阅读说明：本技术 养猪场生猪状态预测系统 (Pig farm live pig state prediction system ) 是由 乔宏哲 顾卫杰 杨保华 于 2021-09-07 设计创作，主要内容包括：本发明属于养猪场监测技术领域,具体涉及一种养猪场生猪状态预测系统,本系统包括：生猪状态预测模型建立模块,根据养猪场样本生猪的样本数据建立生猪状态预测模型；以及生猪状态预测结果输出模块,与生猪状态预测模型建立模块电性连接,以将当前生猪数据代入生猪状态预测模型中得出生猪状态预测结果,本养猪场生猪状态预测系统根据养猪场生猪状态预测模型为养猪场生猪提供了可视化的正常度指数,做到了养猪场每一头生猪的正常度可视化,后续工作人员再对被测的非正常生猪做出判别,减轻了养殖人员的工作量并提供了数据支持,提高了养猪场的管理效率,做到了智能化、远程化、精准化、可视化养猪。(The invention belongs to the technical field of pig farm monitoring, and particularly relates to a pig farm live pig state prediction system, which comprises: the pig state prediction model establishing module is used for establishing a pig state prediction model according to sample data of live pigs of a pig farm sample; and live pig state prediction result output module, establish module electric connection with live pig state prediction model, with obtain live pig state prediction result in substituting live pig state prediction model with current live pig data, this pig farm live pig state prediction system provides visual normality index for the live pig of pig farm according to pig farm live pig state prediction model, it is visual to have done the normality of each live pig of pig farm, follow-up staff makes the judgement to the abnormal live pig that is surveyed again, the work load that has alleviateed the personnel of breeding and provided data support, the management efficiency of pig farm has been improved, it is intelligent to have done, it is remote, it is accurate, visual pig raising.)

1. A pig farm live pig state prediction system is characterized by comprising:

the pig state prediction model establishing module is used for establishing a pig state prediction model according to sample data of live pigs of a pig farm sample; and

and the live pig state prediction result output module is electrically connected with the live pig state prediction model establishing module so as to substitute the current live pig data into the live pig state prediction model to obtain a live pig state prediction result.

2. The pig farm pig status prediction system according to claim 1,

the pig state prediction model building module comprises: the data acquisition unit and the live pig state prediction model unit are electrically connected with the data acquisition unit; wherein

The data acquisition unit acquires sample data of a live pig of a pig farm sample and sends the sample data to the live pig state prediction model unit to establish a live pig state prediction model.

3. The pig farm pig status prediction system according to claim 2,

the data acquisition unit comprises an infrared sensor module, a weighing sensor module and a camera module so as to detect the body temperature data, the weight data, the feeding data and the live pig state of the live pig; and is

The eating data is obtained according to the change of weight data before and after eating; and

the live pig state comprises normal live pigs and abnormal live pigs.

4. The pig farm pig status prediction system according to claim 3,

the pig state prediction model unit is used for establishing a pig state prediction model, namely

According to the sample data, a data vector and a transformation vector are set through fisher judgment so as to obtain the optimal projection direction of the sample data projected to a one-dimensional space through training;

and according to the optimal projection direction, projecting the sample data on the optimal projection direction to obtain data, and solving the mean value and standard deviation of the data types of the normal live pig and the abnormal live pig to obtain a live pig state prediction model and a live pig state prediction result.

5. The pig farm pig status prediction system of claim 4,

according to the sample data, a data vector and a transformation vector are set through fisher judgment so as to obtain the optimal projection direction of the sample data projected to a one-dimensional space through training; namely, it is

Establish data vector x ═ x⁽¹⁾，x⁽²⁾) Wherein

x⁽¹⁾Body temperature variance;

x⁽²⁾is the ratio of food intake to body weight;

defining a transformation vector w ═ (w)⁽¹⁾，w⁽²⁾) Wherein w is⁽¹⁾Is x⁽¹⁾Coefficient vector of (d), w⁽²⁾Is x⁽²⁾A coefficient vector of (a);

finding the best transformation vector w^*＝S_w ^-1(m₂-m₁) Wherein the maximum value w of w^*The optimal transformation vector is the optimal projection direction obtained by training; s_wIs an intra-class covariance matrix; m is₁And m₂Respectively representing the mean vector of normal pigs and abnormal pigs.

6. The pig farm pig status prediction system according to claim 5,

projecting the sample data on the optimal projection direction to obtain data, and obtaining the mean value and standard deviation of the data types of the normal live pig and the abnormal live pig, namely

μ₁＝w^*m₁；

μ₂＝w^*m₂；

Wherein

μ₁Data for normal live pigs are in w^*Mean, mu, of data obtained after vector axis projection₂Data for abnormal live pigs at w^*Mean, delta, of data obtained after vector axis projection₁Data for normal live pigs are in w^*Standard deviation, delta, of data obtained after vector axis projection₂Data for abnormal live pigs at w^*Standard deviation, N, of data obtained after vector axis projection₁Total number of data samples for normal live pigs, N₂Total number of data samples for abnormal live pigs.

7. The pig farm pig status prediction system of claim 6,

the live pig state prediction result output module is suitable for setting the current live pig data as a data vector x_c，Zc＝w^*x_cAnd Zc is a data vector x_cAt w^*The value obtained after vector axis projection;

the normality index D of the current live pig is as follows:

when Zc is less than or equal to mu₁+δ₁If so, judging the pigs as normal pigs, wherein the normality index D is 0;

when Zc is not less than mu₂-δ₂If so, judging the pigs to be abnormal if the normality index D is 1;

when mu is₁+δ₁≤Zc≤μ₂-δ₂Then, a live pig state prediction model is established, namely

Determining the pig is suspected to be abnormal；

The normality index D is a number ranging from 0 to 1, and closer to 0 indicates higher normality, and closer to 1 indicates lower normality.

8. The pig farm pig status prediction system of claim 7,

the live pig state prediction result output module is also connected with the server through a wireless module; wherein

The live pig state prediction result output module is suitable for manually setting a threshold value D₀When the current normality index of the live pig is larger than a preset threshold value D₀Judging the pigs to be abnormal and shooting live pig videos with set duration; or when the normality index is lower than a set threshold D₀Judging the live pig to be normal and shooting a live pig image; and sending the live pig video and the live pig image data to a server for storage.

Technical Field

The invention belongs to the technical field of pig farm monitoring, and particularly relates to a pig farm live pig state prediction system.

Background

The pig farm monitoring system is the inevitable trend of large-scale pig raising in the future, and suspected sick pigs can be determined in time and treated in time by observing the behaviors of the pigs according to the specific life habits of healthy live pigs.

The existing pig farm monitoring system mainly stays in the stage of collecting and storing data, most of the existing pig farm monitoring systems do not carry out deeper analysis and mining on the data, the health degree of live pigs cannot be subjected to data visual management, and manual analysis not only consumes a large amount of manpower and material resources, but also has the defect of strong subjectivity.

Disclosure of Invention

The invention provides a pig farm live pig state prediction system.

In order to solve the technical problem, the invention provides a method for predicting the pig state in a pig farm, which comprises the following steps: the pig state prediction model establishing module is used for establishing a pig state prediction model according to sample data of live pigs of a pig farm sample; and the live pig state prediction result output module is electrically connected with the live pig state prediction model establishing module so as to substitute the current live pig data into the live pig state prediction model to obtain a live pig state prediction result.

The method has the advantages that the visualized normality index is provided for the live pigs in the pig farm according to the established pig farm live pig state prediction model, the normality visualization of each live pig in the pig farm is realized, the follow-up workers judge the tested abnormal live pigs, the workload of the breeding workers is reduced, data support is provided, the management efficiency of the pig farm is improved, and the intelligent, remote, precise and visualized pig raising is realized.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.

FIG. 1 is a flow chart of a pig farm live pig status prediction system of the present invention;

FIG. 2 is a schematic diagram of the operation of the pig farm live pig status prediction system of the present invention;

FIG. 3 is a schematic block diagram of the pig farm live pig status prediction system of the present invention.

Detailed Description

To make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

As shown in fig. 1, the present embodiment provides a pig state prediction system for a pig farm, including: the pig state prediction model establishing module is used for establishing a pig state prediction model according to sample data of live pigs of a pig farm sample; and the live pig state prediction result output module is electrically connected with the live pig state prediction model establishing module so as to substitute the current live pig data into the live pig state prediction model to obtain a live pig state prediction result.

In this embodiment, specifically, the pig state prediction model establishing module includes: the data acquisition unit and the live pig state prediction model unit are electrically connected with the data acquisition unit; the data acquisition unit acquires sample data of a live pig of a pig farm sample and sends the sample data to the live pig state prediction model unit to establish a live pig state prediction model.

As shown in fig. 3, specifically, the data acquisition unit includes an infrared sensor module, a weighing sensor module and a camera module to detect body temperature data, weight data, feeding data and a live pig state of a live pig; and the eating data is obtained according to the change of the weight data before and after eating; and the live pig status comprises normal live pigs and non-normal live pigs.

Specifically, the habit of the live pig is extremely fixed, the body temperature and the food intake are the main observation indexes of the live pig, and the conditions such as oestrus stage and bad physical state of the live pig can be visually expressed; the normal body temperature of the live pig is 38-39.5 ℃, the food intake is the actual body weight multiplied by the coefficient, the coefficient is 0.05 in the piglet stage, the coefficient is 0.04 in the middle pig stage, and the coefficient is 0.03 in the big pig stage.

In this embodiment, specifically, the live pig state prediction model unit is configured to establish a live pig state prediction model, that is, according to sample data, a data vector and a transformation vector are set through fisher discrimination, so as to train to obtain an optimal projection direction of the sample data projected to a one-dimensional space; and according to the optimal projection direction, projecting the sample data on the optimal projection direction to obtain data, and solving the mean value and standard deviation of the data types of the normal live pig and the abnormal live pig to obtain a live pig state prediction model and a live pig state prediction result.

In this embodiment, specifically, according to sample data, a data vector and a transformation vector are set up through fisher discrimination to train to obtain an optimal projection direction of the sample data projected to a one-dimensional space; i.e. to establish a data vector x ═ x⁽¹⁾，x⁽²⁾) Wherein x is⁽¹⁾Body temperature variance; x is the number of⁽²⁾Is the ratio of food intake to body weight; defining a transformation vector w ═ (w)⁽¹⁾，w⁽²⁾) Wherein w is⁽¹⁾Is x⁽¹⁾Coefficient vector of (d), w⁽²⁾Is x⁽²⁾A coefficient vector of (a); finding the best transformation vector w^*＝S_w ^-1(m₂-m₁) Wherein the maximum value w of w^*The optimal transformation vector is the optimal projection direction obtained by training; s_wIs an intra-class covariance matrix; m is₁And m₂Mean values of normal and abnormal pigs, respectivelyAnd (5) vector quantity.

In particular, the optimum variation phasor w^*Obtained by the Lagrangian method.

In this embodiment, the Fisher decision is specifically a projection method established according to the concept of variance analysis, and projects a point in a high-dimensional space to a low-dimensional space; the samples which are difficult to separate under the original coordinate system can be obviously distinguished after projection.

Specifically, based on fisher discrimination, originally collected sample data of a live pig sample can be separated, and the optimal projection direction in the one-dimensional direction is obtained through the method.

In this embodiment, specifically, the data obtained by projecting the sample data on the optimal projection direction is obtained, and the mean and standard deviation, i.e. μ, of the data categories of the two states of the normal live pig and the abnormal live pig are obtained₁＝w^*m₁；μ₂＝w^*m₂；Wherein mu₁Data for normal live pigs are in w^*Mean, mu, of data obtained after vector axis projection₂Data for abnormal live pigs at w^*Mean, delta, of data obtained after vector axis projection₁Data for normal live pigs are in w^*Standard deviation, delta, of data obtained after vector axis projection₂Data for abnormal live pigs at w^*Standard deviation, N, of data obtained after vector axis projection₁Total number of data samples for normal live pigs, N₂Total number of data samples for abnormal live pigs.

In particular, w^*TIs w^*The transposing of (1).

In this embodiment, specifically, the live pig state prediction result output module is adapted to set the current live pig data as the data vector x_c，Zc＝w^*x_cAnd Zc is a data vector x_cAt w^*The value obtained after vector axis projection; the normality index D of the current live pig is as follows: when Zc is less than or equal to mu₁+δ₁When the normality index D is 0, the result is judged to be positiveA live pig; when Zc is not less than mu₂-δ₂If so, judging the pigs to be abnormal if the normality index D is 1; when mu is₁+δ₁≤Zc≤μ₂-δ₂Then, a live pig state prediction model is established, namelyJudging the pigs to be suspected to be abnormal; the normality index D is a number ranging from 0 to 1, and closer to 0 indicates higher normality, and closer to 1 indicates lower normality.

As shown in fig. 2, specifically, the live pig state prediction result output module is further connected to the server through a wireless module; wherein the live pig state prediction result output module is suitable for manually setting a threshold value D₀When the current normality index of the live pig is larger than a preset threshold value D₀Judging the pigs to be abnormal and shooting live pig videos with set duration; or when the normality index is lower than a set threshold D₀Judging the live pig to be normal and shooting a live pig image; and sending the live pig video and the live pig image data to a server for storage.

In this embodiment, optionally, the threshold D₀Set to 0.3.

In this embodiment, optionally, adjust the camera according to the normality index and shoot the invalid inspection work load that can reduce pig farm staff by a wide margin, with each traditional pig observe the strong comparison of subjective nature work load of inspection greatly, through data support in this embodiment, the staff only needs a small amount of work can in time discover sick pig.

Specifically, in the manual judgment stage, the normality index is greater than a preset threshold D₀Each video section needs to be browsed and the specific condition of the live pig is judged; for the normality index being lower than a set threshold value D₀The images of (2) are randomly selected and recorded.

In conclusion, the pig state prediction system for the pig farm obtains the health index of the pig by collecting the sample pig data and establishing the pig state prediction model, and obtains accurate and effective image video information of the pig by combining shooting of the camera, thereby providing effective assistance for judgment of workers in the pig farm, reducing the workload and the working time of low-value information, reducing subjective judgment of pig breeding, and being more accurate and scientific.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

In addition, the functional modules in the embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

In light of the foregoing description of the preferred embodiment of the present invention, many modifications and variations will be apparent to those skilled in the art without departing from the spirit and scope of the invention. The technical scope of the present invention is not limited to the content of the specification, and must be determined according to the scope of the claims.