Health monitoring devices, equipment and media for horses

Abstract

The present application is about a horse health monitoring device, equipment and medium, and specifically relates to the field of smart animal husbandry technology. The method includes: identifying whether a monitoring characteristic behavior appears in a picture in a monitoring video stream, wherein the monitoring characteristic behavior is an action behavior associated with the health status of the horse, and the monitoring characteristic behavior includes: a first type of characteristic behavior related to the physiological needs of the horse, and a second type of characteristic behavior characterizing the abdominal pain of the horse; counting the corresponding behavior distribution parameters of the monitoring video stream in the past first time period, wherein the behavior distribution parameters are used to indicate the abnormality of the first type of characteristic behavior; counting the corresponding behavior frequency parameters of the monitoring video stream in the past second time period, wherein the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior; combining the behavior distribution parameters and the behavior frequency parameters to determine the current health status of the horse.

Description

Technical Field

[0001] The present invention relates to the field of smart animal husbandry technology, and in particular to a horse health monitoring device, equipment and medium. Background Art

[0002] Equine colic, also known as "colic," is a type of abdominal pain syndrome caused by gastrointestinal diseases in horses. It is a common disease that poses a serious threat to the health of horses. If not identified and treated in time, it may cause serious life-threatening consequences.

[0003] Identification of abdominal pain in horses usually requires manual work and records to detect signs of the disease at an early stage. However, this requires staff to have keen observation and rich experience. On the other hand, if only manual observation is relied upon, there will be problems such as shortage of manpower and high costs.

[0004] Therefore, there is an urgent need for a convenient, accurate and low-cost automatic detection method for equine abdominal pain to monitor horse health. Summary of the invention

[0005] The present application provides a horse health monitoring device, equipment and medium, and the technical solution is as follows.

[0006] In one aspect, a method for monitoring the health of a horse is provided, the method comprising:

[0007] Identify whether a monitoring characteristic behavior appears in a picture in a monitoring video stream, wherein the monitoring characteristic behavior is an action behavior associated with the health status of a horse, and the monitoring characteristic behavior includes: a first type of characteristic behavior related to the physiological needs of a horse, and a second type of characteristic behavior representing abdominal pain of a horse;

[0008] Counting the behavior distribution parameters corresponding to the monitoring video stream in the past first time period, where the behavior distribution parameters are used to indicate the abnormality degree of the first type of characteristic behavior;

[0009] Counting the behavior frequency parameters corresponding to the monitoring video stream in the past second time period, where the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior;

[0010] The current health status of the horse is determined by combining the behavior distribution parameter and the behavior frequency parameter.

[0011] In a possible implementation, the counting of behavior distribution parameters corresponding to the monitoring video stream in the past first time period includes:

[0012] Acquiring a standard behavior index, wherein the standard behavior index is used to indicate a behavior index corresponding to a first type of characteristic behavior of a horse in a healthy state;

[0013] Counting behavior indicators to be evaluated, where the behavior indicators to be evaluated are used to indicate behavior indicators corresponding to the first type of characteristic behavior in a past first time period;

[0014] The behavior distribution parameter is determined based on the degree of difference between the behavior indicator to be evaluated and the standard behavior indicator.

[0015] In a possible implementation, determining the behavior distribution parameter based on the difference between the behavior indicator to be evaluated and the standard behavior indicator includes:

[0016] The Mahalanobis distance between the behavior index to be evaluated and the standard behavior index is calculated as the behavior distribution parameter.

[0017] In a possible implementation, combining the behavior distribution parameter and the behavior frequency parameter to determine the current health status of the horse includes:

[0018] Performing weighted summation on the behavior distribution parameter and the behavior frequency parameter to obtain a health assessment value;

[0019] When the health assessment value is greater than the health threshold, determining that the health status of the horse is a risky status;

[0020] When the health assessment value is not greater than the health threshold, it is determined that the health status of the horse is normal.

[0021] In a possible implementation, when the health assessment value is greater than a health threshold, the method further includes:

[0022] A risk warning message is issued, wherein the risk warning message is used to indicate that the health status of the horse is a risky status.

[0023] In a possible implementation, the identifying whether a picture in a surveillance video stream has surveillance characteristic behavior includes:

[0024] Performing key point recognition on a target image in the monitoring video stream to obtain a key point coordinate sequence, wherein the target image is an image containing a horse, and the key points are key position points on the horse's body;

[0025] Based on the key point coordinate sequence, determine whether the target image has the monitored characteristic behavior.

[0026] In a possible implementation, determining whether the target image has the monitoring characteristic behavior based on the key point coordinate sequence includes:

[0027] Performing coordinate processing on the key point coordinate sequence;

[0028] Inputting the processed key point coordinate sequence into a naive Bayes classifier to determine whether the target image has the monitored characteristic behavior;

[0029] Wherein, the naive Bayes classifier is constructed based on key position points of the horse's body.

[0030] In a possible implementation, the performing coordinate processing on the key point coordinate sequence includes:

[0031] Transferring the key point coordinate sequence to a coordinate system with the preset key point as the origin, and performing normalization processing;

[0032] and / or,

[0033] The coordinates of the key position points of the body that have not been identified in the key point coordinate sequence are set as the origin and standardized.

[0034] In a possible implementation, the first type of characteristic behavior includes at least one of the following:

[0035] Horses eat, horses defecate, horses urinate, and horses drink water.

[0036] In a possible implementation, the second type of characteristic behavior includes at least one of the following:

[0037] The horse's head and neck hang down, the horse's front legs dig the ground, the horse rolls, the horse rises and falls, the horse looks back at its belly, and the horse's hind legs kick its belly.

[0038] In another aspect, a horse health monitoring device is provided, the device comprising:

[0039] A behavior recognition module is used to identify whether a monitoring characteristic behavior appears in a picture in a monitoring video stream, wherein the monitoring characteristic behavior is an action behavior associated with the health status of a horse, and the monitoring characteristic behavior includes: a first type of characteristic behavior related to the physiological needs of a horse, and a second type of characteristic behavior representing abdominal pain of a horse;

[0040] A first parameter statistics module, used for counting the behavior distribution parameters corresponding to the monitoring video stream in the past first time period, wherein the behavior distribution parameters are used to indicate the abnormality degree of the first type of characteristic behavior;

[0041] A second parameter statistics module, used to count the behavior frequency parameters corresponding to the monitoring video stream in the past second time period, wherein the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior;

[0042] The health monitoring module is used to determine the current health status of the horse by combining the behavior distribution parameter and the behavior frequency parameter.

[0043] On the other hand, a computer device is provided, comprising a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the above-mentioned horse health monitoring method.

[0044] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction, at least one program, code set or instruction set is stored in the computer-readable storage medium, and the at least one instruction, at least one program, code set or instruction set is loaded and executed by a processor to implement the above-mentioned horse health monitoring method.

[0045] In another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions, the computer instructions being stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device performs the above-mentioned horse health monitoring method.

[0046] The technical solution provided by this application may have the following beneficial effects:

[0047] By monitoring the video stream to identify whether the horse has the first type of characteristic behavior related to the horse's physiological needs and the second type of characteristic behavior that characterizes the horse's abdominal pain, the behavior distribution parameters related to the first type of characteristic behavior and the behavior frequency parameters related to the second type of characteristic behavior are counted, and the daily health of the horse is judged based on the above two types of parameters. This can timely identify health abnormalities of the horse, reduce the cost of identifying horse abdominal pain, and improve the accuracy of identifying horse abdominal pain. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 is a method flow chart of a method for monitoring horse health according to an exemplary embodiment.

[0050] Figure 2 is a method flow chart of a method for monitoring horse health according to an exemplary embodiment.

[0051] Figure 3 is a schematic diagram showing key points of a horse according to an exemplary embodiment.

[0052] Figure 4 is a structural block diagram of a horse health monitoring device according to an exemplary embodiment.

[0053] Figure 5 is a schematic diagram of a computer device provided according to an exemplary embodiment. DETAILED DESCRIPTION

[0054] The technical solution of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0055] It should be understood that the "indication" mentioned in the embodiments of the present application can be a direct indication, an indirect indication, or an indication of an association relationship. For example, A indicates B, which can mean that A directly indicates B, for example, B can be obtained through A; it can also mean that A indirectly indicates B, for example, A indicates C, and B can be obtained through C; it can also mean that there is an association relationship between A and B.

[0056] In the description of the embodiments of the present application, the term "corresponding" may indicate a direct or indirect correspondence between two items, or an association relationship between the two items, or a relationship between indication and being indicated, configuration and being configured, and the like.

[0057] In an embodiment of the present application, "predefinition" can be achieved by pre-saving corresponding codes, tables or other methods that can be used to indicate relevant information in a device (for example, including a terminal device and a network device). The present application does not limit its specific implementation method.

[0058] Currently, the identification of abdominal pain in horses mainly relies on manual work and records, which has the problems of high cost, low convenience and lack of accuracy.

[0059] In view of the above defects, the embodiment of the present application aims to develop a horse health monitoring system through machine learning, which can perform daily health monitoring and abdominal pain warning for horses, greatly saving labor costs. The technical solution provided by the present application is further explained below.

[0060] Figure 1 is a flowchart of a method for monitoring horse health according to an exemplary embodiment. The method is applied in a computer device, such as Figure 1 As shown, the horse health monitoring method may include the following steps:

[0061] Step 110: Identify whether the image in the monitoring video stream shows monitoring characteristic behaviors, where monitoring characteristic behaviors are action behaviors associated with the health status of the horse, and monitoring characteristic behaviors include: first-class characteristic behaviors related to the physiological needs of the horse, and second-class characteristic behaviors representing abdominal pain of the horse.

[0062] The monitoring video stream is a video stream captured by a camera to monitor the horses. For example, a fixed monocular camera is arranged at a top corner of each horse's stable, and the monocular camera is set outside the reach of the horse, and the monitoring video stream is obtained through these cameras.

[0063] In an embodiment of the present application, two types of monitoring characteristic behaviors of horses are pre-set to be monitored, specifically a first type of characteristic behavior related to the physiological needs of the horse, and a second type of characteristic behavior representing abdominal pain in the horse. After acquiring the monitoring video stream, the horse behavior in each frame of the monitoring video stream is identified to identify whether the horse behavior belongs to the above-mentioned monitoring characteristic behavior.

[0064] In a possible implementation, the first type of characteristic behaviors includes at least one of the following: horse eating, horse defecation, horse urination, and horse drinking water.

[0065] In a possible implementation, the second type of characteristic behaviors includes at least one of the following: the horse's head and neck drooping, the horse's front legs digging the ground, the horse rolling, the horse falling, the horse looking back at the abdomen, and the horse's hind legs kicking the abdomen.

[0066] It is understandable that, generally speaking, typical symptoms of abdominal pain in horses include frequent rolling over, digging the ground with the forelimbs, kicking the abdomen with the hind limbs, looking back at the abdomen, sweating, etc. The characteristic behaviors corresponding to the above symptoms can be considered as the second type of characteristic behaviors directly related to abdominal pain in horses, but not all horses will show these obvious symptoms. Some horses may only show subtle changes such as restlessness, agitation, loss of appetite, weakened bowel sounds, and reduced defecation. Therefore, the characteristic behaviors corresponding to these symptoms are recorded as the first type of characteristic behaviors, and both the first and second type of characteristic behaviors are monitored.

[0067] It can be understood that the present application only exemplifies the monitoring of characteristic behaviors including two types of characteristic behaviors, and can further expand other types of characteristic behaviors in addition to the first type of characteristic behaviors and the second type of characteristic behaviors, or divide the first type of characteristic behaviors and the second type of characteristic behaviors into more types of characteristic behaviors.

[0068] Step 120: Count the behavior distribution parameters corresponding to the monitoring video stream in the past first time period, where the behavior distribution parameters are used to indicate the abnormality level of the first type of characteristic behavior.

[0069] In the embodiment of the present application, the abnormality of the first type of characteristic behavior in the past first time period is evaluated to obtain a behavior distribution parameter, wherein the size of the behavior distribution parameter is positively correlated with the abnormality of the first type of characteristic behavior, that is, the larger the behavior distribution parameter, the more abnormal the first type of characteristic behavior is.

[0070] In a possible implementation, step 120 specifically includes the following steps:

[0071] (1) Obtaining standard behavioral indicators, which are used to indicate behavioral indicators corresponding to the first type of characteristic behaviors of horses in a healthy state.

[0072] In this implementation, standard behavior indicators are preset, and the standard behavior indicators may specifically include the following types of behavior indicators: duration, number of times, and the like.

[0073] Exemplarily, standard behavioral indicators include the following: the length of time a horse eats, the number of times a horse defecates, the number of times a horse urinates, and the number of times a horse drinks water.

[0074] For example, the standard behavior index is obtained by counting the duration of the horse's eating, the number of times the horse defecates, the number of times the horse urinates, and the number of times the horse drinks water in 10 days, and calculating the mean value U = (u0, u1, u2, u3), where u0 represents the mean duration of the horse's eating, u1 represents the mean number of times the horse defecates, u2 represents the mean number of times the horse urinates, and u3 represents the mean number of times the horse drinks water. Each mean value is u i Represents the mean of the i-th behavior indicator, Represents the value of the i-th behavioral indicator on the n-th day.

[0075] (2) Counting the behavior indicators to be evaluated, where the behavior indicators to be evaluated are used to indicate the behavior indicators corresponding to the first type of characteristic behavior in the first time period in the past.

[0076] In this implementation, the behavior indicators of the first type of characteristic behavior in the past first time period are statistically calculated to obtain the behavior indicators to be evaluated.

[0077] (3) Determine the behavior distribution parameters based on the degree of difference between the behavior indicator to be evaluated and the standard behavior indicator.

[0078] In this implementation, the behavior distribution parameter can be determined by calculating the difference between the behavior indicator to be evaluated and the standard behavior indicator and characterizing the abnormality of the first type of characteristic behavior through the difference.

[0079] In a possible implementation, the Mahalanobis distance between the behavior indicator to be evaluated and the standard behavior indicator is calculated as the behavior distribution parameter. M represents the behavior distribution parameter, X represents the behavior index to be evaluated, and U represents the standard behavior index.

[0080] Step 130: Count the behavior frequency parameters corresponding to the monitoring video stream in the second time period in the past, where the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior.

[0081] The second time period may be the same as or different from the first time period, and this application does not impose any limitation on this.

[0082] In the embodiment of the present application, the occurrence frequency of the second characteristic behavior in the past second time period is counted to obtain a behavior frequency parameter.

[0083] Specifically, different types of second characteristic behaviors may not be distinguished, and the average of the frequencies of different types of second characteristic behaviors counted may be taken as the final behavior frequency parameter; or the frequencies of different types of second characteristic behaviors counted may be weighted first and then the average may be taken, and the sum of the weights of the second characteristic behaviors of different types is 1.

[0084] Exemplarily, the frequencies of characteristic behaviors such as the horse's head and neck drooping, the horse's front legs digging the ground, the horse's rolling, the horse's standing up and lying down, the horse's looking back at the abdomen, and the horse's hind legs kicking the abdomen within 24 hours are counted to obtain the behavior frequency parameters.

[0085] Step 140: Determine the current health status of the horse by combining the behavior distribution parameter and the behavior frequency parameter.

[0086] In the embodiment of the present application, after the behavior distribution parameter and the behavior frequency parameter are obtained by statistics, these two parameters are used to evaluate the health status of the horse.

[0087] In a possible implementation, step 140 specifically includes the following steps:

[0088] (1) Perform weighted summation of behavior distribution parameters and behavior frequency parameters to obtain the health assessment value.

[0089] For example, the behavior distribution parameter is recorded as x1, the behavior frequency parameter is recorded as x2, and the formula y=ax1+bx2 is input to obtain the health assessment value y, where a and b are the weights of the behavior distribution parameter and the behavior frequency parameter, respectively. The specific values ​​of a and b are not limited in this application.

[0090] Exemplarily, the weight of the behavior distribution parameter is smaller than the weight of the behavior frequency parameter, and whether the horse has abdominal pain is mainly judged by the second type of characteristic behavior that directly represents the horse's abdominal pain.

[0091] (2) When the health assessment value is greater than the health threshold, the health status of the horse is determined to be at risk.

[0092] Furthermore, when the health assessment value is greater than the health threshold, the method further includes the following steps: issuing risk warning information, wherein the risk warning information is used to indicate that the health status of the horse is at risk.

[0093] In this implementation, if the horse is judged to be in a risk state and may have abdominal pain symptoms according to the health assessment value, an early warning is issued through risk warning information. The risk warning information can be in the form of voice, text, picture, vibration, etc., which is not limited in this application.

[0094] (3) When the health assessment value is not greater than the health threshold, the health status of the horse is determined to be normal.

[0095] It is understandable that through the above-mentioned real-time and dynamic horse monitoring method, the health status of the horse can be accurately grasped, and further personalized feeding and training suggestions can be provided for the horse. Compared with the traditional way of understanding the health status of the horse through manual observation and regular physical examination, the system in this application that applies the horse health monitoring method can understand the health status of the horse in real time, and discover and deal with problems in time.

[0096] In summary, the horse health monitoring method provided in this embodiment monitors the video stream to identify whether the horse has a first type of characteristic behavior related to the horse's physiological needs and a second type of characteristic behavior that characterizes the horse's abdominal pain, and counts the behavior distribution parameters related to the first type of characteristic behavior and the behavior frequency parameters related to the second type of characteristic behavior. The above two types of parameters are combined to judge the horse's daily health, and the horse's health abnormalities can be identified in time, the cost of identifying the horse's abdominal pain can be reduced, and the accuracy of identifying the horse's abdominal pain can be improved.

[0097] Furthermore, by calculating the difference between the behavior index to be evaluated and the standard behavior index, the abnormality of the first type of characteristic behavior is characterized by the difference, thereby determining an accurate behavior distribution parameter.

[0098] In an exemplary embodiment, characteristic behaviors of horses are monitored based on the identification of key points in the image.

[0099] Figure 2 FIG. 1 is a flow chart of a method for monitoring horse health according to an exemplary embodiment. The method is applied in a computer device. Figure 2 As shown, the horse health monitoring method may include the following steps:

[0100] Step 210: perform key point recognition on the target image in the monitoring video stream to obtain a key point coordinate sequence, where the target image is an image containing a horse, and the key points are key position points on the horse's body.

[0101] In an embodiment of the present application, horses are identified in each picture in the monitoring video stream, and key points in the horses are identified, and multiple key points identified for each horse form a key point coordinate sequence.

[0102] For example, a target detection algorithm is used to detect whether a horse appears in a picture in a surveillance video stream. If a horse appears in the picture, a posture recognition model is used to estimate key points in the picture. The target detection algorithm and the posture recognition model can be constructed using Yolov8.

[0103] Exemplary, combined with reference Figure 3 , the following 24 key points of the horse's body are used as key points:

[0104] Serial number describe Serial number describe 1 Left Eye 13 Right front elbow 2 Right eye 14 Left posterior knee joint 3 Left ear heel 15 Right knee joint 4 Right ear 16 Left anterior wrist joint 5 tip of the nose 17 Right front wrist joint 6 throat 18 Left rear hock 7 Chest 19 Right rear hock 8 Hairpin 20 Left front ball joint 9 Baihui (highest point of buttocks) 21 Right front ball joint 10 Heel 22 Left rear ball joint 11 Tail tip (end of the tailbone) 23 Right rear ball joint 12 Left front elbow 24 abdomen

[0105] Step 220: Based on the key point coordinate sequence, determine whether the target image has monitoring feature behavior.

[0106] In the embodiment of the present application, whether the horse's behavior belongs to the monitored characteristic behavior is identified by detecting the key point coordinate sequence.

[0107] In a possible implementation, step 220 specifically includes the following steps:

[0108] (1) Perform coordinate processing on the key point coordinate sequence.

[0109] Further, the coordinate processing includes at least one of the following aspects:

[0110] (1.1) The key point coordinate sequence is transferred to a coordinate system with the preset key point as the origin and normalized.

[0111] Since the distance of the horse from the camera will cause the size of the horse in the picture to change, the influence of this change on the subsequent behavior classification results can be reduced by normalizing the key point coordinates. The normalization formula is:

[0112]

[0113] Where (x i ,y i ) is the coordinate of any key point of the horse, (x0, y0) is the coordinate of the preset key point, w = max(X)-min(X), h = max(Y)-min(Y), (X, Y) is the coordinate matrix of all key points of a horse. Further, the preset key point can be the first key point of the horse.

[0114] (1.2) The coordinates of the unrecognized key position points of the body in the key point coordinate sequence are set as the origin and standardized.

[0115] In order to avoid the influence of non-existent or occluded key points on the subsequent behavior classification results, the above-mentioned key points are set to (0.0, 0.0).

[0116] (2) Input the processed key point coordinate sequence into the naive Bayes classifier to determine whether the target image has monitoring feature behavior.

[0117] Among them, the Naive Bayes classifier is constructed based on the key position points of the horse's body.

[0118] Exemplarily, a naive Bayes classifier can construct a behavioral classification based on the following principles, for example: a horse's hind legs kicking the abdomen is classified based on the horse's hind legs being close to the abdomen; a horse looking back at the abdomen is classified based on the horse's head being toward the body; a horse standing up is classified based on the horse's body lying on the ground; a horse rolling is classified based on the horse's limbs facing up; a horse eating is classified based on the horse's head being close to the trough; a horse drinking is classified based on the horse's head being close to the trough; a horse defecating is classified based on the horse stretching and feces appearing nearby; a horse urinating is classified based on the horse stretching and urine appearing nearby; a horse digging the ground with its forelimbs is classified based on the horse's forelimbs being bent and stepping into the air; a horse lowering its head and neck is classified based on the horse's head being at right angles to its back.

[0119] Furthermore, for horse eating and drinking, the SSD (Single Shot MultiBox Detector) recognition algorithm can be used to pre-acquire the location coordinates of the feeding trough and the water trough; for horse defecation and urination, Mobilenet v3 can be used to identify feces and urine after recognizing that the horse is stretching.

[0120] For example, when a naive Bayes classifier is constructed, in order to enhance the generalization of the naive Bayes classifier, data enhancement is performed during training so that the model can achieve scale invariance, occlusion invariance, and rotation invariance.

[0121] Among them, scale invariance and occlusion invariance can be achieved through the above-mentioned normalization and standardization processing.

[0122] Regarding rotation invariance, since the camera shooting angle is fixed and the horse has a high degree of freedom in the stable, the horse's orientation relative to the camera will constantly change, which poses a challenge to behavior recognition. When training the naive Bayes classifier, the input key point coordinates can be randomly rotated by a certain angle to reduce the impact of the horse's rotation on the model's prediction results.

[0123] For example: Let (x c ,y c) is the center coordinate of the horse, (X, Y) is the key point coordinate sequence corresponding to all key points of a horse, then:

[0124]

[0125] Let θ be any radian, (x' i ,y' i ) is the coordinate of the i-th key point (x c ,y c ) is the coordinate after the center rotates by radians θ, then:

[0126] x' i =x c +(x i -x c )×cos(θ)-(y c -y i )×sin(θ);

[0127] y' i =y c -(x i -x c )×sin(θ)-(y c -y i )×cos(θ).

[0128] Step 230: Count the behavior distribution parameters corresponding to the monitoring video stream in the past first time period, where the behavior distribution parameters are used to indicate the abnormality level of the first type of characteristic behavior.

[0129] Step 240: Count the behavior frequency parameters corresponding to the monitoring video stream in the second time period in the past, where the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior.

[0130] Step 250: Determine the current health status of the horse by combining the behavior distribution parameter and the behavior frequency parameter.

[0131] The above steps 230 to 250 can refer to the above embodiments and will not be described in detail here.

[0132] In summary, the horse health monitoring method provided in this embodiment predefines multiple key points, and based on the recognition of key points in the image, the obtained key point coordinate sequence is input into the naive Bayes classifier, so that the monitoring feature behavior can be accurately identified.

[0133] Figure 4 1 is a block diagram of a device for monitoring horse health according to an exemplary embodiment. The device comprises:

[0134] The behavior recognition module 401 is used to recognize whether a monitoring characteristic behavior appears in a picture in a monitoring video stream, wherein the monitoring characteristic behavior is an action behavior associated with the health status of a horse, and the monitoring characteristic behavior includes: a first type of characteristic behavior related to the physiological needs of a horse, and a second type of characteristic behavior representing abdominal pain of a horse;

[0135] A first parameter statistics module 402, used to count the behavior distribution parameters corresponding to the monitoring video stream in the past first time period, wherein the behavior distribution parameters are used to indicate the abnormality degree of the first type of characteristic behavior;

[0136] A second parameter statistics module 403, used to count the behavior frequency parameters corresponding to the monitoring video stream in the past second time period, and the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior;

[0137] The health monitoring module 404 is used to determine the current health status of the horse by combining the behavior distribution parameter and the behavior frequency parameter.

[0138] In a possible implementation, the first parameter statistics module 402 is used to:

[0139] Acquiring a standard behavior index, wherein the standard behavior index is used to indicate a behavior index corresponding to a first type of characteristic behavior of a horse in a healthy state;

[0140] Counting behavior indicators to be evaluated, where the behavior indicators to be evaluated are used to indicate behavior indicators corresponding to the first type of characteristic behavior in a past first time period;

[0141] The behavior distribution parameter is determined based on the degree of difference between the behavior indicator to be evaluated and the standard behavior indicator.

[0142] In a possible implementation, the first parameter statistics module 402 is used to:

[0143] The Mahalanobis distance between the behavior index to be evaluated and the standard behavior index is calculated as the behavior distribution parameter.

[0144] In a possible implementation, the health monitoring module 404 is used to:

[0145] Performing weighted summation on the behavior distribution parameter and the behavior frequency parameter to obtain a health assessment value;

[0146] When the health assessment value is greater than the health threshold, determining that the health status of the horse is a risky status;

[0147] When the health assessment value is not greater than the health threshold, it is determined that the health status of the horse is normal.

[0148] In a possible implementation, the health monitoring module 404 is used to:

[0149] A risk warning message is issued, wherein the risk warning message is used to indicate that the health status of the horse is a risky status.

[0150] In a possible implementation, the behavior recognition module 401 is used to:

[0151] Performing key point recognition on a target image in the monitoring video stream to obtain a key point coordinate sequence, wherein the target image is an image containing a horse, and the key points are key position points on the horse's body;

[0152] Based on the key point coordinate sequence, determine whether the target image has the monitored characteristic behavior.

[0153] In a possible implementation, the behavior recognition module 401 is used to:

[0154] Performing coordinate processing on the key point coordinate sequence;

[0155] Inputting the processed key point coordinate sequence into a naive Bayes classifier to determine whether the target image has the monitored characteristic behavior;

[0156] Wherein, the naive Bayes classifier is constructed based on key position points of the horse's body.

[0157] In a possible implementation, the behavior recognition module 401 is used to:

[0158] Transferring the key point coordinate sequence to a coordinate system with the preset key point as the origin, and performing normalization processing;

[0159] and / or,

[0160] The coordinates of the key position points of the body that have not been identified in the key point coordinate sequence are set as the origin and standardized.

[0161] In a possible implementation, the first type of characteristic behavior includes at least one of the following:

[0162] Horses eat, horses defecate, horses urinate, and horses drink water.

[0163] In a possible implementation, the second type of characteristic behavior includes at least one of the following:

[0164] The horse's head and neck hang down, the horse's front legs dig the ground, the horse rolls, the horse rises and falls, the horse looks back at its belly, and the horse's hind legs kick its belly.

[0165] It should be noted that the horse health monitoring device provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0166] See also Figure 5 , which is a schematic diagram of a computer device provided according to an exemplary embodiment of the present application, the computer device includes a memory and a processor, the memory is used to store a computer program, and when the computer program is executed by the processor, the above-mentioned horse health monitoring method is implemented.

[0167] The processor may be a central processing unit (CPU). The processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0168] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the method in the embodiment of the present invention. The processor executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory, that is, implementing the method in the above method embodiment.

[0169] The memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created by the processor, etc. In addition, the memory may include a high-speed random access memory, and may also include a non-transitory memory, such as at least one disk storage device, a flash memory device, or other non-transitory solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0170] In an exemplary embodiment, a computer-readable storage medium is also provided, which is used to store at least one computer program, and the at least one computer program is loaded and executed by a processor to implement all or part of the steps in the above method. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0171] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any variations, uses or adaptations of the present application, which follow the general principles of the present application and include common knowledge or customary techniques in the art that are not disclosed in the present application. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

Claims

1. A horse health monitoring device, characterized in that: The device comprises: A behavior recognition module is used to identify key points of a target image in a monitoring video stream to obtain a key point coordinate sequence, wherein the target image is an image containing a horse, and the key points are key body position points of the horse; the key point coordinate sequence is transferred to a coordinate system with a preset key point as the origin, and normalized; and / or the coordinates of the key body position points that are not identified in the key point coordinate sequence are set as the origin, and standardized; the key point coordinate sequence after coordinate processing is input into a naive Bayes classifier to determine whether the target image has a monitoring characteristic behavior, wherein the naive Bayes classifier is constructed based on the key body position points of the horse, and the monitoring characteristic behavior is an action behavior associated with the health status of the horse, and the monitoring characteristic behavior includes: a first type of characteristic behavior related to the physiological needs of the horse, and a second type of characteristic behavior characterizing the abdominal pain of the horse; A first parameter statistics module is used to obtain a standard behavior index, which is used to indicate a behavior index corresponding to a first type of characteristic behavior of a horse in a healthy state; to count a behavior index to be evaluated, which is used to indicate a behavior index corresponding to the first type of characteristic behavior in a past first time period; based on the degree of difference between the behavior index to be evaluated and the standard behavior index, determine the behavior distribution parameter, which is used to indicate the degree of abnormality of the first type of characteristic behavior, and the behavior distribution parameter is the Mahalanobis distance between the behavior index to be evaluated and the standard behavior index; A second parameter statistics module, used to count the behavior frequency parameters corresponding to the monitoring video stream in the past second time period, wherein the behavior frequency parameters are used to indicate the frequency of the second type of characteristic behavior; The health monitoring module is used to determine the current health status of the horse by combining the behavior distribution parameter and the behavior frequency parameter.

2. The device according to claim 1, characterized in that In the health monitoring device, the combining of the behavior distribution parameter and the behavior frequency parameter to determine the current health status of the horse includes: The behavior distribution parameter and the behavior frequency parameter are weightedly summed to obtain a health assessment value; when the health assessment value is greater than a health threshold, the health status of the horse is determined to be a risky state; when the health assessment value is not greater than the health threshold, the health status of the horse is determined to be a normal state.

3. The device according to claim 2, characterized in that The health monitoring device is also used to issue risk warning information when the health assessment value is greater than a health threshold, and the risk warning information is used to indicate that the health status of the horse is a risky status.

4. The device according to claim 1, characterized in that The first type of characteristic behavior includes at least one of the following: Horses eat, horses defecate, horses urinate, and horses drink water.

5. The device according to claim 1, characterized in that The second type of characteristic behavior includes at least one of the following: The horse's head and neck hang down, the horse's front legs dig the ground, the horse rolls, the horse rises and falls, the horse looks back at its belly, and the horse's hind legs kick its belly.

6. A computer device, characterized in that: The computer device comprises a processor and a memory, wherein the memory stores the horse health monitoring device according to any one of claims 1 to 5.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores the horse health monitoring device according to any one of claims 1 to 5.

Images