referred to as "per-pig analysis information".
In this embodiment, pig-by-pig analysis information for each of the n pigs P1 to Pn, that is, a total of n pig-by-pig analysis information is obtained for each predetermined type. For example, if the predetermined type is body weight, body weight data for each of n pigs P1 to Pn, that is, a total of n body weight data are obtained.

For example, when the predetermined type is analysis of body weight gain, the type analysis unit 103 analyzes one or more of body weight, body length, muscle mass, etc. for each of the pigs P1 to Pn corresponding to each of the objects OP1 to OPn. Analyze information about the body.
Specifically, the type analysis unit 101 calculates the weight, length, and muscle mass of each of the pigs P1 to Pn based on the area of the outline of each of the objects OP1 to OPn and the weight, length, and muscle mass per preset unit area. Estimate muscle mass, etc. That is, the estimation results for each of the pigs P1 to Pn, that is, the n estimation results in total are an example of the pig-by-pig analysis information.
Here, for example, body weight, body length, muscle mass, and the like have been described as information related to body gain analyzed by the type analysis unit 103, but in addition, bone structure, posture, movement, color, etc. can also be analyzed and measured. can be done.

In addition, the type analysis unit 103 executes analysis of the number of pigs bred, behavior pattern analysis, death/disease analysis, and the like.
For example, when the predetermined type is analysis of the number of pigs bred, the type analysis unit 103 counts the number of bred pigs (individual number) in the target range (per piggery, per fence, etc.) based on the objects OP1 to OPn. .

For example, when the predetermined type is behavior pattern analysis, the type analysis unit 103 determines that if any of one or more conditions set in advance by the transition of the positions of the objects OP1 to OPn is satisfied, the condition is satisfied. Determine the behavior pattern of the corresponding pig.
Specifically, the type analysis unit 103 determines whether or not the transition of the center coordinates of each of the objects OP1 to OPn matches a preset action condition, and determines whether or not the behavior label of the animal corresponding to the matching action condition is determined. (“01” for excretion behavior, “02” for sleeping behavior, “03” for eating behavior, “04” for drinking behavior, etc.) are represented by the objects OP1 to OPn or their original image frames. time information (timestamp) or data frame.

As a result, for example, in an environment where 50 pigs P1 to Pn are housed and active in the pigsty B, the behavior of each pig can be determined when it excreted, when it slept, when it ate food, when it It is possible to grasp detailed actions such as whether or not the user drank water.

For example, if the predetermined type is death/sickness analysis, the type analysis unit 103 analyzes whether the pig is dead or sick based on the objects OP1 to OPn. Specifically, the type analysis unit 103 determines that the pig is dead if the position of the object does not move at all within a certain period of time.
Further, the type analysis unit 103 determines that the movement of the object is a sick behavior when the movement of the object has a movement pattern different from that of the normal movement.

For example, when the predetermined type is detection of a person, the type analysis unit 103 analyzes the presence or absence of objects other than pigs (for example, humans and other animals) based on the objects OP1 to OPn, and detects objects other than pigs. If so, output an alert.
In this way, the pig objects OP1 to OPn and other objects are detected separately, and when an object other than the pig object is detected, it is detected that there is a person or the like, and an alert is output. At this time, for example, by setting a nighttime period (for example, from 21:00 to 7:00 the next morning) as a detection condition, pig thieves, feed thieves, etc. that visit at night can be detected in real time.
Note that the detection conditions shown here are only examples, and other conditions may be used. In addition, by positively putting up a notice saying that the pigsty is being monitored by the camera CA outside the pigsty, it becomes a crime prevention measure against the theft of pigs.

The processing execution unit 104 uses at least a part of each of the plurality of types of analysis information for a total of n pieces of pig-by-pig output from the type analysis unit 103 and the environment information output from the environment information acquisition unit 102 as input parameters. , executes the predetermined process using the input parameter.
The predetermined processing is not particularly limited, but in the example of FIG. 7, the predetermined processing includes feed blending recommendation processing, automatic document processing, body weight gain/meat quality prediction processing, gene/genome analysis processing, sales support processing, and alert/report processing. is adopted.
When the processing execution unit 104 executes the material mixture recommendation process, the feed mixture recommendation unit 110 functions. When the automatic documenting process is executed, the automatic documenting section 111 functions. The body gain/meat quality part 112 functions when the body gain/meat quality prediction process is executed. When gene/genome analysis processing is executed, the gene/genome analysis unit 113 functions. When the sales support process is executed, the sales support unit 114 functions. The alert/report section 115 functions when alert/report processing is executed.

The material combination recommendation unit 110 executes the following process as the material combination recommendation process.
That is, the material combination recommendation unit 110 determines the unit of the pigsty B (first unit ) (hereinafter referred to as “per piggery analysis information”). For example, in the present embodiment, when the total n pieces of analysis information per pig are the weights of the pigs P1 to Pn, the average weight of the pigs P1 to P, etc. are generated as the analysis information per pigsty.
The method of generating (computing) the analysis information for each pig is not particularly limited, that is, the method of obtaining the average value of the weights of n pigs is merely a presentation. (m is an integer value equal to or smaller than n) Any method can be adopted, such as a method of obtaining an average value of body weights of pigs or a method of obtaining a median value other than the average value.
Further, the material combination recommendation unit 110 acquires environmental information indicating the temperature and humidity of the pigsty B (measured data from the sensor E, or predicted data obtained from weather usage, etc.).
Then, the material combination recommendation unit 110 uses the pigsty unit analysis information (average weight of n pigs P1 to Pn, etc.) and the environment information indicating the temperature and humidity of the pigsty B as input parameters to make a recommendation AI_R (model). input.
The material mixture recommendation unit 110 outputs the mixture ratio of various feeds per 100 kg of feed and the amount of feed per pigsty B output from the recommendation AI_R (model) to the display control unit 105 as recommendation information.

The automatic document creation unit 111 executes the following processing as automatic document creation processing.
That is, the automatic reporting unit 111 inputs the number of pigs (the number of individuals) in the target range (per piggery, per fence, etc.) counted by the type analysis unit 103 and the growth situation into a preset management report, Automate input management. This eliminates the need to fill in paper forms or manually enter data on a tablet.

The body gain/meat quality section 112 executes the following process as the body gain/meat quality prediction process.
That is, the weight gain/flesh quality part 112 determines what the weight/length is in the middle of growth (current state) at the time of shipment, based on the weight gain information (current weight/length, etc.) generated by the type analysis unit 103. predict what will happen. The weight gain/meat quality unit 112 adds information such as pig breeds, breeding environment, and feeding method to the weight gain information. sales forecast becomes possible.

The gene/genome analysis unit 113 executes the following processing as gene/genome analysis.
That is, the gene/genome analysis unit 113 creates a database of the individual identified from the moving image of the camera CA, the predicted body weight of the individual, and the kinship relationship of the mother pig and the father pig at the gene level, and based on the database, the body shape and fatness. By crossbreeding healthy pigs and pigs that are less likely to get sick, we can produce high-quality pigs.

The sales support unit 114 executes the following processing as sales support processing.
In other words, the sales support unit 114 predicts the number of days it will take to reach a shippable weight based on the individual identified from the moving image of the camera CA and the predicted weight of the individual. Then, the sales support unit 114 determines the predicted shipping price based on the timing of the shipping prediction and the daily market price.
As a result, based on the price when shipping pigs and the market price at that time, for example, it is better to ship 10 days earlier, or even if it is 10 more days of feeding, it is better to extend the shipment and make it a chief. Sales support can be provided in setting up

The alert/report unit 115 executes the alert/report process as follows.
That is, the alert/report unit 115 presents an alert or report on the screen 80 (see FIG. 8) of the administrator terminal 2 based on the results analyzed or measured by the above units, and notifies the administrator or staff.
Specific examples of alerts are listed below.
Periodically notify the number of pigs in the entire piggery or block unit (fence unit).
Regular notification of weight gain. Regular notification of various weight coefficients. Notify the overall average, median, maximum and minimum weights once a day.
At the timing when the difference diverges by a certain value or more, the median value and the minimum divergence are notified.
Either the slow growing pig or the fast growing pig is detected and notified at the timing when the difference between the slow growing pig and the fast growing pig diverges by a certain value or more. It also informs the median and the maximum divergence.
At the timing when the pig reaches a preset weight suitable for shipping, the pig that has reached that weight is detected and notified. Detects and notifies pigs that have reached a preset weight.
When the weight approaches a preset weight suitable for shipping, it estimates how much carcass can be taken, and notifies the amount of the carcass.
Similarly, at the timing when the weight is close to that suitable for shipping, the amount of fat attached is estimated, and the percentage of fat is expressed as a percentage (%) and notified.
Similarly, when the pigs reach a weight suitable for shipping, the pigs are ranked in 5 levels of best, high, medium, medium, and others, and the number of pigs of each rank is estimated and notified.
Similarly, at the timing when the weight is close to suitable for shipping, the expected sales are displayed along with the market price for each rank. Estimated sales can be grasped in advance by multiplying the estimated number of animals for each rank by the market price.
Periodically notify the number of animals per block and the amount of feed that is optimal for the environment.
Notification of the optimum amount of feed calculated based on environmental information such as the number of animals, degree of growth, temperature and humidity.
Periodically notify the number of animals per block and the blending ratio of raw materials that are optimal for the environment.
Notification of the optimum mixing plan calculated based on environmental information such as the number of animals, degree of growth, temperature and humidity.
At predetermined times (morning, noon, and night), the current temperature and humidity in the pigsty are measured and notified.
Appropriate temperature and humidity in the pigsty are notified for each season at predetermined times (morning, noon, and night).
If it can be predicted that the block will exceed a certain temperature range, it will notify the temperature anomaly on that day.
If it can be predicted that the block will exceed a certain humidity range, it will notify the abnormal value of the humidity for that day.
Notify block by block that pigs are too dense when it can be predicted that they will exceed a certain density range.
In the morning of the day, we will notify you of the previous day's situation regarding the difference between the minimum temperature and maximum temperature (temperature difference).
The prediction of the lowest temperature and the highest temperature (predicted temperature difference) will be notified on the day of the next day's predicted temperature difference.
The previous day's situation regarding the difference (temperature difference) between the outside air and the indoor temperature is notified in the morning of the day.
The prediction of the outside air and indoor temperature (predicted temperature difference) will be notified on the day of the next day's predicted temperature difference.
At predetermined times (3 times in the morning, noon, and night), the measured carbon dioxide concentration is notified.
At the timing when the density (degree) of the fence measured by image analysis becomes higher than the preset allowable density (threshold value), "The density is high and it is better to disperse the pigs. is notified by outputting an alert message to that effect.
The ammonia concentration measured by the sensor SE is notified at predetermined times (three times in the morning, afternoon, and night).
At the timing when the degree of ammonia concentration measured by the sensor SE reaches a preset threshold, an alert message such as "The density is high and it is better to disperse the pigs." .
As a result of the image analysis, if a possible death flag is set, an alert message such as "A pig with a high possibility of being dead has been detected" is notified.
When a condition is met where there is a high possibility that a weak pig is present (when the pig is about to die), it is notified by outputting an alert message to the effect that the risk of death of the pig is increasing.
At the timing when body weight growth has clearly slowed down, the status of body weight growth is analyzed, and signs of disease are identified as follows: "If body weight growth slows down, you are not taking in enough food, so there is a possibility of illness linked to a decrease in appetite." ” message.
When signs of various diseases are detected as a result of analyzing moving images captured by the camera CA, the signs of various diseases are notified.
If the video cannot be obtained, a request for investigation of the operation status of the camera CA is notified.
If the environment information has not been obtained, a request for investigation of the operation status of the sensor SE is sent.
When the communication is cut off, a notification requesting investigation of the state of the network N is made.
If a device such as a camera CA or a sensor SE cannot be remotely accessed, a request is made to investigate whether the device itself is powered on.
When it is determined that there is an adhering matter from the image of the camera CA, an alert message to the effect that the lens of the camera CA is dirty and cannot be photographed normally is presented on the manager terminal 2 and transmitted to the manager or a staff member. In addition, a message indicating how many days have passed since a device such as camera CA or sensor SE was installed or maintained is notified.

The display control unit 105 executes display control for presenting the results of various processes output from the processing execution unit 104 to the administrator terminal 2 .
Specifically, the display control unit 105 presents to the administrator terminal 2 the mixing ratio of various feeds per 100 kg of feed output from the feed mixing recommendation unit 110 of the processing execution unit 104 and the amount of feed per pigsty B. display control.

As described above, according to the functional configuration of the server 1 in the information processing system of the embodiment, the image analysis AI_Q analyzes the moving image D of the pigsty B to recognize the objects OP1 to OPn representing the N pigs P1 to Pn. Then, the type analysis unit 103 analyzes the livestock object for each pig, and generates the weight of each of the 50 pigs for each of the N pig objects OP1 to OPn. Then, the processing execution unit 104 uses the average weight of the pigs in the unit of pigsty B and the measurement data such as temperature and humidity as input parameters to execute calculation processing of, for example, the mixing ratio of feed per 100 kg, and the administrator terminal 2 Therefore, the administrator or staff who browses the feed mixing ratio on the screen of the administrator terminal 2 can mix the feed at the required feed mixing ratio at that time without going around to the piggery B and feed the piggery. It is possible to raise pigs without relying on human experience and intuition.

Here, a screen displayed on the administrator terminal will be described with reference to FIG.
8 is a diagram showing an example of a screen of the administrator terminal of FIG. 1. FIG.
As shown in FIG. 8, the screen 80 of the administrator terminal 2 is a web page disclosed by the server 1 and can be viewed by accessing from the administrator terminal 2 or an application program (hereinafter referred to as an application program) installed in the administrator terminal 2. (referred to as the “App”).
On the screen 80, if the pigsty B is divided by a plurality of fences, an area 81 in which a moving image D of one fence (for example, A block-1) is displayed, and a An area 82 displaying the weight range of the pigs and the number of pigs belonging to that range, an area 83 displaying the latest weight measurement results and statistical weights one week ago, weather, room temperature, and humidity. is displayed, and an area 85 is arranged in which caution comments preset according to the environmental information measured by the sensor SE are displayed.

On this screen 80, the information displayed in each of the areas 81 to 85 described above can be viewed by the manager or the staff, so that the breeding environment and growth status of the pigs P1 to Pn in the piggery B can be viewed without visiting the piggery B. , health status, etc. can be grasped and managed.
For example, by viewing the area 81, a manager or a staff member can view the status of pigs raised in the A block-1 in real time.
Also, in the area 81, when the manager or staff member wants to browse the moving image D of another block, the other block can be browsed by operating a switching button or a pull-down menu (not shown).

In the area 82, the weight range of pigs is divided, for example, in units of 5 kg, and the manager or staff can confirm at a glance how many pigs are in which range, and the number of pigs in the block can be confirmed. You can see the weight gain balance at a glance.

In area 83, the latest measurement results, such as the average weight, minimum weight, and maximum weight of the pigs in the block, are displayed. In addition, the average weight, minimum weight, and maximum weight of the pigs in the block, for example, are displayed as statistical weights of the previous week. This allows the manager or staff to determine how much the pigs have increased between one week ago and now.

Area 84 displays weather marks (sunny, rainy, cloudy, etc.), temperature (outside air temperature), room temperature (temperature inside the pigsty), and humidity (humidity inside the pigsty). and staff can judge the current status of the pig breeding environment at a glance.

In the area 85, when the environmental information satisfies the preset alerting conditions, a message is displayed to alert the manager or staff. The environment of the pigsty B can be improved before the breeding environment deteriorates without relying on the experience of the above.

According to the pigsty monitoring function of the server 1 and the screen 80 of the administrator terminal 2 displaying the monitoring result, the following effects can be obtained.
The weight and length of multiple pigs measured simultaneously in the same fence can be checked in real time.
In this embodiment, the camera CA can estimate and measure the weight, length, and the like of a maximum of 50 pigs at the same time. In addition, since the weight, length, etc. of the pigs are constantly measured from the moving images captured by the camera CA, it is possible to confirm changes in daily gains and the like on graphs.
Until now, this work was done manually for each pig, so there were problems such as taking time to measure weight and length, it was difficult to manage weight gain uniformly, and weight variation at the time of shipment. However, in the present embodiment, weight and length measurements are automatically performed, so daily weight gain measurements can be finely managed. In addition, by putting pigs in fences according to their weight and raising them with feed according to their weight, the weight of the pigs at the time of shipment can be made uniform.
As a result, it is possible to drastically reduce the time and labor required to measure the weight and length of pigs. In addition, by eliminating individual differences in body weight gain as much as possible, it is possible to ship with ideal body weight.

In this embodiment, moving images can be captured by the camera CA to measure the dynamics, and by measuring the temperature, humidity, etc. of the pigsty with the sensor SE, it is possible to analyze the effects on the pigsty environment and dynamics.
By measuring the temperature and humidity inside the piggery for 24 hours and presenting it on the screen 80 of the administrator terminal 2, the manager or staff member who browses it can quickly respond to changes in the environment inside the piggery. can always be kept at the right temperature and humidity. In addition, by analyzing the impact of environmental changes on pigs as the ecology (dynamics) of pig movements, it is possible to grasp abnormalities before the physical condition of pigs deteriorates.
Until now, there have been problems such as checking the humidity and temperature manually due to manual measurement at the piggery, difficulty in understanding the effect of humidity and temperature on pigs, and delays in responding to problems such as affecting the pig's body. In this embodiment, in addition to monitoring by the camera CA, the temperature and humidity of the pigsty can be constantly measured by the sensor SE and automatically totaled. Also, if the temperature or humidity inside the pigsty is not suitable for pigs, an alert is issued so that the manager or staff can quickly grasp the situation. In addition, abnormalities in pigs can be detected from the dynamics of pigs.
As a result, by managing the humidity and temperature in the pigsty, which is indispensable for managing the physical condition of pigs, by using the sensor SE and the server 1, it is possible to quickly deal with the abnormality when an abnormality occurs in the dynamics of the pigs.

In this embodiment, pig death alerts are provided for each block of the piggery.
By catching and notifying dead pigs on a block-by-block basis, it is possible to implement triage measures for the hygiene and epidemic prevention of the piggery.
In the past, there were problems such as not knowing whether the pig was sleeping or dead, there was a delay even though it was desired to dispose of the carcass quickly, and there was a bad influence on the surrounding pigs. By automatically identifying, detecting and outputting alerts for pigs that are doing so, it is possible to deal with them before they have a negative impact on the surroundings.
As a result, the death of pigs is monitored and supplemented by camera CA and sensor SE instead of human visual observation, so that countermeasures can be taken before other pigs in the piggery are affected. lead to decline.

Next, the processing executed by the server 1 will be described with reference to FIG. FIG. 9 is a flow chart illustrating an example of the flow of processing executed by the server 1 having the functional configuration of FIG.
In the information processing system of the embodiment, the server 1 inputs the moving image D of the pigsty B captured by the camera CA and the measurement data such as the temperature and humidity of the pigsty B measured by the sensor SE to the server 1. , identify each individual of one or more pigs P1 to Pn, analyze the types of objects OP1 to OPn representing the respective pigs P1 to Pn, and execute predetermined processing according to the types as follows.

In step S101, the moving image acquisition unit 101 acquires a moving image configured by arranging a plurality of unit images in the time direction, obtained as a result of imaging one or more pigs in action in the pigsty.

In step S102, the image analysis AI_Q analyzes the moving image D of the pigsty B, thereby recognizing the objects OP1 to OPn representing the N pigs P1 to Pn included in the moving image D of the pigsty B.

In step S103, the type analysis unit 103 analyzes the livestock object for each pig, and obtains a predetermined type of information (for example, body weight, muscle mass, presence or absence of death/disease of each of the 50 pigs, etc.) for each pig. The second unit analysis information is generated and output for each of the N pig objects OP1 to OPn.

In step S104, the processing execution unit 104 generates the first unit (the unit of the pigsty B) generated based on the N second unit analysis information (analysis results for each of the 50 pigs) generated by the type analysis unit 103. A predetermined type of information (average weight of 50 pigs, etc.), and one or more of multiple types of first unit information and multiple types of environmental information (temperature and humidity measurement data) as first unit analysis information is used as an input parameter to execute a predetermined process (recommendation process of feed mixing ratio, automatic documenting process, gene/genome analysis process, sales support process, alert/report process, etc.) using the input parameter.
As an example, for example, the feed mixture recommendation unit 110 of the processing execution unit 104 generates the average weight of the 50 pigs based on the analysis results of each of the 50 pigs generated by the type analysis unit 103, and the weight at that time. The measurement data of the temperature and humidity of the pigsty B are input to the recommendation AI_R as input parameters, and the recommendation AI_R is made to perform the arithmetic processing of the mixing ratio per 100 kg of feed using the input parameters, and the output from the recommendation AI_R. It outputs to the display control unit 105 the mixing ratio per 100 kg of feed.

In step S105, the display control unit 105 sends the processing result information (information about the feed mixture for 50 pigs (mixture ratio of multiple types of feed per 100 kg of feed) output from the processing execution unit 104 to the administrator. Output to the screen of terminal 2.

In this way, according to the operation of the server 1, the administrator or staff can calculate the mixing ratio of the feed to be given to the pigs P1 to Pn in the pigsty B on the day without recording the changes in body weight, physical condition, etc. of each pig. is known, the feed for the entire pigsty B can be prepared at that mixing ratio and fed to the pigs P1 to Pn. As a result, working efficiency can be improved.

Here, an example in which the moving image analysis technology of the information processing system of the present embodiment is applied to meat processing will be described with reference to FIG. 10 .
FIG. 10 is a diagram showing how a moving image of a pig is acquired before and after processing.
In addition to the camera CA installed in the piggery B shown in Fig. 1, a camera is installed in the meat processing factory, and the server 1 analyzes the two videos taken by the two cameras, thereby making new use of the analysis results. I can think of a way.

As shown in FIG. 10, a video D1 of a pig (pig before processing) when shipped from the piggery B, a video D2 of a pig after processing at a meat processing plant (pig after processing), and analysis by image analysis AI_Q Then, the information on the pigs before and after processing is made to correspond to each other.
For example, an individual pig imaged before processing and an individual processed meat imaged after processing are managed in association with each other. At this time, the feature information of each part of each individual is associated with each other.

For example, the state of the back of a pig before processing is associated with the fat thickness of processed meat. Also, the characteristics of the outer shell of the pig before processing are associated with the state of the carcass of the processed meat. Furthermore, the foreign substances (protruding portions) attached to the pig before processing are associated with the content of the same foreign substances (for example, fat, etc.) in the processed meat.
By analyzing the information of the pig before and after processing by the image analysis AI_Q, it is possible to estimate the fat thickness and the like of the processed meat from the appearance image of the pig before processing. Also, estimated fat and meat conditions can be estimated. Furthermore, it is possible to determine what kind of foreign matter is attached to the pig before processing.

By providing the results of AI analysis to the meat processing plant, it is possible to compare various estimated measurements and actual measurements to verify whether the data match each other.
In addition, it is possible to carry out on-site foreign matter determination. Furthermore, it is possible to consider the most suitable usage contents according to the operation. Processed meat can be graded based on pre-processed pig information.

According to the above-described embodiment, it is possible to support a series of businesses from pig breeding to sales without relying on human experience and intuition.

The series of processes described above can be executed by hardware or by software.
In other words, the functional configuration of FIG. 7 is merely an example and is not particularly limited.
That is, it is sufficient that the information processing system has a function capable of executing the above-described series of processes as a whole, and what kind of functional blocks and databases are used to realize this function are particularly limited to the example of FIG. not.
Also, the locations of the functional blocks and the database are not particularly limited to those shown in FIG. 7, and may be arbitrary. For example, the functional blocks and database of the server 1 may be transferred to the administrator terminal 2, camera CA, sensor SE, and the like. Furthermore, camera CA and sensor SE may be the same hardware.

Further, for example, when a series of processes is executed by software, a program constituting the software is installed in a computer or the like from a network or a recording medium.
The computer may be a computer built into dedicated hardware. Also, the computer may be a computer capable of executing various functions by installing various programs, such as a server, a general-purpose smart phone, or a personal computer.

Further, for example, a recording medium containing such a program is not only configured by a removable medium (not shown) that is distributed separately from the device main body in order to provide the program to the user, but also is pre-installed in the device main body. It consists of a recording medium or the like provided to the user in a state.

In this specification, the steps of writing a program recorded on a recording medium are not only processes that are performed chronologically in that order, but also processes that are not necessarily chronologically processed, and that are performed in parallel or individually. It also includes the processing to be executed.
Further, in this specification, the term "system" means an overall device composed of a plurality of devices, a plurality of means, or the like.

In the above embodiment, the pigsty B is the predetermined place and the target of individual identification is the pig P. However, by expanding the data of the image analysis AI_Q and the recommendation AI_R, livestock such as cattle, sheep, and chickens can also be analyzed. can be done. Furthermore, various animals such as dogs, cats, monkeys, and humans can be targeted. That is, N (N is an integer value equal to or greater than 1) livestock managed in a predetermined place is treated as a first unit, and a predetermined process may be executed for the first unit.
In the above embodiment, one or more physical quantities related to the environment of the field are temperature and humidity data and are detected by the sensor SE, but other data such as carbon dioxide concentration may be used. It is sufficient for the detection device to detect one or more physical quantities relating to the field environment and output information including the detection result as field environment information.
In the above-described embodiment, an example in which one type analysis unit 103 analyzes a plurality of types of information has been described.
In the above embodiment, the average weight of 50 pigs P1 to Pn is shown as an example of the mass of the first unit generated based on the mass of N, but in addition to this, for example, the total weight may be used, or based on the deviation It may be a statistical value or the like.

In the above embodiment, one camera CA is installed in the pigsty B so that all the pigs Pk in the pigsty B are imaged.
Therefore, as shown in FIG. 11, a wire 92 may be stretched over the pillar 91 of the pigsty B, and a camera 93 may be installed on the wire 92 so as to be movable in a predetermined direction (for example, the horizontal direction W).
The camera 93 is provided with a lidar sensor (LiDAR sensor) and is capable of measuring the distance to the imaging target. Although the wire 92 is used in this example, a rail or the like may be used.
The server 1 generates a 3D model by image analysis AI_Q from a plurality of images acquired from different positions by the camera 93, measures the body length and width of the pig, and estimates the body weight from the measured body length and width.
According to this example, a camera 93 that moves on a wire 92 stretched on a pillar 91 of the pigsty B is installed, and from a plurality of images captured from different positions by the camera 93, all the images in the pigsty B are captured without blind spots. The weight of each pig Pk can be estimated.

Further, in the above example, the camera CA and the camera 93 are installed in the pigsty B, but in addition to this, for example, if the administrator terminal 2 is a smartphone equipped with a lidar camera (a function to measure the distance to the subject), a dedicated By installing an application program (hereinafter referred to as an "app"), an administrator or a staff member uses a smartphone to image each pig Pk in pigsty B and display the weight of each pig Pk on the screen of the smartphone application. You may do so.

FIG. 12 shows that embodiment.
In this case, as shown in FIG. 12, when a staff member taps the camera icon 94 displayed on the top screen G1 of the smartphone application, the screen of the application transitions to the imaging screen G2.

On the imaging screen G2, a frame 95 indicating the imaging area and a message such as "Please tap the position on the ground" are displayed.
When the attendant taps the position of the ground (the position of the double circle), the distance to the pig including the ground is measured.
Then, the application transmits to the server 1 data including the measured distance data to the pig including the ground, and angle data and image data when the pig was photographed. Note that if there is other data (data such as imaging time, temperature, humidity, etc.), that data is also transmitted together.

The server 1 analyzes the received data using weight analysis AI, etc., and returns the estimated weight data of the analysis results to the smartphone.

On the smartphone, the estimated weight data received from the server 1 is displayed in the weight display frame 96 superimposed on the captured image of the pig on the screen G3 of the application. A text icon 97 is also displayed on the screen G3.

When the clerk taps the text icon 97, the application screen transitions to the next screen G4.
Since a text input frame 98 is displayed on the screen G4, the staff inputs the information (pig name, individual number, state of the pig, etc.) that he/she observed or noticed at that time as text, and taps the save button 99. Thus, the application transmits image data, weight data, text data, and time data of the pig to the server 1 .
In the server 1, each data received from the smartphone is associated with an identifier such as an individual number or name of the pig, and stored in the storage unit 18 as a management log, breeding record, or the like.

According to this example, by installing a distance measurement function and an application that are pre-installed in a smartphone, pigs can be captured with a simple operation of capturing an image of a pig with a smartphone without installing a camera in the piggery B. Head weight can be measured and managed. As a result, it is possible to realize a management function for raising pigs at low cost.

In the above-described embodiment, as the predetermined processes of the process execution unit 110, feed blending recommendation process, automatic form processing, weight gain/meat quality prediction process, gene/genome analysis process, sales support process, and alert/report process are adopted. However, in addition to this, various services can be provided by adopting the following processing.
For example, a sales agent service can be provided by recommending supplements, processed feeds, medicines, etc. according to the condition of pigs.
It is possible to predict the growth of pigs by AI or the like, and provide financial services in which the pigs themselves are movable properties based on the predicted growth data of the predicted pigs.
By using breeding information as evidence for insurance claims, an insurance referral service can be provided.
By predicting sales from meat quality prediction and predicting feed costs, it is possible to provide services that support the business management of pig farms.
It is possible to provide M&A intermediary services by understanding the business conditions of pig farms and calculating the value when selling the business.
In addition, it is possible to provide leasing of the equipment necessary for the business and financial services after grasping the financial condition of the pig farmers.
It is possible to realize a service that provides information on pigs raised by pig farms to meat manufacturers that purchase pigs.

In other words, the information processing apparatus to which the present invention is applied can take various embodiments having the following configurations.
That is, the information processing system of the present invention (for example, the information processing system in FIG. 5, etc.)
N (N is an integer value of 1 or more) livestock (for example, pigs P1 to Pn in FIG. 1) managed in a predetermined place (for example, pigsty B in FIG. 1) is a first unit (for example, in the above specification) In an information processing system that executes a predetermined process for the first unit as a pigsty unit),
an imaging device (e.g., camera CA in FIG. 4, etc.) that captures an image of the scene (e.g., pigsty B in FIG. 1) and outputs the captured image obtained as a result as a scene image (e.g., movie D in FIG. 1); ,
An output device (e.g., sensor SE in FIG. 1, or future prediction) that outputs information including one or more physical quantities (e.g., temperature, humidity, etc.) related to the environment of the field (e.g., pigsty B in FIG. 1) as field environment information is a device that acquires and outputs forecast information such as weather forecasts, etc.),
The predetermined processing is executed based on at least a part of the field image (for example, images of pigs P1 to Pn in pigsty B in FIG. 1) and the field environment information (for example, measured values such as temperature and humidity). an information processing device (for example, the server 1 in FIG. 7);
including
The information processing device (for example, the server 1 in FIG. 7, etc.)
By analyzing the field image, N objects representing the N livestock (for example, pigs P1 to Pn) included in the field image are converted to N livestock objects (for example, pig objects OP1 to OPn in FIG. 1). etc.), and a recognition means (for example, image analysis AI_Q in FIG. 7),
Using the livestock object as a second unit (for example, one pig unit as referred to in the above specification), a predetermined type of information (weight, height, muscle mass, presence or absence of death/illness, etc.) related to the second unit is analyzed as a second unit. Analysis means (for example, type analysis unit 103 etc.) and
The predetermined type of information (for example, if the second unit analysis information is body weight, the first unit (for example, the piggery unit referred to in the above specification) generated based on the N second unit analysis information, n pigs with one or more of the plurality of types of first unit information and the plurality of types of environmental information as input parameters (for example, the average weight of n pigs P1 to Pn and the pigsty B temperature and humidity), the predetermined processing using the input parameter (as a result of inputting the input parameter into the recommendation AI_R (model), the mixing ratio of various feeds per 100 kg of feed output from the recommendation AI_R (model) and the pigsty a predetermined process execution means (for example, the process execution unit 104 in FIG. 7, etc.) for executing a process of outputting the amount of feed per B unit as recommendation information;
Prepare.
In this way, by analyzing the field image, N objects representing N livestock (pigs P1 to Pn, etc. in the pigsty B) are converted into N livestock objects (for example, pig objects OP1 to OPn, etc. in FIG. 1). , and the livestock object as a second unit (one pig unit), and a predetermined type of information related to the second unit (weight, height, muscle mass, presence or absence of death/illness, etc.) as second unit analysis information, N livestock objects (for example, pig objects OP1 to OPn in FIG. 1), and a first unit (for example, the above-mentioned A predetermined type of information (for example, if the second unit analysis information is body weight, the average weight of n pigs) is used as the first unit analysis information, and multiple types of first unit information and One or more of a plurality of types of environmental information (e.g., data such as temperature and humidity of pigsty B) are used as input parameters (e.g., the average weight of n pigs P1 to Pn and the temperature and humidity of pigsty B). (As a result of inputting the input parameter into the recommendation AI_R (model), the mixing ratio of various feeds per 100 kg of feed output from the recommendation AI_R (model) and the amount of feed per pigsty B are output as recommendation information. By executing the process of raising and selling livestock, the staff does not need to manage information about each of the large number of livestock in the piggery, etc. Of all the lines of business, at least it can support the raising of livestock.

The output device (e.g., sensor SE in FIG. 1, etc.) outputs external environment parameters (e.g., temperature and humidity measurement data, etc.) related to the heat dissipation of the livestock (pigs P1 to Pn, etc. in pigsty B) in the field (e.g., pigsty B, etc.). ), and
The analysis means (for example, the type analysis unit 103 in FIG. 7, etc.) converts the mass (for example, weight, muscle mass, etc.) based on the muscles of the livestock to the N livestock objects (for example, FIG. 1) as the second unit analysis information. generated and output for each pig object OP1 to OPn, etc.),
The predetermined processing execution means (for example, the feed mixture recommendation unit 110 and recommendation AI_R in FIG. 7)
The mass of the first unit generated based on the mass of N (in this example, the average weight of 50 pigs P1 to Pn, etc.), and the external environment parameters (e.g., temperature and humidity measurement data, etc.) is input as the input parameter, and the processing of outputting information (mixing ratio of multiple types of feed in 100 kg of feed, etc.) regarding the feed composition of the first unit (for example, the piggery unit referred to in the above specification) (feed recommendation, etc. ) is executed as the predetermined process.
As a result, staff such as administrators and attendants can know the mixing ratio of feed given to pigs P1 to Pn in pigsty B on that day without having to record changes in body weight and physical condition of each pig. Since the feed for the entire pigsty B can be made at the mixing ratio and given to the pigs P1 to Pn, the feed given to the pigs P1 to Pn every day can be properly mixed without relying on human experience and intuition, and work efficiency can be improved. can be improved.

CA: camera, SE: sensor, Q: image analysis AI, R: recommendation AI, U: administrator, 1: server, 2: administrator terminal, 11. CPU 18 Storage unit 19 Communication unit 101 Moving image acquisition unit 102 Environment information acquisition unit 103 Type analysis unit 104 Process execution unit , 105... display control unit, 111... automatic reporting unit, 112... weight gain/meat quality prediction unit, 113... gene/genome analysis unit, 114... sales support unit, 115...・Alert/report section

Claims (5)

1. In an information processing system for executing a predetermined process on N (N is an integer value of 1 or more) livestock managed in a predetermined place as a first unit,
   an imaging device that captures the state of the field and outputs the captured image obtained as a result as a field image;
   an output device that detects one or more physical quantities related to the field environment and outputs information including the detection result as field environment information;
   an information processing device that executes the predetermined process based on at least part of the field image and the field environment information;
   including
   The information processing device is
   recognition means for recognizing, as N livestock objects, N objects representing each of the N livestock contained in the field image by analyzing the field image;
   analysis means for generating and outputting the livestock object as a second unit and a predetermined type of information relating to the second unit as second unit analysis information for each of the N livestock objects;
   Using the predetermined type of information of the first unit generated based on the N second unit analysis information as the first unit analysis information, one or more of the plurality of types of first unit information and the plurality of types of environment information a predetermined process executing means for executing the predetermined process using the input parameter as an input parameter;
   An information processing system comprising
2. The output device outputs, as the field environment information, an external environment parameter related to heat dissipation from the livestock in the field,
   The image analysis means generates and outputs, as the second unit analysis information, the mass based on the muscles of the livestock for each of the N livestock objects,
   The predetermined process execution means is
   The process of inputting the mass of the first unit generated based on the mass of N and the external environment parameter as the input parameters and outputting information on the feed composition of the first unit is performed by the predetermined run as a process,
   The information processing system according to claim 1.
3. the external environmental parameters include at least one of temperature and humidity;
   The information processing system according to claim 2.
4. An information processing system for executing a predetermined process on N (N is an integer value of 1 or more) livestock managed in a predetermined place as a first unit,
   an imaging device that captures the state of the field and outputs the captured image obtained as a result as a field image;
   an output device that detects one or more physical quantities related to the field environment and outputs information including the detection result as field environment information;
   an information processing device that executes the predetermined process based on at least part of the field image and the field environment information;
   In the information processing method executed by the information processing device of the information processing system comprising
   a recognition step of analyzing the field image to recognize N objects representing the N domestic animals included in the field image as N livestock objects;
   an analysis step of generating and outputting, for each of the N livestock objects, the livestock object as a second unit and a predetermined type of information relating to the second unit as second unit analysis information;
   Using the predetermined type of information of the first unit generated based on the N second unit analysis information as the first unit analysis information, one or more of the plurality of types of first unit information and the plurality of types of environment information a predetermined process execution step of executing the predetermined process using the input parameter as an input parameter;
   Information processing method including.
5. An information processing system for executing a predetermined process on N (N is an integer value of 1 or more) livestock managed in a predetermined place as a first unit,
   an imaging device that captures the state of the field and outputs the captured image obtained as a result as a field image;
   an output device that detects one or more physical quantities related to the field environment and outputs information including the detection result as field environment information;
   an information processing device that executes the predetermined process based on at least part of the field image and the field environment information;
   to a computer that controls the information processing device of an information processing system including
   a recognition step of analyzing the field image to recognize N objects representing the N domestic animals included in the field image as N livestock objects;
   an analysis step of generating and outputting, for each of the N livestock objects, the livestock object as a second unit and a predetermined type of information relating to the second unit as second unit analysis information;
   Using the predetermined type of information of the first unit generated based on the N second unit analysis information as the first unit analysis information, one or more of the plurality of types of first unit information and the plurality of types of environment information a predetermined process execution step of executing the predetermined process using the input parameter as an input parameter;
   A program that executes control processing including

Images