Livestock tracking method, device and medium based on industrial internet

Abstract

The embodiment of the specification discloses a livestock tracking method and device based on an industrial internet and a medium, relates to the technical field of livestock tracking, and comprises the following steps: acquiring livestock tag information and a plurality of livestock monitoring data in a current time interval through an RFID tag arranged on each livestock and a plurality of internet of things devices, so as to determine a plurality of specified livestock monitoring data corresponding to each livestock in the current time interval; analyzing specified livestock video data to obtain motion data and body posture data of each livestock, analyzing a plurality of sensor data in the specified livestock monitoring data to obtain feeding data and vital sign data of each livestock, determining a livestock disease probability of each livestock through a tree structure analysis model, determining at least one risk livestock in close contact with a specified livestock when the livestock disease probability of the specified livestock is greater than a preset probability threshold, generating a risk livestock tracking group, and tracking the risk livestock tracking group.

Description

TECHNICAL FIELD

[0001] The present specification relates to the livestock tracking technical field, and particularly relates to a livestock tracking method and device based on an industrial internet and a medium. BACKGROUND

[0002] With the development of industrial internet technology and internet of things technology, new agricultural and livestock production lines have given birth to intelligent agriculture, which can improve productivity, reduce the impact on the environment, and record information essential to the agricultural process. Due to the nature of data collection, storage, and transmission by internet of things devices, a powerful information security and protection system is needed. As a result, an industrial internet platform is born to control the management and analysis of the entire internet of things data. In the field of livestock breeding, the uncontrollability of livestock makes the management or tracking process of livestock difficult. Through the internet platform, it can help livestock workers save time and resources and change the management of livestock.

[0003] In the process of livestock breeding, the monitoring of livestock is mainly to solve the problem of livestock getting lost and the health problem of livestock. In the health problem of livestock, the current physical sign data of livestock is usually monitored to obtain the physical condition of livestock. However, since the onset process of livestock is slow, especially infectious diseases, when an abnormality is detected through the current physical sign data, it may have spread and spread in the breeding farm, which will pose an infection risk to other livestock in the breeding farm and increase additional losses. Therefore, the existing technology monitors the health of livestock through immediate physical sign data, which is not suitable for slow-onset infectious diseases, and does not consider the potential risk of abnormal livestock to other normal livestock in the breeding farm, resulting in one-sidedness in livestock tracking and monitoring, increasing the risk of infectious disease transmission in the breeding farm. SUMMARY

[0004] One or more embodiments of the present specification provide a livestock tracking method and device based on an industrial internet, which solves the technical problem that the existing technology monitors the health of livestock through immediate physical sign data, which is not suitable for slow-onset infectious diseases, and does not consider the potential risk of abnormal livestock to other normal livestock in the breeding farm, resulting in one-sidedness in livestock tracking and monitoring, increasing the risk of infectious disease transmission in the breeding farm.

[0005] One or more embodiments of the present specification adopt the following technical solutions:

[0006] The one or more embodiments of the specification provide an industrial internet-based livestock tracking method, characterized in that the method comprises: acquiring livestock tag information and a plurality of livestock monitoring data in a current time interval through an RFID tag arranged on each livestock and a plurality of Internet of Things devices to determine a plurality of designated livestock monitoring data corresponding to each livestock in the current time interval; wherein the livestock tag information comprises tag identification and livestock basic information, and the livestock monitoring data comprises livestock video data and a plurality of sensor data; using a preset video analysis algorithm to analyze designated livestock video data in the designated livestock monitoring data to obtain motion data and body posture data of each livestock, and using a preset multi-sensor fusion algorithm to analyze a plurality of sensor data in the designated livestock monitoring data to obtain feeding data and vital sign data of each livestock; acquiring a tree structure analysis model corresponding to each livestock pre-constructed, and determining a livestock disease probability of each livestock through the tree structure analysis model according to the motion data, the body posture data, the feeding data and the vital sign data of each livestock, wherein the tree structure analysis model is obtained through historical monitoring data of each livestock, and comprises a trunk structure and a plurality of branch structures, the trunk structure is used to represent the livestock disease probability, and each branch structure corresponds to a type of monitoring data; when the livestock disease probability of a designated livestock is greater than a preset probability threshold, at least one risk livestock in close contact with the designated livestock is determined according to the motion data of the designated livestock and the livestock tag information, and a risk livestock tracking group is generated, so as to control the disease transmission risk of the breeding area by tracking the risk livestock tracking group.

[0007] Further, the preset video analysis algorithm is used to analyze the specified livestock video data in the specified livestock monitoring data to obtain motion data and body posture data of each livestock, specifically including: acquiring the specified livestock video data, wherein the specified livestock video data includes multiple frames of livestock images, and each livestock image includes at least one livestock; extracting features in each frame of livestock image through a feature extraction algorithm to obtain multiple livestock features and reference features, wherein the reference features include fence features and feeding bowl features; according to current livestock label information of a current livestock, performing corresponding relationship lookup in a preset feature library to obtain current livestock features corresponding to the current livestock, wherein the feature library includes multiple livestock features and livestock label information corresponding to each livestock feature; performing feature comparison between the current livestock features and the multiple livestock features to set a current livestock mark for the current livestock in each frame of livestock image; based on the order of each frame of livestock image in the specified livestock video data, the current livestock mark in each frame of livestock image and the reference features, motion data of the current livestock in a current time interval is generated, wherein the motion data includes a running track, a displacement length, at least one stopping place and a stopping time of each stopping place; through the current livestock mark, multiple body key points of the current livestock are extracted in each frame of livestock image to obtain key point position coordinates of each body key point; based on the key point position coordinates of each body key point in the specified livestock video data, body posture data of the current livestock is determined, wherein the body posture data includes a posture type and a posture duration, and the posture type includes a walking posture, a sitting posture, a lateral lying posture and a prone posture.

[0008] Further, a preset multi-sensor fusion algorithm is used to analyze the plurality of sensor data in the designated livestock monitoring data to obtain the feeding data and the physical sign data of each livestock, specifically including: determining the feeding food bowl identifier corresponding to the current livestock and the feeding time interval of the current livestock within the current time interval through at least one stay place in the motion data and the stay time of each stay place; determining, according to the feeding food bowl identifier, a plurality of food bowl weighing data corresponding to the feeding food bowl identifier in the plurality of sensor data through corresponding relationship lookup, wherein each food bowl weighing data includes a food bowl weight and a weighing time; performing key weight screening in the plurality of food bowl weights based on the feeding time interval of the current livestock within the current time interval and the weighing time in each food bowl weighing data to obtain a start food bowl weight at a start time and an end food bowl weight at an end time corresponding to the feeding time interval; performing algebraic operation on the start food bowl weight and the end food bowl weight to obtain the feeding weight of each livestock, taking the feeding weight and the feeding duration corresponding to the feeding time interval as the feeding data of the current livestock; obtaining a plurality of temperature data in the plurality of sensor data, wherein each temperature data includes a temperature value and a temperature collection time; determining a plurality of temperature correction times according to the stay time of each stay place in the motion data and the posture duration in the body posture data to perform temperature correction on a plurality of designated temperature values according to each temperature correction time and the temperature collection time to obtain the livestock temperature data.

[0009] Further, before obtaining the pre-constructed tree structure analysis model corresponding to each livestock, the method further comprises: obtaining a pre-constructed livestock state database, wherein the livestock state database comprises a plurality of historical disease monitoring data of livestock suffering from infectious diseases and a plurality of historical health monitoring data of healthy livestock, wherein the monitoring data comprises a plurality of monitoring dimensions, including a motion dimension, a posture dimension, a feeding dimension, and a physical sign dimension; performing weight analysis according to the plurality of historical disease monitoring data and the plurality of historical health monitoring data to obtain a monitoring weight of each monitoring dimension, wherein the monitoring weight is positively correlated with the disease probability; constructing an initial tree structure, wherein the initial tree structure comprises a trunk structure arranged from top to bottom and a plurality of branch structures arranged along the trunk structure, and the bottom end of the trunk structure is a root node, wherein each branch structure corresponds to a monitoring dimension; according to the monitoring weight of each monitoring dimension, setting the position of the branch structure corresponding to the monitoring dimension in the trunk structure to generate an initial tree structure model, wherein the greater the monitoring weight, the greater the distance between the branch structure corresponding to the monitoring dimension and the root node; obtaining a plurality of historical monitoring data corresponding to each livestock, optimizing the initial tree structure model through the historical monitoring data, determining a branch structure of each branch structure in the initial tree structure model to construct a required tree structure analysis model, wherein the branch structure comprises a change curve of the monitoring data under the monitoring dimension corresponding to the branch structure and a pre-set data reference range, and the data reference range is a reference range of monitoring data when suffering from a disease.

[0010] Further, through the tree structure analysis model, the disease probability of each livestock is determined according to the motion data, body posture data, feeding data, and physical sign data of each livestock, specifically comprising: according to the motion data, body posture data, feeding data, and physical sign data of each livestock, data filling is performed in the branch structure corresponding to the tree structure analysis model, the change curve in the branch structure is updated, and a current branch structure corresponding to each branch structure is obtained; through the current change curve in the current branch structure and the data reference range, a quantization factor of a branch node corresponding to each branch structure is generated, wherein the quantization factor is positively correlated with the distance between the current change curve and the data reference range; based on the quantization factor of each branch node and the distance between the branch node and the root node, probability calculation is performed to generate the disease probability of the livestock.

[0011] Further, according to the motion data of the specified livestock and the livestock tag information, at least one risk livestock in close contact with the specified livestock is determined, and a risk livestock tracking group is generated, specifically including: acquiring the motion data of the specified livestock, wherein the specified motion data includes a running track, a displacement length, at least one stay location, and a stay time at each stay location; identifying the RFID tag of each livestock through a pre-set tag reader to acquire position data and identification information of each livestock at multiple time points, so as to generate a synchronous running track of each livestock in the current time interval based on the position data at the multiple time points; aligning the synchronous running track of each livestock in the current time interval with the specified running track of the specified livestock according to time, so as to calculate a track distance between the motion data of the specified livestock and each synchronous running track at the same time, and obtain multiple risk synchronous running tracks with a track interval less than a pre-set distance threshold; setting a livestock identification for the multiple synchronous running tracks according to the identification information, and determining multiple risk livestock based on the livestock identification corresponding to the multiple risk synchronous running tracks.

[0012] Further, the disease transmission risk of the breeding area is controlled by tracking the risk livestock tracking group, specifically including: determining a risk tag identification corresponding to the multiple risk livestock in the risk livestock tracking group; acquiring multiple real-time tag position data collected by a tag reader, and determining a current activity position of each risk livestock in the real-time tag position data according to the risk tag identification; based on the current activity position of each risk livestock, setting an isolation risk area, so as to control the disease transmission risk of the breeding area through the isolation risk area.

[0013] Further, after generating the motion data of the current livestock in the current time interval based on the order of each frame of the livestock image in the specified livestock video data, the current livestock mark in each frame of the livestock image, and the reference feature, the method further includes: acquiring multiple real-time tag position data corresponding to the current livestock collected by a tag reader, and generating tag motion data according to the real-time tag position data; correcting the motion data of the current livestock in the current time interval through the tag motion data, and determining actual motion data of the current livestock.

[0014] One or more embodiments of the present specification provide an industrial internet-based livestock tracking device, including:

[0015] at least one processor; and,

[0016] a memory in communication connection with the at least one processor; wherein,

[0017] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above method.

[0018] The one or more embodiments of the present specification provide a non-volatile computer storage medium, which stores computer executable instructions, and the computer executable instructions are configured to: acquire livestock tag information and a plurality of livestock monitoring data in a current time interval through an RFID tag arranged on each livestock and a plurality of Internet of Things devices, to determine a plurality of designated livestock monitoring data corresponding to each livestock in the current time interval; wherein the livestock tag information includes tag identification and livestock basic information, and the livestock monitoring data includes livestock video data and a plurality of sensor data; analyze designated livestock video data in the designated livestock monitoring data by using a preset video analysis algorithm, to obtain motion data and body posture data of each livestock, and analyze a plurality of sensor data in the designated livestock monitoring data by using a preset multi-sensor fusion algorithm, to obtain feeding data and vital sign data of each livestock; acquire a tree structure analysis model corresponding to each livestock which is constructed in advance, and determine a livestock disease probability of each livestock according to the motion data, the body posture data, the feeding data and the vital sign data of each livestock through the tree structure analysis model, wherein the tree structure analysis model is obtained through historical monitoring data of each livestock, and includes a trunk structure and a plurality of branch structures, the trunk structure is used to represent the livestock disease probability, and each branch structure corresponds to a type of monitoring data; when the livestock disease probability of a designated livestock is greater than a preset probability threshold, at least one risk livestock in close contact with the designated livestock is determined according to the motion data of the designated livestock and the livestock tag information, a risk livestock tracking group is generated, so as to control the disease transmission risk of the breeding area by tracking the risk livestock tracking group.

[0019] The above at least one technical solution adopted by the embodiments of the present specification can achieve the following beneficial effects: Through the above technical solution, the livestock is monitored through the motion data, body posture data, feeding data and vital sign data of each livestock, multi-dimensional monitoring can be realized, and the large amount of complex multiple data is integrated into multi-dimensional monitoring data, which is more intuitive and representative. The tree structure analysis model corresponding to each livestock is obtained through historical monitoring data, the historical monitoring data is referred to, and is suitable for slowly occurring infectious diseases; in combination with the tree structure analysis model and the multi-dimensional data, various data are quantified to obtain a disease probability, which can be suitable for prediction of slowly occurring infectious diseases, and the accuracy and comprehensiveness of the disease probability are ensured. When the disease probability of the specified livestock is greater than the preset probability threshold, the risk livestock in close contact with the specified livestock is determined, a risk livestock tracking group is generated, and the livestock in the risk livestock tracking group is tracked, so that the data tracking amount is reduced, the tracking speed is improved, and the risk of rapid spread of infectious diseases can be coped with; through tracking of the livestock in the risk livestock tracking group, the disease spread risk of the breeding area is controlled, the potential risk of the risk livestock to other normal livestock in the breeding farm is avoided, and the risk of spread of infectious diseases in the breeding farm is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present specification or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments described in the present specification, and those skilled in the art can also obtain other drawings according to these drawings without creative labor. In the drawings:

[0021] Figure 1 A flowchart of a livestock tracking method based on an industrial internet provided by the embodiments of the present specification is shown in the figure.

[0022] Figure 2 A flowchart of a tree structure analysis model provided by the embodiments of the present specification is shown in the figure.

[0023] Figure 3 A structural diagram of a livestock tracking device based on an industrial internet provided by the embodiments of the present specification is shown in the figure. DETAILED DESCRIPTION

[0024] In order for those skilled in the art to better understand the technical solutions in the specification, the technical solutions in the specification will be clearly and completely described below in combination with the drawings in the specification. Obviously, the described embodiments are only a part of the embodiments of the specification, not all the embodiments. Based on the embodiments of the specification, all other embodiments obtained by those skilled in the art without creative labor should be within the protection scope of the specification.

[0025] With the development of industrial internet technology and Internet of Things technology, new agricultural and livestock production lines have given birth to intelligent agriculture, which can improve productivity, reduce environmental impact, and record information critical to the agricultural process. Due to the nature of Internet of Things devices collecting, storing and transmitting data, a strong information security and protection system is needed. Thus, the industrial internet platform is born to control the management and analysis of the entire Internet of Things data. In the field of livestock breeding, the uncontrollability of livestock makes it difficult to manage or track livestock. Through the Internet platform, it can help livestock workers save time and resources and change livestock management.

[0026] In the process of livestock breeding, the monitoring of livestock is mainly to solve the problem of livestock getting lost and the health problem of livestock. In the health problem of livestock, the current physical sign data of livestock such as body temperature is usually monitored to obtain the physical condition of livestock. However, since the onset process of livestock is slow, especially infectious diseases, when an abnormality is detected through the current physical sign data, it may have spread and spread in the breeding farm, which will pose an infection risk to other livestock in the breeding farm and increase additional losses. Therefore, the existing technology monitors the health of livestock through immediate physical sign data, which is not suitable for slow-onset infectious diseases, and does not consider the potential risk of abnormal livestock to other normal livestock in the breeding farm, resulting in one-sidedness in livestock tracking and monitoring and increasing the risk of infectious disease transmission in the breeding farm.

[0027] The embodiment of the specification provides a livestock tracking method based on an industrial internet. It should be noted that the execution subject in the embodiment of the specification can be a server or any device with data processing capability. Figure 1 A flowchart of a livestock tracking method based on an industrial internet provided by the embodiment of the specification is shown in Figure 1 As shown, it mainly includes the following steps:

[0028] Step S101, through the RFID tag set on each livestock and a plurality of Internet of Things devices, the livestock tag information and a plurality of livestock monitoring data in the current time interval are obtained to determine a plurality of specified livestock monitoring data corresponding to each livestock in the current time interval.

[0029] The embodiment of the present specification is applied to an industrial internet platform. Before step S101 is executed, all nodes are activated, including cloud, edge server, radio frequency identification (RFID) reader, RFID tag, and related nodes belonging to each livestock, breeding management personnel, and platform. According to actual needs, different permissions are given to users. All internet of things data from sensors are collected, including temperature, humidity, feed consumption, and animal behavior patterns, etc. Then, the data are transmitted to the platform for uploading to the cloud, thereby creating an unchangeable, secure animal information module. Then, data transmission between the RFID reader and the sensor tag is performed: first, the RFID reader locates the tag data; the tag responds to the radio frequency signal sent by the reader and establishes communication with the RFID reader; the RFID reader receives data from the sensor tag; the RFID reader aggregates all sensor tag data and transmits the data to the edge server, i.e., the gateway node; the data are transmitted to the cloud server, i.e., the industrial internet platform, through the gateway node; data management, analysis, and prediction are performed through the platform, and the results are displayed in the visualization module. All collected data can be displayed in real time through a smart device supporting Bluetooth. Breeding management users can analyze and observe the health status of animals at any location. Livestock management users and medical care management users can view the status of animals through mobile devices. Due to continuous monitoring, any signs of animal health problems will be immediately discovered.

[0030] Specifically, in one embodiment of the present specification, an RFID tag is arranged on each livestock, for example, through a collar or a neck strap, or in the form of an ear tag. The internet of things device can be an image acquisition device such as a camera and various sensors. The image acquisition device is arranged in the breeding area, and the number and acquisition parameters are set to cover the breeding area. The sensor device is arranged at the corresponding position, for example, a temperature sensor for collecting livestock temperature data can be arranged on the livestock, and a weight sensor for collecting food bowl weight data is arranged at the bottom of the food bowl. That is, the setting position of each sensor is related to the collected sensor data. When the RFID tag is set, each tag has a unique tag identifier, and the livestock information is written into the tag, so that when the reader scans the tag, the livestock information corresponding to the tag can be obtained through the tag identifier, that is, the tag is bound to the livestock. In addition, the sensor data arranged on the livestock is bound in advance. The corresponding relationship between the sensor data and the identification tag can be determined by adding the tag identifier corresponding to the livestock in the sensor data, and stored.

[0031] The livestock tag information and the plurality of livestock monitoring data in the current time interval are acquired through the RFID tag arranged on each livestock and the plurality of Internet of Things devices, the livestock tag information includes tag identification and livestock basic information, and the livestock monitoring data includes livestock video data and a plurality of sensor data. According to the corresponding relationship between the sensor data and the identification tag in the above steps, the plurality of specified livestock monitoring data corresponding to each livestock in the current time interval is determined. Since the image collected by the camera or the image collection device is the image of the entire breeding area, the livestock has mobility in the breeding area, and the plurality of videos collected can be regarded as the specified video data in the plurality of specified livestock monitoring data. According to the corresponding relationship between the sensor data and the identification tag, the sensor data belonging to the same livestock is classified into a category, so as to obtain the plurality of specified sensor data corresponding to each livestock.

[0032] In step S102, the specified livestock video data in the specified livestock monitoring data is analyzed by using a preset video analysis algorithm to obtain the motion data and the body posture data of each livestock, and the plurality of sensor data in the specified livestock monitoring data is analyzed by using a preset multi-sensor fusion algorithm to obtain the feeding data and the vital sign data of each livestock.

[0033] Specifically, in one embodiment of the present specification, first, the specified livestock video data in the specified livestock monitoring data is analyzed by using a preset video analysis algorithm to obtain the motion data and the body posture data of each livestock.

[0034] The specified livestock video data is obtained, and the specified livestock video data can be obtained by preliminary screening of videos collected by multiple video collection devices to ensure that each frame of livestock image in the specified livestock video data includes at least one livestock. In addition, in each frame of livestock image, in addition to the livestock, there are also environmental objects such as fences, feeding bowls and the like. By a feature extraction algorithm, livestock features and reference features in each frame of livestock image are extracted to obtain multiple livestock features and reference features, and the reference features include fence features and feeding bowl features. Taking a current livestock as a livestock to be tracked as an example, the livestock label information corresponding to the current livestock is determined, and the livestock label information includes a label identifier and livestock basic information of the current livestock. It should be noted that when matching the label with the livestock, a feature library is constructed in advance, and the livestock features corresponding to the label identifier are stored in the feature library to obtain the corresponding relationship of the livestock features corresponding to each label, that is, the feature library includes multiple livestock features and the livestock label information corresponding to each livestock feature. According to the label identifier in the current livestock label information, the corresponding relationship is searched in the preset feature library to obtain the current livestock feature corresponding to the current livestock. The current livestock feature is compared with multiple livestock features extracted from each frame of image, and the livestock feature in the image with the largest similarity is determined as the current livestock, and the current livestock is marked in each frame of image, for example, in the form of a bounding box or in the form of setting a numerical identifier. In the above manner, the position of the current livestock in each frame of image in the specified livestock video data is obtained.

[0035] By the order of each frame of livestock image in the specified livestock video data, the current livestock mark in each frame of livestock image and the reference features, the motion data of the current livestock in the current time interval is generated, and the motion data includes a running track, a displacement length, at least one stop location and a stop time of each stop location. When the running track is obtained, the current livestock mark of the current livestock in each frame of image in the specified livestock video data can be used as a position point at each time, and according to the change of the position point, the running track is obtained in combination with the time sequence of each frame of livestock image data, and the displacement data in the image is obtained in combination with the running track, the displacement data is scaled, and the displacement length of the current livestock in the actual breeding area is obtained. In addition, in combination with the position of the reference object and the position of the current livestock mark in each frame of image, when the two positions coincide and the position coincidence appears in multiple consecutive frames of image, it is determined that the current livestock has stopped at the position of the reference object, that is, the reference object is a stop location, and the stop time is obtained through the time interval of the multiple frames of image in which the position coincidence appears.

[0036] In addition, in order to ensure the accuracy of the motion data, the motion data obtained by the video is corrected by the positioning function of the RFID tag. Specifically, in one embodiment of the present specification, after generating the motion data of the current livestock within the current time interval based on the order of the livestock image in each frame of the specified livestock video data, the current livestock marker in each frame of the livestock image, and the reference feature, the multiple real-time tag position data corresponding to the current livestock collected by the tag reader is obtained, and the motion position data of the current livestock is determined through the real-time position data. According to the real-time tag position data, the tag motion data is generated. The data generation process here can obtain the running track through the real-time tag position data and its data collection time, and obtain the displacement length through the length calculation of the running track; the position data of each reference object is set in advance, and the reference object position data is compared with the real-time tag position data, and when the two coincide, the reference object is used as the stopping place, and the duration of the stopping place is obtained by combining the duration of the coincidence of the two.

[0037] The motion data of the current livestock within the current time interval is corrected by the tag motion data to determine the actual motion data of the current livestock. The correction process includes correction of the running track, correction of the displacement length, correction of the stopping place, and correction of the stopping time. In order to facilitate description, the motion data obtained by the video data is referred to as first motion data, i.e., including first running track, first displacement length, first stopping place and first stopping time, and the motion data obtained by the tag is referred to as second motion data, i.e., including second running track, second displacement length, second stopping place and second stopping time. Multiple reference points are taken in the first running track and the second running track, i.e., multiple first reference points and multiple second reference points, the correspondence between each first reference point and each second reference point is established, the midpoint of the line connecting the first reference point and the corresponding second reference point is taken, and the multiple midpoints obtained are connected in turn to obtain the corrected running track; the corrected displacement length is obtained according to the corrected running track. Generally, the first stopping place and the second stopping place are the same, if they are different, the first stopping place is used as the reference, in addition, the stopping time is used as the reference.

[0038] After obtaining the motion data of the current livestock, the current livestock is marked, and a plurality of body key points of the current livestock, such as shoulder, limb knee point, and trunk center point, are extracted in each frame of livestock image, and the key point position coordinates of each body key point are obtained. The body posture data of the current livestock is determined by the key point position coordinates of each body key point in the specified livestock video data, and the body posture data includes a posture type and a posture duration. The posture type includes a walking posture, a sitting posture, a lying posture, and a crawling posture. It should be noted that after the posture type is generated, if the same posture is recognized in a plurality of image data of consecutive frames, the time of the consecutive frames is taken as the posture duration.

[0039] Specifically, in one embodiment of the present specification, a plurality of sensor data in the specified livestock monitoring data is analyzed by using a preset multi-sensor fusion algorithm to obtain the feeding data and the vital sign data of each livestock.

[0040] The feeding bowl identifier corresponding to the current livestock and the feeding time interval of the current livestock in the current time interval are determined by at least one stop location in the motion data and the stop time of each stop location. Generally, there may be multiple feeding bowls in the breeding area, and only one livestock can feed in each feeding bowl at the same time. In this case, the position of the stop feeding bowl in at least one stop location in the motion data is compared with the preset position of each feeding bowl to obtain the feeding bowl identifier. In this case, there may be multiple feeding bowls, and the total stop time of each feeding bowl is taken as the feeding time.

[0041] A weight sensor is arranged at the bottom of each feeding bowl to upload the weighing data of the feeding bowl in real time. In order to identify the attribution of the weighing data, the feeding bowl identifier can be added when uploading the weighing data, or the corresponding relationship between the sensor identifier and the feeding bowl identifier can be set. According to the feeding bowl identifier, the plurality of feeding bowl weighing data corresponding to the feeding bowl identifier is determined in the plurality of sensor data by looking up the corresponding relationship. Each feeding bowl weighing data includes the weight of the feeding bowl and the weighing time. The key weight is selected from the plurality of feeding bowl weights by the feeding time interval of the current livestock in the current time interval and the weighing time in each feeding bowl weighing data. According to the first end time and the last end time of the feeding time interval, the two key weight weighing times are determined to obtain the starting feeding bowl weight corresponding to the starting time and the end feeding bowl weight corresponding to the ending time. The starting feeding bowl weight and the end feeding bowl weight are subtracted to obtain the feeding weight of each livestock. The feeding weight and the feeding time corresponding to the feeding time interval are taken as the feeding data of the current livestock. If multiple feeding bowls feed, the plurality of feeding weights and the feeding time are superimposed.

[0042] Obtain a plurality of temperature data in a plurality of sensor data, each of the temperature data comprising a temperature value and a temperature collection time, the temperature value representing a real-time body temperature of the livestock. Since the body temperature of the livestock is affected by movement and posture, for example, when the running track of the livestock is long, the corresponding body temperature is slightly higher than the normal body temperature; similarly, when the livestock is in a lateral recumbent posture and in a sleep state, the body temperature at this moment is slightly lower than the normal body temperature. Therefore, in order to avoid the body temperature deviation caused by the movement and the posture of the livestock, the temperature in this case can be corrected, and a plurality of temperature correction times are determined according to the stay time of each stay place in the movement data and the posture duration in the body posture data, so as to correct a plurality of specified temperature values according to each temperature correction time and the temperature collection time, and obtain the livestock temperature data. When correcting, if the experience data shows that the corresponding body temperature is too high, the deviation amount is reduced on the basis of the collected temperature data, and the deviation amount can be obtained based on the experience data. If the experience data shows that the corresponding body temperature is too low, the deviation amount is increased on the basis of the collected temperature data, and the deviation amount is obtained based on the experience data.

[0043] Through the above technical solution, the specified livestock video data in the specified livestock monitoring data is analyzed to obtain the movement data and the body posture data of each livestock, the plurality of sensor data in the specified livestock monitoring data is analyzed to obtain the feeding data and the physical sign data of each livestock, and the livestock is monitored through the movement data, the body posture data, the feeding data and the physical sign data of each livestock. The multi-dimensional monitoring can be realized, the large amount of data and the complex plurality of data are integrated into multi-dimensional monitoring data, and the multi-dimensional monitoring data is more intuitive and representative.

[0044] In step S103, a tree structure analysis model corresponding to each livestock is obtained in advance, and through the tree structure analysis model, the livestock disease probability of each livestock is determined according to the movement data, the body posture data, the feeding data and the physical sign data of each livestock.

[0045] The tree structure analysis model is obtained through the historical monitoring data of each livestock, and comprises a trunk structure and a plurality of branch structures. The trunk structure is used to represent the livestock disease probability, and each branch structure corresponds to a type of monitoring data.

[0046] Before step S103, a livestock state database pre-constructed is acquired, the livestock state database including a plurality of historical illness monitoring data of livestock suffering from infectious diseases and a plurality of historical health monitoring data of healthy livestock, the monitoring data including a plurality of monitoring dimensions, the monitoring dimensions including a motion dimension, a posture dimension, a feeding dimension and a sign dimension. According to the plurality of historical illness monitoring data and the plurality of historical health monitoring data, weight analysis is performed on monitoring data corresponding to each monitoring dimension to obtain a monitoring weight of each monitoring dimension, the monitoring weight being positively correlated with a probability of illness, and if the monitoring weight of a certain monitoring dimension is larger, it indicates that the corresponding probability of illness is greater when the monitoring data of the dimension is abnormal. It should be noted that the sum of the four monitoring weights is 1.

[0047] An initial tree structure is constructed, Figure 2 A structural diagram of a tree structure analysis model provided by an embodiment of the present specification is shown in Figure 2 As shown in Figure 2 The initial tree structure includes a trunk structure arranged from top to bottom and a plurality of branch structures arranged along the trunk structure, and the bottom end of the trunk structure is a root node, and each branch structure corresponds to a monitoring dimension. According to the monitoring weight of each monitoring dimension, the position of the branch structure corresponding to the monitoring dimension in the trunk structure is set to generate an initial tree structure model. It should be noted that the larger the monitoring weight, the farther the branch structure corresponding to the monitoring dimension from the root node, i.e., away from the root node. A plurality of historical monitoring data corresponding to each livestock is acquired, and the initial tree structure model is optimized by the historical monitoring data to determine the branch structure of each branch structure in the initial tree structure model to construct a tree structure analysis model meeting the requirements.

[0048] The motion dimension data, the feeding dimension data, the posture dimension data and the sign dimension data of the plurality of historical monitoring data are acquired, the motion dimension data taking displacement length as the motion amount under the motion dimension, the feeding dimension data taking feeding weight as the feeding amount under the feeding dimension, the posture dimension data taking the duration corresponding to the lateral posture as the posture abnormal duration under the posture dimension, and the sign dimension data being a temperature value. Through empirical data, a motion amount reference range, a feeding amount reference range, a posture abnormal reference range and a temperature reference range in an illness state are respectively acquired, and the change curves of the motion dimension, the feeding dimension, the posture dimension and the sign dimension are drawn in combination with the monitoring time of the plurality of historical monitoring data, the ordinate of each change curve being the dimension monitoring data and the abscissa being the monitoring time. The change curve of the monitoring data under the monitoring dimension and the pre-set data reference range form a branch corresponding to each dimension, i.e., the branch structure includes the change curve of the monitoring data under the monitoring dimension corresponding to the branch structure and the pre-set data reference range, and the data reference range is a reference range of the monitoring data in an illness state.

[0049] Specifically, in one embodiment of the present specification, the livestock disease probability of each livestock is determined according to the motion data, body posture data, feeding data and vital sign data of each livestock by a tree structure analysis model. According to the motion data, body posture data, feeding data and vital sign data of each livestock, data is filled in the branch structure corresponding to the tree structure analysis model, and the change curve in the branch structure is updated to obtain a current branch structure corresponding to each branch structure. A quantization factor of a branch node corresponding to each branch structure is generated by a current change curve in the current branch structure and a data reference range, wherein the quantization factor is positively correlated with the distance between the current change curve and the data reference range. When generating the quantization factor, the vertical distance between the latest dimension data at the latest time in the current branch structure and the lowest point or the highest point in the data reference range is calculated. Specifically, the relative position relationship between the current change curve and the data reference range can be determined in the selection of the lowest point or the highest point. The data reference range is composed of two parallel lines parallel to the horizontal axis, corresponding to two reference limits of the disease state. If the current change curve is below the data reference range, the minimum value, that is, the lowest point of the reference range, is taken. The data interval of the data reference range in the current branch structure, that is, the difference between the two longitudinal coordinates corresponding to the two parallel lines, is obtained, and the ratio of the vertical distance to the data interval is taken as the quantization factor. The quantization factor corresponding to the branch node output of each branch structure and the main branch structure is output, and the livestock disease probability of the livestock is generated based on the quantization factor of each branch node and the distance between the branch node and the root node. The distance coefficient is determined by the distance between the branch node and the root node. The distance coefficient can be set by the monitoring weight during model construction. The larger the monitoring weight, the greater the distance, and the larger the distance coefficient. The product of the quantization factor of each branch node and the distance coefficient between the branch node and the root node is calculated, and the multiple products are combined to output the livestock disease probability of the livestock through the root node.

[0050] Through the above technical solution, the tree structure analysis model corresponding to each livestock is obtained by historical monitoring data, which refers to the case of historical monitoring data and is suitable for slow-onset infectious diseases. The livestock disease probability of each livestock is determined by combining the tree structure analysis model, the motion data, body posture data, feeding data and vital sign data of each livestock. The tree structure analysis model associates each data with historical monitoring data, quantizes various data by combining multi-dimensional data, and obtains the disease probability, which can be suitable for the prediction of slow-onset infectious diseases, and ensures the accuracy and comprehensiveness of the disease probability.

[0051] Step S104, when the livestock disease probability of the specified livestock is greater than the preset probability threshold, at least one risk livestock in close contact with the specified livestock is determined according to the motion data and the livestock tag information of the specified livestock, and a risk livestock tracking group is generated, so as to control the disease transmission risk of the breeding area by livestock tracking on the risk livestock tracking group.

[0052] In an embodiment of the present specification, the probability threshold is set according to user needs, and can refer to the onset time of various infectious diseases. The probability threshold can be set in different time periods, for example, a smaller probability threshold is set in a time period when infectious diseases are prone to occur, the number of livestock tracking is increased, and the safety of the farm is ensured. When the livestock disease probability of the specified livestock is greater than the preset probability threshold, at least one risk livestock in close contact with the specified livestock is determined according to the motion data and the livestock tag information of the specified livestock, and a risk livestock tracking group is generated, so as to control the disease transmission risk of the breeding area by livestock tracking on the risk livestock tracking group.

[0053] Specifically, in an embodiment of the present specification, at least one risk livestock in close contact with the specified livestock is determined according to the motion data and the livestock tag information of the specified livestock, and a risk livestock tracking group is generated.

[0054] In the motion data of multiple livestock, the motion data of the specified livestock is obtained, and the specified motion data includes a running track, a displacement length, at least one stay location, and a stay time of each stay location. The RFID tag of each livestock in the breeding area is identified by a pre-set tag reader, and the position data and identification information of each livestock at multiple time points are obtained, so as to generate a synchronous running track of each livestock in the current time interval based on the position data at multiple time points. The synchronous running track of each livestock in the current time interval is aligned with the specified running track of the specified livestock according to time, so as to calculate the track distance between the motion data of the specified livestock and each synchronous running track at the same time, and obtain multiple risk synchronous running tracks with a track interval less than a preset distance threshold. If the distance is less than the preset distance threshold, it means that the distance is close at the same time, and it is close contact livestock. According to the identification information, the livestock identification of the multiple synchronous running tracks is set, and the multiple risk livestock is determined based on the livestock identification corresponding to the multiple risk synchronous running tracks. Since there are many livestock in the farm, the data amount is large, and when a risk occurs, it takes a long time to track the data of each livestock, and the risk of infectious disease spreading increases during this period, therefore, by generating a risk livestock tracking group, only the data corresponding to the generated risk livestock tracking group is analyzed, the data tracking amount is reduced, and the tracking speed is improved. In addition, when determining the risk livestock, the risk level of the risk livestock can also be set according to the duration of the close distance state, the longer the duration, the higher the corresponding risk level, indicating a greater risk of infection.

[0055] Specifically, in one embodiment of the present specification, the disease transmission risk of the breeding area is controlled by tracking the livestock of the risk livestock tracking group. The risk label identification corresponding to the multiple risk livestock in the risk livestock tracking group is determined; the multiple real-time tag position data collected by the tag reader is obtained, and the current activity position of each risk livestock is determined in the real-time tag position data according to the risk label identification; the area composed of the current activity position of each risk livestock is set as an isolation risk area, so that the risk livestock is isolated by the breeding management user according to the isolation risk area, to avoid the infection risk of other livestock, and to control the disease transmission risk of the breeding area.

[0056] Through the above technical solution, when the livestock disease probability of the specified livestock is greater than the preset probability threshold, the risk livestock in close contact with the specified livestock is determined, the risk livestock tracking group is generated, and the livestock in the risk livestock tracking group is tracked, thereby reducing the data tracking amount, improving the tracking speed, and being able to cope with the risk of rapid spread of infectious diseases; by tracking the livestock of the risk livestock tracking group, the disease transmission risk of the breeding area is controlled, and the potential risk of the risk livestock to other normal livestock in the farm is avoided, thereby reducing the disease transmission risk of the farm.

[0057] Through the technical solution, the livestock is monitored through the motion data, body posture data, feeding data and vital sign data of each livestock, multi-dimensional monitoring can be realized, and multiple data with large and complex data quantity are integrated into multi-dimensional monitoring data, which is more intuitive and representative. The tree structure analysis model corresponding to each livestock is obtained through historical monitoring data, the historical monitoring data is referred to, and the model is suitable for slow-onset infectious diseases; in combination with the tree structure analysis model and the multi-dimensional data, various data are quantified to obtain a disease probability, which can be applied to the prediction of slow-onset infectious diseases, and the accuracy and comprehensiveness of the disease probability are ensured. When the disease probability of the specified livestock is greater than a preset probability threshold, the risk livestock in close contact with the specified livestock is determined, a risk livestock tracking group is generated, and the livestock in the risk livestock tracking group is tracked, so that the data tracking amount is reduced, the tracking speed is improved, the risk of rapid spread of infectious diseases can be coped with, the disease spread risk of the breeding area is controlled through the tracking of the risk livestock tracking group, the potential risk of the risk livestock to other normal livestock in the breeding farm is avoided, and the risk of the spread of infectious diseases in the breeding farm is reduced.

[0058] The embodiments of the present specification also provide an industrial internet-based livestock tracking device, as shown in the accompanying drawings. Figure 3 The device includes at least one processor and a memory connected in communication with the at least one processor; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the above method.

[0059] The embodiment of the specification also provides a non-volatile computer storage medium, which stores computer executable instructions, and the computer executable instructions are configured to: acquire livestock tag information and a plurality of livestock monitoring data in a current time interval through an RFID tag arranged on each livestock and a plurality of Internet of Things devices, to determine a plurality of specified livestock monitoring data corresponding to each livestock in the current time interval; wherein the livestock tag information comprises a tag identifier and livestock basic information, and the livestock monitoring data comprises livestock video data and a plurality of sensor data; using a preset video analysis algorithm, analyzing specified livestock video data in the specified livestock monitoring data to obtain motion data and body posture data of each livestock, using a preset multi-sensor fusion algorithm, analyzing a plurality of sensor data in the specified livestock monitoring data to obtain feeding data and vital sign data of each livestock; acquiring a tree structure analysis model corresponding to each livestock which is constructed in advance, and determining a livestock disease probability of each livestock according to the motion data, the body posture data, the feeding data and the vital sign data of each livestock through the tree structure analysis model, wherein the tree structure analysis model is obtained through historical monitoring data of each livestock, and comprises a trunk structure and a plurality of branch structures, the trunk structure is used to represent the livestock disease probability, and each branch structure corresponds to a type of monitoring data; when the livestock disease probability of a specified livestock is greater than a preset probability threshold, at least one risk livestock in close contact with the specified livestock is determined according to the motion data of the specified livestock and the livestock tag information, and a risk livestock tracking group is generated, so as to control the disease transmission risk of the breeding area through livestock tracking of the risk livestock tracking group.

[0060] Each of the embodiments in the specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. Especially, for the device, equipment and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the related parts can be referred to the part of the method embodiment.

[0061] The above describes specific embodiments of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different than the order in the embodiments and still achieve the desired result. In addition, the processes depicted in the figures do not necessarily require the particular order shown or sequential order to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous or possible.

[0062] The device and medium provided by the embodiments of the present specification are one-to-one corresponding, and therefore the device and medium also have similar beneficial technical effects to the corresponding method. Since the beneficial technical effects of the method have been described in detail above, the beneficial technical effects of the device and medium will not be described here again.

[0063] Those skilled in the art will appreciate that embodiments of the present specification can be provided as methods, systems, or computer program products. Therefore, the present specification can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present specification can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer usable program code.

[0064] The present specification is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present specification. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing apparatus produce an apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the functions specified in the flowchart

[0065] These computer program instructions can also be stored in a computer-readable memory that can direct the computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the functions specified in the flowchart

[0066] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable data processing apparatus to produce a computer-implemented process, so that the instructions executed on the computer or other programmable data processing apparatus provide a process for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the functions specified in the flowchart

[0067] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memories.

[0068] Memory can include non-persistent memory, Random Access Memory (RAM), and / or non-volatile memory, such as Read Only Memory (ROM) or flash memory, in computer readable media. Memory is an example of computer readable media.

[0069] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile discs (DVDs) or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0070] It should also be noted that the terms "comprising", "containing", or any other similar term are intended to encompass non-exclusive inclusions, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements in the list, but can also include other elements not expressly listed, or can also include elements inherent in such process, method, article, or apparatus. Without more limitations, an element defined by the phrase "comprising a" does not exclude the existence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0071] The above description is merely one or more embodiments of the present specification and is not intended to limit the present specification. One or more embodiments of the present specification can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. within the spirit and principles of one or more embodiments of the present specification should be included in the scope of the claims of the present specification.

Claims

1. A livestock tracking method based on an industrial internet, characterized by, The method comprises: obtaining livestock tag information and a plurality of livestock monitoring data in a current time interval through an RFID tag arranged on each livestock and a plurality of Internet of Things devices to determine a plurality of specified livestock monitoring data corresponding to each livestock in the current time interval; wherein the livestock tag information comprises tag identification and livestock basic information, and the livestock monitoring data comprises livestock video data and a plurality of sensor data; using a preset video analysis algorithm to analyze specified livestock video data in the specified livestock monitoring data to obtain motion data and body posture data of each livestock, and using a preset multi-sensor fusion algorithm to analyze a plurality of sensor data in the specified livestock monitoring data to obtain feeding data and vital sign data of each livestock; obtaining a tree structure analysis model corresponding to each livestock constructed in advance, and determining a livestock disease probability of each livestock according to the motion data, the body posture data, the feeding data and the vital sign data of each livestock through the tree structure analysis model, wherein the tree structure analysis model is obtained through historical monitoring data of each livestock, comprises a trunk structure and a plurality of branch structures, and the trunk structure is used to represent the livestock disease probability, and each branch structure corresponds to a type of monitoring data; when the livestock disease probability of a specified livestock is greater than a preset probability threshold, determining at least one risk livestock in close contact with the specified livestock according to the motion data of the specified livestock and the livestock tag information, and generating a risk livestock tracking group, so as to control the disease transmission risk of the breeding area by tracking the risk livestock tracking group; Before obtaining the tree structure analysis model corresponding to each livestock constructed in advance, the method further comprises: obtaining a livestock state database constructed in advance, wherein the livestock state database comprises a plurality of historical disease monitoring data of livestock suffering from infectious diseases and a plurality of historical health monitoring data of healthy livestock, wherein the monitoring data comprises a plurality of monitoring dimensions, and the monitoring dimensions comprise a motion dimension, a posture dimension, a feeding dimension and a vital sign dimension; performing weight analysis according to the plurality of historical disease monitoring data and the plurality of historical health monitoring data to obtain a monitoring weight of each monitoring dimension, wherein the monitoring weight is positively correlated with the disease probability; constructing an initial tree structure, wherein the initial tree structure comprises a trunk structure arranged from top to bottom and a plurality of branch structures arranged along the trunk structure, and the bottom end of the trunk structure is a root node, wherein each branch structure corresponds to a monitoring dimension; setting a position of the branch structure corresponding to the monitoring dimension in the trunk structure according to the monitoring weight of each monitoring dimension to generate an initial tree structure model, wherein the greater the monitoring weight, the greater the distance between the branch structure corresponding to the monitoring dimension and the root node. Obtaining a plurality of historical monitoring data corresponding to each livestock, optimizing the initial tree structure model through the historical monitoring data, determining a branch structure of each branch structure in the initial tree structure model to construct a required tree structure analysis model, wherein the branch structure includes a change curve of the monitoring data in the monitoring dimension corresponding to the branch structure and a pre-set data reference range, and the data reference range is a reference range of the monitoring data when being ill; Through the tree structure analysis model, the livestock illness probability of each livestock is determined according to the motion data, body posture data, feeding data and body sign data of each livestock, specifically including: According to the motion data, body posture data, feeding data and body sign data of each livestock, data filling is performed in the branch structure corresponding to the tree structure analysis model, the change curve in the branch structure is updated, and a current branch structure corresponding to each branch structure is obtained; Through the current change curve in the current branch structure and the data reference range, a quantization factor of a branch node corresponding to each branch structure is generated, wherein the distance between the quantization factor and the current change curve and the data reference range is positively correlated; Based on the quantization factor of each branch node and the distance between the branch node and the root node, probability calculation is performed to generate the livestock illness probability of the livestock. 2.The livestock tracking method based on industrial internet according to claim 1, wherein, Using a preset video analysis algorithm, the specified livestock video data in the specified livestock monitoring data is analyzed to obtain the motion data and body posture data of each livestock, specifically including: Obtaining the specified livestock video data, wherein the specified livestock video data includes a plurality of livestock images, and each livestock image includes at least one livestock; Through a feature extraction algorithm, features in each frame of livestock image are extracted to obtain a plurality of livestock features and reference features, wherein the reference features include fence features and feeding bowl features; According to the current livestock tag information of the current livestock, corresponding relationship lookup is performed in a preset feature library to obtain the current livestock feature corresponding to the current livestock, wherein the feature library includes a plurality of livestock features and livestock tag information corresponding to each livestock feature; The current livestock feature is compared with the plurality of livestock features, and a current livestock mark is set for the current livestock in each frame of livestock image; Based on the order of each frame of the livestock image in the specified livestock video data, the current livestock mark in each frame of livestock image and the reference features, the motion data of the current livestock in the current time interval is generated, wherein the motion data includes a running track, a displacement length, at least one stopping place and a stopping time of each stopping place; Through the current livestock mark, a plurality of body key points of the current livestock are extracted in each frame of livestock image to obtain key point position coordinates of each body key point; Determine the body posture data of the current livestock based on the key point position coordinates of each of the body key points in the specified livestock video data, wherein the body posture data includes a posture type and a posture duration, and the posture type includes a walking posture, a sitting posture, a lying posture, and a lying-on-the-ground posture. 3.The livestock tracking method based on industrial internet according to claim 2, wherein, Analyze the multiple sensor data in the specified livestock monitoring data using a preset multi-sensor fusion algorithm to obtain the feeding data and the vital sign data of each livestock, specifically including: Determine the feeding food bowl identifier of the current livestock and the feeding time interval of the current livestock within the current time interval based on at least one stopover location and the stopover time of each stopover location in the motion data; Determine the multiple food bowl weighing data corresponding to the feeding food bowl identifier from the multiple sensor data based on the corresponding relationship lookup, wherein each of the food bowl weighing data includes a food bowl weight and a weighing time; Perform key weight screening among the multiple food bowl weights based on the feeding time interval of the current livestock within the current time interval and the weighing time in each of the food bowl weighing data to obtain the initial food bowl weight at the start time and the final food bowl weight at the end time corresponding to the feeding time interval; Perform algebraic operation on the initial food bowl weight and the final food bowl weight to obtain the feeding weight of each livestock, and take the feeding weight and the feeding duration corresponding to the feeding time interval as the feeding data of the current livestock; Obtain the multiple temperature data in the multiple sensor data, wherein each of the temperature data includes a temperature value and a temperature collection time; Determine multiple temperature correction times based on the stopover time of each stopover location in the motion data and the posture duration in the body posture data, and perform temperature correction on the multiple specified temperature values based on each temperature correction time and the temperature collection time to obtain the livestock temperature data. 4.The livestock tracking method based on industrial internet according to claim 1, wherein, Determine at least one risk livestock in close contact with the specified livestock based on the motion data of the specified livestock and the livestock tag information, and generate a risk livestock tracking group, specifically including: Obtain the motion data of the specified livestock, wherein the motion data includes a running trajectory, a displacement length, at least one stopover location, and a stopover time of each stopover location; Identify the RFID tag of each livestock using a pre-set tag reader to obtain the position data and the identifier information of each livestock at multiple time points, and generate the synchronous running trajectory of each livestock within the current time interval based on the position data at the multiple time points; Align the synchronous running trajectory of each livestock within the current time interval with the specified running trajectory of the specified livestock according to time, so as to calculate the trajectory distance between the motion data of the specified livestock and each of the synchronous running trajectories at the same time to obtain multiple risk synchronous running trajectories with a trajectory interval less than a preset distance threshold; Set the livestock identifier for multiple synchronous running trajectories according to the identifier information, and determine multiple risk livestock based on the livestock identifier corresponding to the multiple risk synchronous running trajectories. 5.The livestock tracking method based on industrial internet according to claim 1, wherein, The disease transmission risk of the breeding area is controlled by livestock tracking on the risk livestock tracking group, specifically including: Determining risk tag identifiers corresponding to a plurality of risk livestock in the risk livestock tracking group; Obtaining a plurality of real-time tag position data collected by a tag reader, and determining a current active position of each risk livestock in the real-time tag position data according to the risk tag identifiers; Based on the current active position of each risk livestock, a quarantine risk area is set to control the disease transmission risk of the breeding area through the quarantine risk area. 6.The livestock tracking method based on industrial internet according to claim 2, wherein, After generating the motion data of the current livestock within the current time interval based on the order of each frame of the livestock image in the specified livestock video data, the current livestock marker in each frame of the livestock image, and the reference feature, the method further includes: Obtaining a plurality of real-time tag position data corresponding to the current livestock collected by a tag reader, and generating tag motion data according to the real-time tag position data; The motion data of the current livestock within the current time interval is corrected through the tag motion data to determine the actual motion data of the current livestock. 7.An industrial internet-based livestock tracking device, characterized by, The device includes: At least one processor; and The memory is in communication with the at least one processor; wherein The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

8. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer executable instructions are configured to perform the method of any one of claims 1-6.

Images