A method and system for early identification of lameness in cattle

Abstract

This application relates to the field of image recognition technology, specifically to a method and system for early identification of bovine hoof lameness. The method includes: acquiring visible light and infrared video of each cow walking; identifying and segmenting the coronary band region, spinal key points, and hoof key points frame by frame in the visible light video; quantifying the interhoof temperature difference to assess local physiological thermal abnormalities to calculate a first evaluation value for each cow; quantifying the spinal curvature to assess arched back posture abnormalities to calculate a second evaluation value for each cow; assessing gait abnormalities to calculate a third evaluation value for each cow; determining the evaluation coefficient for each cow; and evaluating and identifying early hoof lameness. This application improves the sensitivity and accuracy of early lameness identification.

Description

Technical Field

[0001] This application relates to the field of image recognition technology, specifically to a method and system for early identification of bovine hoof lameness. Background Technology

[0002] In intensive animal husbandry, the pathological basis of lameness in cattle is mostly hoof diseases, which is one of the most common and impactful health problems in cattle herds. It not only causes pain and discomfort in cattle, but also significantly reduces their production performance. When cattle exhibit obvious lameness that is visible to the naked eye, their hoof lesions have often progressed to the middle or late stages, resulting in deep tissue infection or structural damage. At this point, treatment is costly and ineffective.

[0003] Traditional methods for identifying lameness in cattle mainly rely on the naked eye observation of breeders. However, lameness in its early stages manifests as extremely subtle physiological and behavioral changes. The naked eye cannot quantitatively detect the underlying physiological abnormalities in the early stages of lameness; for example, it cannot perceive the temperature rise caused by changes in local microcirculation due to early hoof inflammation. Furthermore, the subtle behavioral changes accompanying early lameness, such as slight gait cycle disturbances and spinal curvature, are difficult for the human eye to continuously and stably track and judge. Additionally, the behavioral manifestations of early lameness can easily appear similar to other normal behaviors in cattle, such as restlessness during estrus, leading to biases in the identification of early lameness and affecting the accuracy of identifying lameness in cattle. Summary of the Invention

[0004] To address the aforementioned technical problems, a method and system for early identification of bovine hoof lameness are provided to solve the existing issues.

[0005] The solution to the technical problem of this application is to provide a method and system for early identification of bovine lameness, including the following steps: In a first aspect, embodiments of this application provide a method for early identification of bovine hoof lameness, the method comprising the following steps: Acquire visible light and infrared videos of each cow while it is walking; for the visible light videos, identify and segment the coronary band region, key points of the spine, and key points of the hooves frame by frame. For each cow's different hooves, the differences in temperature distribution in the corresponding coronary band area in infrared video were analyzed, the interhoof temperature difference was quantified to assess local physiological thermal anomalies, and the first assessment value for each cow was calculated. For visible light video, the spatial distribution characteristics of key points of the spine in each frame are used to quantify the spinal curvature to assess the abnormality of hunched posture and obtain a second evaluation value for each cow. Based on key points of the cattle hoof in different frames, the periodic changes in the contact between the cattle hoof and the deviation of the hoof posture direction are evaluated to assess gait abnormalities. A third evaluation value is determined for each cattle. The evaluation coefficient for each cattle is determined by combining the first and second evaluation values ​​to assess and identify early lameness of the cattle hoof.

[0006] Preferably, the key points of the spine include the coccyx point, the cervical point, and a plurality of intermediate points evenly selected between the two along the spinal line.

[0007] Preferably, the key points of the cow hoof include the tip of each hoof and the center point of the coronary band area.

[0008] Preferably, the calculation of the first assessment value for each cow includes: For each cow's visible light video, based on each coronary band region in each frame, find its corresponding region in the same frame of the infrared video, and take the average temperature of all pixels contained in that region as the local temperature; calculate the average local temperature of each coronary band region across all frames as the average temperature. The first evaluation value is the result of normalizing the maximum difference in average temperature between any two crown band regions.

[0009] Preferably, obtaining the second assessment value for each cow includes: For each frame of the visible light video, the tailbone point and the neck point are connected by a line, and the distance from the tailbone point to the neck point is recorded as the line distance. The mean of the line distances of all frames of the visible light video is calculated as the average distance. Calculate the vertical distance from each midpoint to the connecting line in each frame, and select the maximum vertical distance among all frames of the visible light video. The second evaluation value is positively correlated with the maximum vertical distance, but negatively correlated with the average distance.

[0010] Preferably, determining the third assessment value for each cow includes: For visible light video, the ordinate of the center point of the coronary band region corresponding to each hoof in each frame is extracted. The frame sequence and ordinate of the same hoof in each frame are combined into a two-dimensional array. Curve fitting is performed on all two-dimensional arrays for each hoof. The frame sequence corresponding to the trough on the fitted curve is defined as the landing frame sequence of each hoof, and the frame sequence corresponding to the first trough is defined as the first landing frame sequence of each hoof. The unit direction vector between the position coordinates of the two hoof key points on each hoof in each frame is obtained. Collect visible light videos of multiple healthy cows walking, analyze the periodic distribution of adjacent landing frames of the same cow hoof, and the average level of the unit direction vector in the corresponding frames of the landing frame sequence, and calculate the standard gait period and standard landing vector of each cow hoof. For each cow, starting from the first landing frame of each hoof, and using the standard gait cycle of that hoof as the time interval, all theoretical landing frames are obtained. The angle between the unit direction vector of each hoof and the standard landing vector of that hoof in the corresponding frame of each theoretical landing frame is calculated. The maximum angle is selected and normalized, and used as the third evaluation value for each cow. Among them, the The first of the cow's hooves A theoretical landing frame sequence for: , For the first The first frame sequence of a cow's hoof landing. For the first The standard gait cycle of a cow's hoof. Indicates the number of cycles.

[0011] Preferably, the process of obtaining the standard gait cycle and standard landing vector is as follows: traverse the visible light video of each healthy cow to obtain all landing frame sequences and the unit direction vector of each hoof; calculate the average time interval between all two adjacent landing frame sequences for each hoof of each healthy cow in the visible light video as the average interval; take the average interval of the same hoof among all healthy cows as the standard gait cycle of each hoof; calculate the average unit direction vector of each hoof of each healthy cow in all landing frame sequences as the average vector; and calculate the average vector of the same hoof among all healthy cows as the standard landing vector of each hoof.

[0012] Preferably, the evaluation coefficient is the result of a positive fusion of the first evaluation value, the second evaluation value, and the third evaluation value.

[0013] Preferably, the assessment and identification of early lameness in cattle hooves includes: if the assessment coefficient of each cow is less than a preset first value, the cow does not have early lameness in its hooves; if the assessment coefficient is greater than or equal to the preset first value and less than a preset second value, the cow has a risk of lameness in its hooves; if the assessment coefficient is greater than the preset second value, the cow has early lameness in its hooves; wherein the preset first value is less than the preset second value.

[0014] Secondly, embodiments of this application also provide an early identification system for bovine hoof lameness, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any of the above-described methods for early identification of bovine hoof lameness.

[0015] This application has at least the following beneficial effects: This application introduces visible light and infrared video to subsequently assess the behavioral and physiological characteristics of cattle. It identifies and segments the coronary band region, spinal key points, and hoof key points frame by frame. The beneficial effect is that it achieves automated, high-precision localization and segmentation of key areas through a neural network model. A first assessment value is calculated for each cow, which is beneficial because it extracts quantitative physiological indicators from the infrared video and can sensitively capture early hoof inflammation through interhoove temperature differences, thus providing a preliminary assessment of the likelihood of lameness caused by hoof inflammation. A second assessment value is obtained for each cow, which is beneficial because it allows for analysis of the spine during walking from a static behavioral perspective. The bending characteristics are used to reflect the protective arching of the back that cattle often exhibit due to pain, further assessing the likelihood of early hoof lameness; a third assessment value is determined for each cow, which is beneficial in analyzing whether the gait cycle is disordered and whether the posture is correct at the theoretical landing moment from a dynamic behavioral dimension, in order to capture dynamic gait incoordination caused by pain, further assessing the likelihood of early hoof lameness; an assessment coefficient is determined for each cow to assess and identify early hoof lameness, which is beneficial in comprehensively assessing early hoof lameness from both physiological and two behavioral levels, significantly improving the sensitivity and accuracy of early lameness identification. Attached Figure Description

[0016] The following is a detailed description of a method for early identification of bovine hoof lameness according to the present application, with reference to the accompanying drawings.

[0017] Figure 1 A flowchart illustrating the steps of an early identification method for bovine hoof lameness provided in this application embodiment; Figure 2 A flowchart illustrating the method for obtaining evaluation coefficients provided in this application embodiment. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of a method and system for early identification of bovine hoof lameness, in conjunction with the accompanying drawings and embodiments, is provided. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0020] Please see Figure 1 The diagram illustrates a flowchart of a method for early identification of bovine hoof lameness according to an embodiment of this application. The method includes the following steps: Step 1: Obtain visible light and infrared video of each cow while it is walking.

[0021] Lameness is the most common and serious health problem in cattle. It is often an early sign of hoof disease or injury, primarily a painful condition caused by hoof lesions. It directly affects the cattle's mobility and behavior. If left untreated, it can worsen, leading to more serious health problems such as hoof infections and laminitis, and even significantly reducing milk production and impairing reproductive performance. Therefore, early identification of hoof lameness allows for timely detection and treatment, alleviating the cattle's pain and discomfort.

[0022] Based on the above analysis, a dedicated detection channel was built at the entrance and exit of the cattle shed. The channel is 5m long to ensure the acquisition of continuous walking data of the cattle. The channel is 1m wide to limit the passage to only one cow at a time, thereby avoiding obstruction and data confusion between individuals. High-resolution visible light cameras and infrared thermal imaging cameras were deployed on one side of the channel to simultaneously collect walking videos of each cow after it enters the channel, including visible light video and infrared video. In this implementation, the frame rate of both the visible light camera and the infrared thermal imaging camera is set to 24fps, and the resolution is set to 1920*1080. As for other implementation methods, the implementer can set them according to the actual situation. Secondly, the visible light video and the infrared video are acquired synchronously. Therefore, each frame in the visible light video can be found in the infrared video.

[0023] Secondly, the video frames at the beginning and end of the visible light and infrared videos are removed to ensure that the cattle are fully presented in the final video and to avoid situations where only part of the cattle's body is shown.

[0024] At this point, we have obtained visible light and infrared video of each cow's movement.

[0025] Step 2: For visible light video, identify and segment the coronary band region, spinal key points, and hoof key points frame by frame; for different hooves of each cow, analyze the differences in temperature distribution of the corresponding coronary band region in infrared video, quantify the temperature difference between hooves to assess local physiological thermal anomalies, and calculate the first assessment value for each cow.

[0026] The causes of lameness in cattle are mostly diseases of the hoof, or a small number of infections and inflammations caused by exogenous wounds. In the early stages, cattle will only show subtle changes in physiology and behavior (lame posture) due to pain. Other normal behaviors of cattle may have some similarities to the early lame posture. For example, cattle may also show some unusual behaviors when in estrus, which may interfere with the judgment of early lameness and lead to a high false alarm rate.

[0027] Secondly, before disease or wound inflammation causes lameness in cattle, it will first manifest at the physiological level. For example, inflammation of the hoof will cause the local blood flow to increase, which will cause the local temperature of the inflamed area to be higher than the local temperature of the corresponding area of ​​other healthy hooves. Secondly, inflammation will also cause pain, which will cause subtle changes in the cattle's behavior, including uncoordinated gait and pain causing the cattle to involuntarily arch their backs when walking on the inflamed hoof.

[0028] Based on the above analysis, the area where the hoof crown is located and key points in the video were segmented, specifically as follows: Multiple cows were driven through a dedicated detection channel in turn, and visible light video of each cow's walk was acquired. The video was then processed by frame segmentation, and multiple frames were uniformly extracted from the visible light video of each cow to form a dataset. In one embodiment, visible light videos of 100 cows walking are acquired, and 10 frames are uniformly extracted from the visible light videos of each cow to form a dataset. As another implementation method, the implementer can set it according to the actual situation.

[0029] The images in the dataset were manually annotated using an image annotation tool. The annotation objects were: Mark the coronal band area of ​​each cow hoof in the image, which is the ring-shaped band at the junction of the hoof shell and the skin; Mark multiple key points of the spine in the image, namely the tailbone point, the neck point, and three intermediate points evenly selected between the two along the spine line of the cattle. Mark the key points of the cow hooves in the image, namely the hoof tip of each hoof and the center point of the coronary band area; The neural network is trained based on the labeled dataset; In this embodiment, the Mask R-CNN model is used for training. The optimizer uses stochastic gradient descent (SGD), and the loss function is the sum of three loss functions: classification loss, bounding box regression loss, and mask segmentation loss. The classification loss uses cross-entropy loss, the bounding box regression loss uses L1 loss, and the mask segmentation loss uses binary cross-entropy loss. The Mask R-CNN model is a well-known technique and will not be described in detail here.

[0030] The visible light video is processed frame by frame, and the trained neural network is used to identify and segment the coronary band region, spinal key points and bovine hoof key points in each frame. Secondly, from a physiological perspective, the coronary band is the area most prone to and sensitive to inflammation, and temperature changes are first manifested in this area. Contact with dew and grass affects the absolute temperature of the coronary band. Therefore, the temperature differences between different coronary band areas in each frame of the infrared video are assessed, and a first evaluation value is calculated, specifically: For any hoof and crown region in each frame of the visible light video, find its corresponding region in the same frame of the infrared video, and take the average temperature of all pixels contained in that region as the local temperature. Calculate the average local temperature of any crown band region across all frames, and use it as the average temperature; For each cow, calculate the maximum difference in average temperature between any two crown band regions, normalize it, and use it as the first evaluation value for each cow. In this embodiment, the maximum absolute value of the difference between the average temperatures of any two coronary band regions is calculated and normalized, and this result is used as the first evaluation value for each cow. Next, the maximum-minimum normalization method is used for normalization processing. The maximum-minimum normalization method is a well-known technique and will not be described in detail here.

[0031] It should be noted that a higher local temperature indicates a higher temperature of the hoof corresponding to the coronary band area in that frame; a higher average temperature indicates a higher average temperature of the coronary band area in the overall video, which may indicate persistent inflammation or other pathological changes; a higher first evaluation value indicates a greater temperature difference between the four coronary band areas of the cow, a poorer temperature uniformity among the four hooves, reflecting a more severe inflammation or infection of the hoof, and a higher probability of early lameness.

[0032] This gives the first assessment value for each cow.

[0033] Step 3: For visible light video, the spinal curvature is quantified to assess the abnormality of hunched posture by using the spatial distribution characteristics of key points of the spine in each frame, and a second evaluation value is obtained for each cow.

[0034] Furthermore, at the behavioral level of cattle, considering the degree of arching during gait, the more severe the inflammation and the more intense the pain, the more subconsciously the cattle will reduce the load on the affected hoof and shift their center of gravity to the healthy side by adjusting their body posture. This protective mechanism leads to abnormal curvature of the spine, resulting in arching of the back. Therefore, by analyzing whether there is curvature in the spinal posture represented by key points of the spine, a second assessment value is calculated, specifically: For each frame of the visible light video, the tailbone point and the neck point are connected by a line, and the distance from the tailbone point to the neck point is recorded as the line distance. The mean of the line distances of all frames of the visible light video is calculated as the average distance. In this embodiment, the distance is measured by calculating the Euclidean distance from the coccyx point to the neck point. The Euclidean distance is a well-known technique and will not be described in detail here.

[0035] Calculate the vertical distance from each midpoint to the connecting line in each frame, and select the maximum vertical distance among all frames of the visible light video. It should be noted that the calculation process for the distance from a point to a line is a well-known technique and will not be elaborated here.

[0036] The second assessment value for each cow is positively correlated with the maximum vertical distance, but negatively correlated with the average distance; It should be noted that a positive correlation means that the dependent variable increases as the independent variable increases and decreases as the independent variable decreases; a negative correlation means that the dependent variable decreases as the independent variable increases and increases as the independent variable decreases.

[0037] In this embodiment, the ratio of the maximum vertical distance to the average distance is used as the second evaluation value for each cow.

[0038] It should be noted that the smaller the average distance, the more the cow's overall body posture is relatively hunched and its spinal baseline is short when walking, which may be due to long-term chronic pain leading to habitual arching of the back. The larger the maximum vertical distance, the larger the vertical distance from the midpoint to the line connecting the two points during the cow's walk, reflecting that the cow exhibits a significant arching behavior when walking. The larger the obtained second assessment value, the more significant the arching abnormality is, reflecting that the cow is more likely to have an arched posture due to pain, and is more likely to develop lameness later.

[0039] This yields the second assessment value for each cow.

[0040] Step 4: Based on the key points of the cow's hoof in different frames, evaluate the periodic changes in the contact between the cow's hoof and the deviation of the hoof posture direction to assess gait abnormality, determine the third evaluation value for each cow, and combine the first and second evaluation values ​​to determine the evaluation coefficient for each cow, so as to assess and identify early lameness of the cow's hoof.

[0041] Accordingly, at the behavioral level of cattle, the coordination of their gait is further considered. More severe inflammation leads to more intense pain and a more disordered gait, which is more likely an early sign of lameness. Therefore, a third assessment value is calculated by analyzing the incoordination of the cattle's gait.

[0042] First, we analyze the gait and walking posture of cattle, specifically: For visible light video, the ordinate of the center point of the coronary band region corresponding to each hoof in each frame is extracted. The frame sequence and ordinate of the same hoof in each frame are combined into a two-dimensional array. Curve fitting is performed on all two-dimensional arrays corresponding to each hoof. The frame sequence corresponding to the trough on the fitted curve is defined as the landing frame sequence of each hoof, and the frame sequence corresponding to the first trough is defined as the first landing frame sequence of each hoof. In this embodiment, the least squares method is used for curve fitting, which is a well-known technique and will not be described in detail here. Secondly, the AMPD (Automatic multiscale-based peak detection) algorithm is used to obtain the troughs, which is a well-known technique and will not be described in detail here.

[0043] It should be noted that the fitted curve shows a trend of rising and falling, rising and falling again over time, and continuously changing. The lowest point on the fitted curve is the moment when the cow's hoof touches the ground.

[0044] Extract the unit direction vector between the position coordinates of two key points on the hoof corresponding to any coronary band region in each frame; In this embodiment, the two key points on each hoof refer to the hoof tip and the center point of the coronary band region of each hoof. The direction vector is the vector pointing from the hoof tip to the center point of the coronary band region. The magnitude of the direction vector is calculated, and the ratio of each component of the direction vector to the magnitude of the direction vector is used as the unit direction vector.

[0045] Collect visible light videos of multiple healthy cows walking, iterate through the visible light videos of each healthy cow, and obtain the sequence of all its landing frames. In this embodiment, visible light videos of 100 healthy cows walking are collected. As another implementation method, the implementer can set it according to the actual situation.

[0046] For each healthy cow, the average time interval between all two adjacent grounding frames of each cow's hoof in the visible light video is calculated as the average interval of each hoof for each healthy cow. The average interval of the same hoof in all healthy cattle is taken as the standard gait cycle for each hoof. For the visible light video of each healthy cow, extract the unit direction vector between the coordinates of the two key points on each hoof in the frame corresponding to the landing frame sequence; calculate the mean vector of the unit direction vector of each hoof in all frames corresponding to the landing frame sequence, and use it as the average vector of each hoof under each healthy cow; calculate the mean vector of the average vector of the same hoof in all healthy cows, and use it as the standard landing vector of each hoof. It should be noted that the mean vector is the average of all values ​​in the same dimension of all unit direction vectors. The vector formed by the average values ​​of all dimensions is the mean vector. Among them, the standard gait cycle reflects the gait characteristics of a healthy herd, and the standard landing vector reflects the landing posture characteristics of a healthy herd.

[0047] Secondly, for the four hooves of a normal cow, taking the left forehoove as an example, the time it takes to lift and lower the hoof during walking is roughly the same and has a stable periodicity. By quantifying this periodicity, after each cycle, the left forehoove will repeat the state before the cycle began; that is, from the moment the left forehoo lands, after one cycle, the left forehoo will land on the ground again. If a cow's left forehoove exhibits an early lameness posture, this periodicity will be disrupted. After one cycle, the left forehoove may not land on the ground as it normally would, or only the tip of the hoof may just touch the ground. At this point, the hoof's state differs from its ground-landing state. Specifically, under normal circumstances, when the hoof is suspended in the air, it will naturally move forward and fully extend to land on the ground. However, during lameness, due to pain or discomfort, the hoof may be pulled backward by the ankle, and the hoof may only partially contact the ground, or it may not land properly at all. Therefore, by comparing the differences in the walking posture of each cow's hoof under different gait cycles, a third evaluation value is calculated, specifically: For each cow's visible light video, starting with the first landing frame of each hoof, and using the standard gait cycle of that hoof as the time interval, all theoretical landing frames are obtained. The first... The first of the cow's hooves A theoretical landing frame sequence for: , For the first The first frame sequence of a cow's hoof landing. For the first The standard gait cycle of a cow's hoof. , indicating the number of cycles; in, A value of 0 indicates the first landing frame sequence. A value of 1 indicates the theoretical second landing frame sequence after one cycle. The value 2 indicates the theoretical third landing frame sequence after two cycles, and so on. ,in, This represents the overall frame order of the visible light video. For each cow's visible light video, the angle between the unit direction vector of each cow's hoof and the standard landing vector of the hoof in the corresponding frame of each theoretical landing frame sequence is calculated. The maximum angle of all cow hooves under all theoretical landing frame sequences is selected and normalized, and used as the third evaluation value for each cow. In this embodiment, the normalization process is as follows: the ratio of the maximum included angle to 180° is used as the third evaluation value; secondly, the calculation of the vector included angle is a well-known technique and will not be described in detail here.

[0048] It should be noted that the larger the included angle, the greater the difference between the hoof placement direction of the cow and the posture of a healthy herd. By adding a standard gait cycle to the first landing frame sequence to assess the posture difference, the disorder of the cow's hoof gait is reflected. Therefore, the larger the obtained third evaluation value, the more severe the abnormal posture of the cow's hoof at the critical moment of ground contact during walking, and the less healthy its gait.

[0049] Furthermore, based on the first evaluation value, the second evaluation value, and the third evaluation value, the evaluation coefficient is determined as follows: The first, second, and third evaluation values ​​are positively integrated to obtain the evaluation coefficient for each cow. In this embodiment, the calculation process of the evaluation coefficient is as follows: For each healthy cow, the first assessment value, second assessment value, and third assessment value are obtained according to the calculation process of the first assessment value, the second assessment value, and the third assessment value. Calculate the mean of the first assessment values ​​for all healthy cattle. and standard deviation According to the Raida criterion, the threshold interval is... Its upper limit value is used as the first threshold; Accordingly, the mean of the second assessment values ​​for all healthy cattle was calculated. and standard deviation According to the Raida criterion, the threshold interval is... Its upper limit is used as the second threshold; Calculate the mean of the third assessment values ​​for all healthy cattle. and standard deviation According to the Raida criterion, the threshold interval is... Its upper limit is used as the third threshold; in, For the first The evaluation coefficient of a cow For the first The first assessment value of a cow, The first threshold, For the first The second assessment value of the cow, The second threshold, For the first The third assessment value of a cow, The third threshold, , as well as These represent the preset weights. Represents a non-linear activation function; In this embodiment, The value is 0.3. The value is 0.3. The value is 0.4. As for other implementation methods, the implementer can set it according to the actual situation. Among them, nonlinear activation function Represented as: Wherein, the nonlinear activation function is the independent variable of the sigmoid function. Move 5 units to the left, while the range of values ​​for the sigmoid function is... Independent variable The range is any real number, when When the function value is 0.9933, at... As the function value approaches 1 infinitely, the value continues to increase. As the value increases, the function value rapidly transitions from a nonlinear region close to 0 to a saturation region close to 1, exhibiting a clear threshold-switching characteristic. Therefore, by introducing a constant of 5 to shift the function to the left, the input independent variable will change as long as the evaluated value exceeds the threshold. If the value is greater than 0, the function's output will approach 1, thus dynamically controlling the function's output.

[0050] It should be noted that the Raida criterion is a well-known technique and will not be elaborated upon here. Secondly, a larger evaluation coefficient indicates a greater deviation between the cow's gait and its normal state, reflecting a potential serious lameness problem. The flowchart for obtaining the evaluation coefficient provided in this embodiment is as follows: Figure 2 As shown.

[0051] Furthermore, based on the evaluation coefficient, lameness in cattle is assessed and identified, specifically as follows: If the evaluation coefficient of each cow is less than the preset first value, then the cow does not have early lameness; if the evaluation coefficient is greater than or equal to the preset first value and less than the preset second value, then the cow has a risk of lameness; if the evaluation coefficient is greater than or equal to the preset second value, then the cow has early lameness; wherein, the preset first value is less than the preset second value. In this embodiment, the first preset value is 0.3 and the second preset value is 0.6. As for other implementation methods, the implementer can set them according to the actual situation.

[0052] For cattle at risk of lameness, human intervention is conducted for examination and assessment, while for cattle with early-stage lameness, medical treatment is provided.

[0053] Based on the same inventive concept as the above methods, this application also provides an early identification system for bovine hoof lameness, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described methods for early identification of bovine hoof lameness.

[0054] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0055] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0056] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application, without departing from the content of the technical solution of this application, shall fall within the protection scope of the technical solution of this application.

Claims

1. A method for early identification of lameness in cattle, characterized in that, The method includes the following steps: Acquire visible light and infrared videos of each cow while it is walking; for the visible light videos, identify and segment the coronary band region, key points of the spine, and key points of the hooves frame by frame. For each cow's different hooves, the differences in temperature distribution in the corresponding coronary band area in infrared video were analyzed, the interhoof temperature difference was quantified to assess local physiological thermal anomalies, and the first assessment value for each cow was calculated. For visible light video, the spatial distribution characteristics of key points of the spine in each frame are used to quantify the spinal curvature to assess the abnormality of hunched posture and obtain a second evaluation value for each cow. Based on key points of the cattle hoof in different frames, the periodic changes in the contact between the cattle hoof and the deviation of the hoof posture direction are evaluated to assess gait abnormalities. A third evaluation value is determined for each cattle. The evaluation coefficient for each cattle is determined by combining the first and second evaluation values ​​to assess and identify early lameness of the cattle hoof.

2. The method for early identification of bovine lameness as described in claim 1, characterized in that, The key points of the spine include the coccyx point, the cervical point, and a number of intermediate points evenly selected between the two along the spinal line.

3. The method for early identification of bovine lameness as described in claim 1, characterized in that, The key points of the cow hoof include the tip of each hoof and the center point of the coronary band area.

4. The method for early identification of bovine hoof lameness as described in claim 1, characterized in that, The calculation of the first assessment value for each cow includes: For each cow's visible light video, based on each coronary band region in each frame, find its corresponding region in the same frame of the infrared video, and take the average temperature of all pixels contained in that region as the local temperature; calculate the average local temperature of each coronary band region across all frames as the average temperature. The first evaluation value is the result of normalizing the maximum difference in average temperature between any two crown band regions.

5. The method for early identification of bovine lameness as described in claim 2, characterized in that, The second assessment value obtained for each cow includes: For each frame of the visible light video, the tailbone point and the neck point are connected by a line, and the distance from the tailbone point to the neck point is recorded as the line distance. The mean of the line distances of all frames of the visible light video is calculated as the average distance. Calculate the vertical distance from each midpoint to the connecting line in each frame, and select the maximum vertical distance among all frames of the visible light video. The second evaluation value is positively correlated with the maximum vertical distance, but negatively correlated with the average distance.

6. The method for early identification of bovine hoof lameness as described in claim 3, characterized in that, The determination of the third assessment value for each cow includes: For visible light video, the ordinate of the center point of the coronary band region corresponding to each hoof in each frame is extracted. The frame sequence and ordinate of the same hoof in each frame are combined into a two-dimensional array. Curve fitting is performed on all two-dimensional arrays corresponding to each hoof. The frame sequence corresponding to the trough on the fitted curve is defined as the landing frame sequence of each hoof, and the frame sequence corresponding to the first trough is defined as the first landing frame sequence of each hoof. The unit direction vector between the coordinates of the corresponding positions of the two hoof key points on each hoof in each frame is obtained. Collect visible light videos of multiple healthy cows walking, analyze the periodic distribution of adjacent landing frames of the same cow hoof, and the average level of the unit direction vector in the corresponding frames of the landing frame sequence, and calculate the standard gait period and standard landing vector of each cow hoof. For each cow, starting from the first landing frame of each hoof, and using the standard gait cycle of that hoof as the time interval, all theoretical landing frames are obtained. The angle between the unit direction vector of each hoof and the standard landing vector of that hoof in the corresponding frame of each theoretical landing frame is calculated. The maximum angle is selected and normalized, and used as the third evaluation value for each cow. Among them, the The first of the cow's hooves A theoretical landing frame sequence for: , For the first The first frame sequence of a cow's hoof landing. For the first The standard gait cycle of a cow's hoof. Indicates the number of cycles.

7. The method for early identification of bovine hoof lameness as described in claim 6, characterized in that, The process of obtaining the standard gait cycle and standard landing vector is as follows: traverse the visible light video of each healthy cow to obtain all its landing frame sequences and the unit direction vector of each hoof; calculate the average time interval between all two adjacent landing frames for each hoof of each healthy cow in the visible light video, as the average interval; calculate the average interval of the same hoof among all healthy cows as the standard gait cycle of each hoof; calculate the average unit direction vector of each hoof of each healthy cow in all landing frame sequences, as the average vector; calculate the average vector of the same hoof among all healthy cows as the standard landing vector of each hoof.

8. The method for early identification of bovine hoof lameness as described in claim 1, characterized in that, The evaluation coefficient is the result of a positive fusion of the first evaluation value, the second evaluation value, and the third evaluation value.

9. The method for early identification of bovine hoof lameness as described in claim 1, characterized in that, The assessment and identification of early lameness in cattle hooves includes: if the assessment coefficient of each cow is less than a preset first value, the cow does not have early lameness in its hooves; if the assessment coefficient is greater than or equal to the preset first value and less than a preset second value, the cow has a risk of lameness in its hooves; if the assessment coefficient is greater than the preset second value, the cow has early lameness in its hooves; wherein the preset first value is less than the preset second value.

10. An early detection system for bovine hoof lameness, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for early identification of bovine hoof lameness as described in any one of claims 1-9.

Images