Low-yield laying hen detection system and method based on multi-target behavior tracking

Abstract

The invention discloses a low-yield laying hen detection system and method based on multi-target behavior tracking. The low-yield laying hen detection system comprises a behavior action device and a behavior detection device, laying hens perform behavior actions through the behavior action device, and behavior video frames of the behavior actions of the laying hens are obtained through the movably or fixedly installed behavior action device; the behavior video frames are preprocessed and then processed through a laying hen behavior recognition model and a laying hen behavior tracking model to obtain a laying hen behavior tracking data set, and then low-yield laying hens are detected after being processed in the host through a laying hen behavior detection method and displayed on a displayer. According to the method, the breeding management precision and efficiency can be improved, resource utilization can be optimized, intelligent decision support can be carried out, an efficient and accurate low-yield laying hen detection method is realized, the laying hen production efficiency can be improved, and automation and intelligence of the breeding industry are promoted.

Description

Technical Field

[0001] The present invention relates to a detection method for laying hens, belonging to the field of poultry breeding industry, and particularly to a low-yield laying hen detection system and method based on multi-object behavior tracking. Background Art

[0002] With the rapid development of the poultry breeding industry, the breeding scale of laying hens has been continuously expanding, putting forward higher requirements for the automation and refinement of breeding management. Traditional low-yield laying hen detection methods mainly rely on manual observation, which has problems such as low efficiency, poor accuracy, and causing stress to chickens, and it is difficult to meet the needs of modern breeding industry. In recent years, automated detection methods based on computer vision and machine learning have gradually attracted attention, but there are still deficiencies in the accuracy of caged chicken behavior recognition and the effectiveness of classification models in the existing technologies. Therefore, there is an urgent need for an automated detection method that can accurately identify and distinguish low-yield laying hens to improve the efficiency of breeding management. Summary of the Invention

[0003] In order to solve the problems existing in the background art, the present invention provides a low-yield laying hen detection system and method based on multi-object behavior tracking. The present invention can improve the efficiency of chicken coop management and resource optimization through automated and intelligent means.

[0004] The technical solution adopted by the present invention is as follows: I. A low-yield laying hen detection system based on multi-object behavior tracking, comprising: A number of behavior action devices, which are respectively installed in each chicken cage of the chicken coop and are used for each laying hen to perform behavior actions.

[0005] A behavior detection device, which is installed on the side of each chicken cage in the chicken coop in a movable or fixed manner, and is used to detect low-yield laying hens in a manner of moving inspection or fixed detection according to the behavior actions of each laying hen at the behavior action device. A low-yield laying hen specifically refers to a laying hen with a monthly egg production lower than a preset threshold.

[0006] The described behavior acting device includes a feeding trough and a drinking fountain. The feeding trough is installed at the lower outer side of the chicken coop with its trough opening facing upwards, and the drinking fountain is installed at the upper outer side of the chicken coop with its outlet facing downwards. A number of laying hens are placed in each chicken coop; The behavior detection device includes a high-definition camera, an LED lamp tube, a camera bracket, a chassis, a display, a switch, a hard disk video recorder, and a host. The chassis is movably or fixedly arranged on the ground between each chicken coop. The camera bracket is installed on the chassis, and the high-definition camera is installed at the top of the camera bracket and oriented towards the positions of the feeding trough and the drinking fountain. The LED lamp tube is installed on the casing of the high-definition camera with its light source oriented towards the positions of the feeding trough and the drinking fountain; Multiple high-definition cameras are electrically connected to the hard disk video recorder through network cables and a switch. The hard disk video recorder, the host, and the display are interconnected. The switch, the hard disk video recorder, and the high-definition camera form a chicken behavior video acquisition platform. The switch and the hard disk video recorder are powered by direct current, and the high-definition camera is powered by POE (Power over Ethernet). The behavior video data stream is output through the hard disk video recorder to the chicken behavior tracking and statistics platform.

[0007] II. A low-yield laying hen detection method based on multi-object behavior tracking, including: Step 1: Construct a behavior recognition model and a behavior tracking model and install them in the host.

[0008] Step 2: Use the high-definition camera to collect several behavior video frames of laying hens performing behavior actions in the behavior acting device, and then sequentially input them into the host through network cables, switches, and hard disk video recorders for preprocessing and behavior annotation processing to construct a caged laying hen behavior dataset. Use the caged laying hen behavior dataset to train the behavior recognition model and the behavior tracking model to obtain a trained laying hen behavior recognition model and a trained laying hen behavior tracking model.

[0009] Step 3: Use the high-definition camera to collect several behavior video frames of each laying hen in the chicken coop to be detected performing behavior actions in the behavior acting device. After performing the same preprocessing as in Step 2, input them into the laying hen behavior recognition model and the laying hen behavior tracking model in sequence for processing and output a laying hen behavior tracking diagram. Each laying hen behavior tracking diagram is constructed into a laying hen behavior tracking dataset.

[0010] Step 4: Use the laying hen behavior detection method in the host to process the laying hen behavior tracking dataset, detect low-yield laying hens, and display them on the display to complete the low-yield detection.

[0011] In the above-mentioned step 1, the behavior recognition model adopts the improved object detection YOLOV11n ("You Only Look Once 11n-oriented bounding box") model. The improved object detection YOLOV11n model connects the backbone network and each feature extraction C3k2 module in the detection head of the object detection YOLOV11n model, and then connects the iterative attention feature fusion iAFF (Iterative Attention Feature Fusion) module, so as to be reconstructed into a feature extraction and fusion C3k2_iAFF composite module. The loss function of the improved object detection YOLOV11n model adopts the loss function WloU_v1 with a dynamic non-monotonic focusing mechanism. During training, until the loss function WloU_v1 converges, a trained laying hen behavior recognition model is obtained. The behavior tracking model adopts the multi-object tracking ByteTrack model.

[0012] In the above-mentioned step 2, the behavior actions of the laying hens in the behavior action device include the action of lowering the head to feed in the feeder and the action of raising the head to drink water using the waterer. For each behavior video frame, the behavior video frame is preprocessed. First, the behavior video frames with laying hens are screened out, and then adjusted to a preset contrast and preset size, and then normalized to obtain a normalized image to improve the image quality. Then, the behavior of the laying hens in the normalized image is labeled using the Labelimg tool. When the head of the laying hen is in the feeder, it is marked as the feeding action. At this time, the lower edge of the beak of the laying hen crosses the upper edge of the feeder and has an obvious downward movement trend. When the beak of the laying hen is less than the preset pixel distance from the outlet of the waterer, it is marked as the drinking action. At this time, the beak of the laying hen is no more than 50 pixels away from the nipple waterer and the chicken head has a tendency to move obliquely upward. Other actions that are not feeding or drinking are marked as other actions, thus constructing a dataset with rich characteristics and accurate annotation information, providing high-quality training and verification data for the automatic recognition of laying hen behaviors. The preprocessing process and the two models form a chicken behavior tracking and statistics platform.

[0013] The above-mentioned step 4 is as follows: Step 4.1: For each laying hen in the continuous N frames of laying hen behavior tracking diagrams within the preset total time period, use the laying hen behavior detection method for processing. Regard the feeding, drinking, and feeding-coupled drinking behaviors of the laying hens as state detection behaviors. For each state detection behavior, first, according to the behavior trajectory in the continuous N frames of laying hen behavior tracking diagrams of the laying hens, obtain the frame ratio of the state detection behavior of the laying hens, and obtain the average speed of the state detection behavior of the laying hens.

[0014] Step 4.2: Establish a comprehensive behavior intensity model, input the number of frames of the state detection behavior of the laying hens into the comprehensive behavior intensity model, and output the comprehensive behavior intensity index after processing.

[0015] Step 4.3: Establish a low-yield laying hen fusion detection model, normalize the proportion of the number of frames, average speed, and comprehensive behavior intensity index of the state detection behavior of the laying hens, input them into the low-yield laying hen fusion detection model, output the low-yield laying hen detection parameters after processing, and detect and realize the laying hens according to the low-yield laying hen detection parameters.

[0016] In the above-mentioned step 4.1, the proportion of the number of frames of the state detection behavior of the laying hens is the proportion of the total number of frames of the laying hen behavior tracking diagram of the laying hens performing the feeding action in the continuous N number of frames of the laying hen behavior tracking diagram P e , the proportion of the total number of frames of the laying hen behavior tracking diagram of the laying hens performing the drinking action in the continuous N number of frames of the laying hen behavior tracking diagram P d , or the proportion of the feeding coupled with drinking action P c , P c = P e + P d .

[0017] In the above-mentioned step 4.1, when obtaining the average speed of the state detection behavior of the laying hens, first, according to the behavior trajectory in the continuous N number of frames of the laying hen behavior tracking diagram, obtain the instantaneous moving speed of the feeding or drinking behavior of the laying hens with each n frame as the unit time window, and obtain the average value of the instantaneous moving speeds of each feeding behavior, drinking behavior, or feeding coupled with drinking behavior of the laying hens within the preset total time period as the average speed; the instantaneous moving speed of the feeding coupled with drinking behavior includes the instantaneous moving speed of the feeding behavior and the instantaneous moving speed of the drinking behavior.

[0018] In the above-mentioned step 4.2, the comprehensive behavior intensity model is as follows: S = λ h I h + λ m I m + λ l I l Ih = N h / N , T > T h I m = N m / N , T l ≤ T ≤ T h I l = N l / N , T < T l Among them, S is the comprehensive behavior intensity index; I h , I m and I l are the high, medium, and low intensity weighting parameters respectively; I h , I m and I l are the high, medium, and low intensity time indices respectively; T is the duration of a single state detection behavior, T h and T l are the first and second duration thresholds respectively; N h , N m and N l are the total number of frames in which the duration of a single state detection behavior is higher than the first duration threshold T h , between the first duration threshold T l and the second duration threshold T h and lower than the second duration threshold T l respectively.

[0019] In step 4.3 described above, the low - laying hen fusion detection model is as follows: Q = w 1 P n + w 2 S n + w 3 V n Among them, Q is the detection parameter for low - laying hens; w 1 、 w 2 and w 3 are the frame weight, speed weight, and intensity weight respectively; P n 、 S n and V n are the proportion of frames of the state detection behavior of the normalized laying hens, the comprehensive behavior intensity index, and the average speed respectively.

[0020] In the current state detection behavior, if the low - laying hen detection parameter is not lower than the preset threshold, it is judged as a non - low - laying hen, otherwise it is a low - laying hen.

[0021] The beneficial effects of the present invention are: 1. The present invention can improve the accuracy and efficiency of breeding management: By combining an automated data acquisition platform, multi - target behavior tracking technology, and a deep learning model, the present invention realizes the precise detection of low - laying hens. Compared with traditional manual methods, the system significantly improves the accuracy and efficiency of breeding management, reduces human error and labor intensity, and has a low detection cost and high detection efficiency.

[0022] 2. The present invention can optimize resource utilization: By intelligently identifying low - laying hens, the present invention can achieve early detection and elimination of low - producing individuals, avoid resource waste, and optimize the structure and production efficiency of the chicken flock. It reduces the feed consumption, space occupation, and environmental pressure of low - laying hens, improves the utilization efficiency of resources, and thus promotes the development of the farm towards a more efficient, environmentally friendly, and sustainable model.

[0023] 3. The present invention can provide intelligent decision-making support: The present invention provides intelligent decision-making support. The system can analyze the behavioral characteristics of low-producing laying hens in real time and automatically put forward management optimization suggestions according to the detection results. For example, after detecting low-producing laying hens, the system can automatically adjust management strategies such as feeding and feed delivery, optimize the production plan, and provide decision-making references for breeders. The intelligent decision-making mechanism not only improves the management level of the farm but also enhances the scientificity and flexibility of breeding decisions, further improving production efficiency and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 is a schematic structural diagram of the system of the present invention; Figure 2 is a training result diagram of the behavior recognition model of the present invention; Figure 3 is a flow chart of the behavior tracking model of the present invention; Figure 4 is a general flow chart of the present invention; Figure 5 is a flow chart of behavior statistics and output of the present invention; In the figure: 1, chicken coop; 2, feeding trough; 3, drinking fountain; 4, high-definition camera; 5, LED lamp tube; 6, camera bracket; 7, chassis; 8, network cable; 9, display; 10, switch; 11, digital video recorder. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0026] As Figure 4 shown, the present invention first enters the platform construction and dataset establishment stage, as Figure 1As shown in the figure, a low-yield laying hen detection system based on multi-object behavior tracking including a chicken behavior video acquisition platform is built. The low-yield laying hen detection system includes a number of behavior action devices and behavior detection devices. Each behavior action device is respectively installed in each chicken cage 1 in the chicken house and is used for each laying hen to perform behavior actions. The behavior detection device is movably or fixedly installed on the side of each chicken cage 1 in the chicken house, and detects low-yield laying hens in a way of mobile patrol or fixed detection according to the behavior actions of each laying hen at the behavior action device. The low-yield laying hen is specifically a laying hen with a monthly egg production lower than a preset threshold. The behavior action device includes a feeding trough 2 and a waterer 3. The feeding trough 2 is installed at the lower part outside the chicken cage 1 with the trough opening facing upwards, and the waterer 3 is installed at the upper part outside the chicken cage 1 with the outlet facing downwards. A number of laying hens are placed in each chicken cage 1. The behavior detection device includes a high-definition camera 4, an LED lamp tube 5, a camera bracket 6, a chassis 7, a monitor 9, a switch 10, a hard disk video recorder 11 and a host. The chassis 7 is movably or fixedly arranged on the ground between each chicken cage 1. The camera bracket 6 is installed on the chassis 7. The high-definition camera 4 is installed at the top of the camera bracket 6 and faces the positions of the feeding trough 2 and the waterer 3. The LED lamp tube 5 is installed on the casing of the high-definition camera 4 and the light source faces the positions of the feeding trough 2 and the waterer 3. A plurality of high-definition cameras 4 are electrically connected to the hard disk video recorder 11 through a network cable 8 and a switch 10. The hard disk video recorder 11, the host and the monitor 9 are connected to each other. The switch 10, the hard disk video recorder 11 and the high-definition camera 4 form a chicken behavior video acquisition platform. The switch 10 and the hard disk video recorder 11 are powered by direct current, and the high-definition camera 4 is powered by POE (Power over Ethernet). The behavior video data stream is output from the hard disk video recorder 11 to the chicken behavior tracking and statistics platform. When installing the high-definition camera 4, if it is necessary to detect the laying hens in the second-layer chicken cage 1, first manually adjust the high-definition camera 4 to an appropriate height according to the height of the second-layer chicken cage 1, and then adjust the field of view range to the best angle by changing the depression angle of the high-definition camera 4. Use the preview function of the high-definition camera 4 to ensure that the chicken cage 1 is located in the center of the field of view, aligned with the feeding trough 2 and the waterer 3, and the chicken head can be photographed when the laying hen eats. During the adjustment process, ensure that the feeding trough 2 in the picture remains horizontal and is parallel to the long axis of the picture to improve the accuracy and efficiency of image processing. Before collecting the video, debug the high-definition camera 4, set the video encoding format to MJPG, the resolution aspect ratio to 3840×2160, and the frame rate to 20FPS to ensure the clarity and smoothness of the video. For the convenience of data storage, the video is compressed using the H.264 standard. The video data collected by the high-definition camera 4 is stored in the hard disk video recorder 11 through a high-definition multimedia interface HDMI cable, and the captured pictures can be browsed on the monitor 9 in the chicken house and video browsing software.

[0027] Based on the low-egg-production hen detection system, the present invention uses computer vision to focus on identifying two behaviors of hens, namely feeding and drinking, monitors the feeding situation, drinking situation, and the coupling situation of feeding and drinking of 200-day-old hens within 3 minutes after feeding, and finally compares the accuracy rates of three methods for identifying low-egg-production hens. The specific implementation manner of the low-egg-production hen detection method based on multi-object behavior tracking is as follows: First, construct a behavior recognition model and install it in the host computer. The behavior recognition model uses the improved object detection YOLOV11n model. The improved object detection YOLOV11n model connects the backbone network and each feature extraction C3k2 module in the detection head and then connects the iterative attention feature fusion iAFF module, thus reconstructing it into a feature extraction and fusion C3k2_iAFF composite module. The iterative attention feature fusion iAFF module includes two multi-scale channel attention modules. On the premise of retaining the original C3k2 basic structure, the C3k2_iAFF module embeds a dual-branch attention weight generation unit. The first branch uses a channel attention mechanism composed of global average pooling and a fully connected layer to perform channel dimension compression-activation operations on the input features, generating a channel enhancement coefficient matrix; the second branch extracts spatial context information through 3×3 depthwise separable convolution and generates a spatial focus heat map in combination with the Sigmoid function; the attention weights output by the two branches are dynamically weighted and fused through a learnable scale factor to form a final feature selection mask. After this mask is element-wise multiplied with the original features, it is input into the feature fusion gating unit, and is adaptively weighted and summed with the original path features through residual connection. Further, the improved method deploys four groups of C3k2_iAFF modules in the feature pyramid construction stage of the backbone network, which act on the P3-P5 level feature extraction paths respectively, uses C3k2_iAFF modules in the 2nd, 4th, 6th, and 8th layers of the backbone network, and deploys three groups of C3k2_iAFF modules in the multi-scale feature fusion stage of the detection head. Through the cascaded dynamic feature calibration mechanism, the network automatically optimizes the fusion ratio coefficient of cross-level features, the channel dimension feature enhancement coefficient matrix, and the spatial dimension feature focus weight during the training process, so as to significantly improve the feature representation ability for occluded targets, small-scale targets, and complex background interference on the premise of maintaining the original network depth and parameter efficiency; the present invention also proposes an adaptive bounding box regression loss function WloU_v1 that fuses the decoupled attention mechanism of distance and aspect ratio to address the over-punishment problem of the existing loss function for high-quality samples. As the loss function WloU_v1 of the dynamic non-monotonic focusing mechanism of the improved object detection YOLOV11n model, it is as follows: L WIoU =R WIoU L IoU R WIoU =exp((x - xgt ) 2 +(y - y gt ) 2 ) / (W g 2 + H g 2 ) L IoU = 1 - IoU Where L WIoU represents the loss function WloU_v1, R WIoU and L IoU represent the first and second calculated values respectively; exp( ) represents the exponential operation; x and x gt represent the abscissas of the predicted box and the ground truth box respectively, y and y gt represent the ordinates of the predicted box and the ground truth box respectively, W g and H g represent the length and width of the minimum quadrilateral enclosing the predicted box and the ground truth box respectively; IoU represents the Intersection over Union.

[0028] The loss function WloU_v1 first calculates the minimum enclosing size (W´, H´) of the predicted box and the target box, where W´ and H´ are the width and height of the enclosing size respectively, and separates it from the computational graph to eliminate gradient anomalies; subsequently, an exponential distance attention term exp[( △ x² + △ y²) / (W´ + H´)] is constructed based on the normalized center distance, △ x and △ y are the differences in the horizontal and vertical coordinates respectively, △ x = x - x gt , △ y = y - y gt , and then dynamically adjusts the geometric penalty weight: when the overlap between the anchor box and the target box is low, the distance sensitivity is enhanced to improve the localization accuracy, and when the overlap is high, the geometric constraint is weakened to reduce the overfitting risk for high-quality samples. The loss function WloU_v1 realizes a differential learning strategy for different quality samples by constructing a geometric constraint decoupler and a dynamic sensitivity regulator, and its core modules include a geometric constraint decoupled attention generation unit, a dynamic loss sensitivity adjustment unit, and a gradient collaborative optimization mechanism. The geometric constraint decoupled attention generation unit includes a distance attention factor and an aspect ratio attention factor, and makes the aspect ratio penalty decay significantly when the Intersection over Union (IOU) is high through the IOU correlation truncation threshold. The dynamic loss sensitivity adjustment unit constructs a dual attention gating weight G to dynamically fuse the distance and aspect ratio correlations. The gradient collaborative optimization mechanism designs a main optimization path and an auxiliary optimization path to ensure the dynamic stability of the attention mechanism.

[0029] When training the behavior recognition model until the loss function WloU_v1 converges, a trained laying hen behavior recognition model is obtained. When training the behavior recognition model, after preprocessing the collected images, the Labeling tool is used to carefully label the feeding and drinking behaviors of the laying hens. Then, the PyTorch deep learning framework and the Yolov11n model are used for training, and the model weights are updated through forward propagation and backward propagation to optimize the model performance. On the premise of retaining the original C3k2 basic structure, a dual-branch attention weight generation unit is embedded, and an adaptive bounding box regression loss function that fuses the distance and aspect ratio decoupled attention mechanism is used to improve the feature representation ability for occluded targets, small-scale targets, and complex background interference while maintaining the original network depth and parameter efficiency. Finally, various metrics of the training set and validation set are used to evaluate the model performance, including the loss curve, accuracy, recall rate, and the mean average precision mAP50 (mean Average Precision at 50% Intersection over Union) when the intersection over union IoU (Intersection over Union) reaches 50%, to ensure that the model reaches the expected accuracy level in the behavior recognition task.

[0030] To ensure the effectiveness and good generalization ability of model training, the collected 4,800-image dataset is divided into a training set, a validation set, and a test set in a ratio of 8:1:1. The categories of three behaviors are declared in the category annotation file, with a total of 29,334 annotation boxes, which are respectively labeled as class0, class1, and class2 to support the recognition function of the model. The PyTorch deep learning framework based on Python is adopted, and YOLOV11n is used as the basic architecture to construct a target detection and behavior recognition network model. Configure the model training environment, then initialize the model parameters, calculate the prediction output through forward propagation, and use the loss function to evaluate the difference between the prediction result and the true label. Subsequently, the improved YOLOV11n model is used to adjust the model weights to improve the overall performance of the model. At the end of each training cycle, the performance of the model on the validation set is evaluated, the early stopping strategy is implemented or the model checkpoint is saved, and the learning rate and other hyperparameters are dynamically adjusted according to the validation results to optimize the training process. During the training process, the recognition accuracy of the feeding and drinking behaviors is monitored to ensure that the model has the ability to effectively distinguish various behaviors. After training is completed, the generalization ability of the model is evaluated using the test set, and necessary fine-tuning is performed to prevent the model from overfitting and improve the performance of the optimized model in practical applications. In addition, to ensure the stability and accuracy of the model in practical applications, the training process needs to be monitored throughout to prevent overfitting or underfitting, and the training strategy is flexibly adjusted according to the training feedback.

[0031] After the behavior recognition model is trained, the model accuracy needs to be evaluated. Training and validation loss evaluation is required. During the model training process, key metrics such as bounding box loss, classification loss, and depth loss of the training set and validation set are recorded. As Figure 2 shown, the training results of the model are, in sequence, the bounding box loss, classification loss, depth loss, accuracy, recall rate of the training set, and the bounding box loss, classification loss, depth loss, accuracy, and recall rate of the validation set. Then, an analysis of precision and recall rate is carried out to evaluate the accuracy and recall rate recall of the bounding boxes of the model on the training set and validation set to measure the recognition accuracy and integrity of the model. It is also necessary to plot the mean average precision mAP curve, and plot the mean average precision mAP50 (Intersection over Union IoU threshold is 0.5) and mean average precision mAP50-95 (Intersection over Union IoU threshold range is from 0.5 to 0.95) curves of the model on the training set and validation set to comprehensively evaluate the performance of the model under different IoU thresholds. Through the comprehensive analysis of the above metrics, a preliminary evaluation of the recognition performance of the model is carried out to ensure that it reaches the expected precision level in the behavior recognition task.

[0032] Then, a behavior tracking model is built and installed in the host. The behavior tracking model uses the multi-object tracking ByteTrack model. Specifically, when implemented, the detection boxes with behavior labels output by the behavior recognition model are output to the ByteTrack model, and the number of behavior frames of each target is calculated. The ByteTrack model includes Kalman filter detection box prediction, IOU matching, Hungarian matching, and the creation, deletion, and merging of chicken trajectories. The multi-object tracking ByteTrack model will divide the detection boxes output by the behavior model into high-score detection boxes Pi > Thigh and low-score detection boxes Pi < Tlow according to two confidence thresholds Thigh and Tlow, and then use the Kalman filter to predict the motion state of the target and update it through the detection box to provide the accurate position and trajectory of the target. Then, calculate the overlap degree IOU between the current detection box and the target position box predicted by the Kalman filter, and then complete the best match through the Hungarian algorithm, and finally complete the creation, deletion, and merging of the trajectory. As Figure 3 shown, specifically as follows: 1.1. Create a tracking trajectory for each target before starting tracking, and then predict the next frame bounding box of each tracking trajectory through the Kalman filter, and obtain the detection box B of the target through the detector i , each detection box B i = (X i , Y i , W i , H i , P i ), where X i , Y i , Wi , H i , P i represent the horizontal and vertical coordinates of the center point of the prediction box, the length and width of the detection box, and the confidence of the behavior classification respectively. According to two confidence thresholds T high and T low , the detection boxes are divided into high-score detection boxes (P i > T high ) and low-score detection boxes (P i < T low ), where the high-score detection boxes are used for priority matching, and the low-score detection boxes are used for supplementary matching. The detection boxes with confidence levels between the confidence thresholds T low and T high are classified as low-score detection boxes and the background is filtered out.

[0033] First, for the high-score boxes, calculate the IOU between the high-score boxes and the prediction boxes, and use the Hungarian algorithm to match the IOU to obtain three results: the matched trajectories, the unmatched trajectories, and the unmatched detection boxes. After successful matching, update the boxes in the tracking trajectories to high-score detection boxes, and update the matched trajectories to the active state. Then, for the low-score boxes, calculate the IOU between the low-score boxes and the prediction boxes that were not matched in the previous step, and use the Hungarian algorithm to match the IOU to obtain three results: the matched detection boxes and trajectories, the unmatched trajectories, and the unmatched low-score detection boxes. After successful matching, update the boxes in the tracking trajectories to detection boxes. Finally, for the high-score detection boxes that were not matched, match them with the trajectories with inactive states to obtain three results: the matched trajectories, the unmatched trajectories, and the unmatched detection boxes. For the matched ones, update the status; for the unmatched trajectories, mark them for deletion; for the unmatched detection boxes, if the confidence is greater than the high threshold + 0.1, create a new tracking trajectory, otherwise discard them.

[0034] 1.2. State prediction and update of Kalman filter: Kalman filter is used to predict the motion state of the target and update it through the detection box to provide the accurate position and trajectory of the target.

[0035] 1.2.1. State prediction: At each frame, Kalman filter predicts the target state of the current frame based on the target state X k-1 of the previous frame and the state transition model.

[0036] 1.2.2. State update: When the detection box Z k is available, Kalman filter corrects the predicted state.

[0037] 1.3. IOU matching and Hungarian matching algorithm: Calculate the overlap degree IOU between the current detection box and the target position box predicted by Kalman filter as follows: IOU ( B i ,T j )=( B i ∩ T j ) / ( B i ∪ T j ) Among them, B i represents the current detection box, T j represents the target position box predicted by Kalman filter, B i =( w B , h B ) and T j =( w T , h T ) respectively represent the set of boundary information of the current detection box and the predicted target detection box, w B and h B respectively represent the width and height of the current detection box, w T and h T respectively represent the width and height of the predicted target detection box.

[0038] The cost matrix is obtained through IOU matching C ij , that is, the detection box B i in the current frame T j and the matching cost of the target

[0039] in the previous frame. The Hungarian algorithm completes the matching by minimizing the total cost value, as follows: min∑ i=1 D ∑ j=1 J C ij · X ij Among them, D represents the number of detection boxes in the current frame, that is, the total number of predicted targets generated by Kalman filter currently; J represents the number of un-matched tracking trajectories in the previous frame, that is, the total number of existing tracking trajectories to be associated with the current detection box.

[0040] The constraints are as follows: ∑ j=1 T X ij = 1, indicating that each detection box can be paired with at most one trajectory ∑ i=1 D X ij = 1, indicating that each trajectory can accept at most one prediction box After the matching is completed, update the ID of the target and generate a new trajectory.

[0041] 1.3.1. First IOU matching: Perform Kalman filtering prediction on the trajectories t_s (trajectories being currently tracked) and l_s (temporarily lost trajectories) in the trajectory pool to obtain the predicted positions of each trajectory in the current frame. Calculate the IOU matrix between the predicted boxes of the trajectories and the high-confidence detection boxes, and use the Hungarian algorithm based on the IOU matrix to complete the matching of the trajectories and the detection boxes. Finally, update the trajectory status: For the trajectories with successful matching, update the bounding box position, status, and confidence score of the trajectories, and put the matched trajectories into a_s (activated tracking trajectories) or r_s (trajectories that did not match in the previous frame but match a target in the current frame); for the unmatched trajectories, put these trajectories into l_s (missing trajectories) and wait for the next round of matching.

[0042] 1.3.2. Second IOU matching: Calculate the IOU matrix between the trajectories that were not matched in the first matching T r and the low-confidence detection boxes D l and complete the matching using the Hungarian algorithm after filtering by Tlow. Update the status and position of the successfully matched trajectories; directly delete the unmatched low-confidence detection boxes; mark the unmatched trajectories as Lost and add them to l_s (missing trajectories).

[0043] 1.4. Trajectory creation, deletion, and merging in the tracking model: 1.4.1. Create a new trajectory: If the confidence of an unmatched high-confidence detection box is higher than the threshold Thigh, call the trajectory activation activate method, initialize the unmatched high-confidence detection box as a new trajectory, assign it a unique trajectory ID, add it to the tracking list, allocate a new trajectory ID, create a new trajectory and add it to the activated tracking trajectories.

[0044] 1.4.2. Trajectory deletion: If the number of missing trajectories exceeds the maximum allowed value, call the trajectory removal mark_removed method, mark the trajectories that have not been matched for a long time as invalid status, remove the trajectory from the tracking list, and remove the trajectory from the trajectory pool.

[0045] 1.4.3. Trajectory merging: After all trajectories are updated, use the IOU threshold to determine whether there are duplicates between the tracked trajectories and the missing trajectories. If they are the same target, keep the trajectory with a longer duration and remove the shorter one.

[0046] 1.4.4. Trajectory output: Use the trajectories with the status of Tracked and active as the output trajectories of the current frame, and return the bounding boxes and their corresponding IDs.

[0047] During specific implementation, use the high-definition camera 4 to collect a number of behavioral video frames of laying hens performing behavioral actions in the behavioral action device, including the action of lowering the head to feed in the feeding trough 2 and the action of raising the head to drink water using the water dispenser 3; then sequentially input them into the host through the network cable 8, switch 10, and hard disk recorder 11 for preprocessing. For each behavioral video frame, first screen out the behavioral video frames with laying hens, then adjust them to the preset contrast and preset size, and then perform normalization processing to obtain a normalized image to improve the image quality. Then use the Labelimg tool to annotate the behaviors of the laying hens in the normalized image. When the head of the laying hen is in the feeding trough 2, it is marked as the feeding action. At this time, the lower edge of the chicken beak of the laying hen crosses the upper edge of the feeding trough 2 and has an obvious downward movement trend; when the chicken beak of the laying hen is less than the preset pixel distance from the outlet of the water dispenser 3, it is marked as the drinking action. At this time, the chicken beak of the laying hen is no more than 50 pixels away from the nipple water dispenser 3 and the chicken head has a tendency to move obliquely upward; mark other actions that are not feeding and drinking as other actions, thus constructing a dataset with rich features and accurate annotation information, providing high-quality training and verification data for the automatic recognition of laying hen behaviors. The preprocessing and other processes and the two models form a chicken behavior tracking and statistics platform. After processing, it is constructed into a caged laying hen behavior dataset, and the caged laying hen behavior dataset is used to train the behavior recognition model and the behavior tracking model to obtain the trained laying hen behavior recognition model and laying hen behavior tracking model.

[0048] Then use the high-definition camera 4 to collect a number of behavioral video frames of each laying hen in the chicken cage 1 to be detected performing behavioral actions in the behavioral action device. After performing the same preprocessing in step 2, input them into the laying hen behavior recognition model and the laying hen behavior tracking model in sequence, and then output the laying hen behavior tracking map. The behavior tracking model can finally obtain the behavior trajectories of the laying hens, and each laying hen behavior tracking map is constructed into a laying hen behavior tracking dataset.

[0049] As Figure 5As shown, finally, the behavior statistics and output of laying hens are carried out. The laying hen behavior tracking dataset output by the chicken behavior tracking and statistics platform is further input into the low-yield laying hen detection platform based on behavior tracking for detection. In the host, the laying hen behavior detection method is used to process the laying hen behavior tracking dataset. For each laying hen in the continuous N frames of laying hen behavior tracking diagrams within the preset total time period, the laying hen behavior detection method is used for processing. The feeding, drinking, and feeding-coupled drinking behaviors of laying hens are all regarded as state detection behaviors. For each state detection behavior, first, according to the behavior trajectories in the continuous N frames of laying hen behavior tracking diagrams, the frame ratio of the state detection behavior of the laying hen is obtained. The frame ratio of the state detection behavior of the laying hen is the ratio of the total number of frames of the laying hen behavior tracking diagrams in which the laying hen performs feeding actions to the number of frames in the continuous N frames of laying hen behavior tracking diagrams P e , the ratio of the total number of frames of the laying hen behavior tracking diagrams in which the laying hen performs drinking actions to the number of frames in the continuous N frames of laying hen behavior tracking diagrams P d , or the ratio of the feeding-coupled drinking actions P c , P c = P e + P d ; and the average speed of the state detection behavior of the laying hen is obtained. When obtaining the average speed of the behavior trajectory of the state detection behavior of the laying hen, first, according to the behavior trajectories in the continuous N frames of laying hen behavior tracking diagrams, the instantaneous moving speed of the feeding or drinking behavior of the laying hen is obtained with every n frames as a unit time window. The average value of the instantaneous moving speeds of each feeding behavior, drinking behavior, or feeding-coupled drinking behavior of the laying hen within the preset total time period is used as the average speed; the instantaneous moving speed of the feeding-coupled drinking behavior includes the instantaneous moving speed of the feeding behavior and the instantaneous moving speed of the drinking behavior. It is necessary to calculate the instantaneous moving speed during the feeding or drinking behavior through the coordinates of the center point of the chicken body output by the behavior tracking model, with every 10 frames as a time window. During the 3-minute detection period, the instantaneous speed sequences of the chicken in the feeding and drinking behaviors are respectively statistically analyzed and the average speed of the behavior is calculated.

[0050] Then, a comprehensive behavior intensity model is established as follows: S = λ h I h + λ mI m + λ l I l I h = N h / N , T > T h I m = N m / N , T l ≤ T ≤ T h I l = N l / N , T < T l Among them, S is the comprehensive behavior intensity index; I h , I m and I l are the high, medium, and low intensity weighting parameters respectively; I h , I m and I l are the high, medium, and low intensity time indices respectively; T is the duration of a single state detection behavior, T h and T l are the first and second duration thresholds respectively; N h , N m and N l are respectively the duration of a single state detection behavior higher than the first duration threshold within the preset total time period T h , within the first duration threshold T l and the second duration threshold T hand below a second duration threshold T l The total number of frames, that is, when the laying hens perform 3 high-intensity feeding behaviors exceeding 120 s within a preset total time period, the total number of frames obtained for the three feeding behaviors is N h ; In specific implementation λ h = 0.6, λ m = 0.3 , λ l = 0.1, T h = 120 s, T l = 60 s.

[0051] For feeding and drinking behaviors, the duration T of a single behavior can be divided into high intensity I h , medium intensity I m and low intensity I l . For the feeding or drinking behaviors within each monitoring period, calculate the proportion of the number of frames in each intensity level respectively, and finally use a weighted formula to fuse the proportion of the number of frames in different intensity levels. Input the number of frames of the state detection behavior of the laying hens into the comprehensive behavior intensity model, and output the comprehensive behavior intensity index after processing S .

[0052] Finally, conduct behavior feature comparison and recognition evaluation. Apply the data such as the proportion of the number of behavior frames calculated above to the groups of low-producing laying hens and normal laying hens respectively, conduct comparative analysis of behavior features, and calculate the recognition accuracy of low-producing laying hens. The specific steps are as follows: 1. Behavioral feature fusion analysis: Compare the differences in behavioral features of low-producing laying hens and normal laying hens in three dimensions: the proportion of the number of behavior frames, comprehensive behavior intensity, and behavior trajectory speed through statistical methods, and select the weights of each index w 1 , w 2 , w 3 , and establish a fusion detection model for low-producing laying hens as follows: Q = w 1 P n + w 2 S n + w 3 Vn Among them, Q is the detection parameter for low-laying hens; w 1 , w 2 and w 3 are the frame weight, speed weight, and intensity weight respectively; P n , S n and V n are the proportion of frames of the state detection behavior of the normalized laying hens, the comprehensive behavior intensity index, and the average speed respectively. Perform data normalization calculation, linearly normalize the proportion of behavior frames, comprehensive behavior intensity, and average speed of the behavior trajectory, and normalize the values of the three indicators to [0,1].

[0053] After normalizing the proportion of frames of the state detection behavior of the laying hens, the average speed, and the comprehensive behavior intensity index, input them into the low-laying hen fusion detection model. After processing, output the low-laying hen detection parameter.

[0054] 2. Threshold setting: Detect laying hens according to the low-laying hen detection parameter. In the current state detection behavior, if the low-laying hen detection parameter is not lower than the preset threshold, it is judged as a non-low-laying hen, otherwise it is a low-laying hen. Finally, detect the low-laying hens and display them on the display 9. Specifically, by analyzing the behavior characteristic distribution of the low-laying and normal populations, optimize with the Receiver Operating Characteristic (ROC) curve to maximize the classification accuracy, and set the judgment thresholds for feeding, drinking, and feeding coupled with drinking respectively T e , T d and T c . Compare the data obtained by fusing different judgment thresholds and behavior characteristics. Those greater than or equal to the judgment threshold are normal laying hens, and those less than the judgment threshold are low-laying hens, thus completing the low-laying detection.

[0055] 3. Application of identification methods: Three identification methods (such as single feeding behavior determination method, single drinking water behavior determination method and feeding coupled drinking water behavior determination method) were used to identify low-producing laying hens and normal laying hens. A total of 3 chickens in 1 cage from 4 cages (1 low-producing and 2 high-producing) and 4 chickens in 1 cage from 4 cages (1 low-producing and 3 high-producing) were selected. A 3-minute video clip after feeding every day is taken out to represent a data. Feeding is done 4 times a day, so 1 chicken has 4 data in 1 day, for a total of 28 chickens (8 low-producing and 20 normal), 112 data in 1 day, and a total of 7 days of monitoring, for a total of 784 data. The final accuracy rate was obtained. A as follows: A =( N lc + N nc ) / N in, N lc and N nc They represent the number of hens correctly identified as low-laying hens and the number of hens correctly identified as normal laying hens, respectively. K Indicates the total amount of data tested, K =784.

[0056] By comparing the method of the present invention with the existing method, the overall monitoring results of the behavior of caged laying hens were obtained, as shown in Table 1.

[0057] Table 1

[0058] Among them, P and R represent precision and recall respectively.

[0059] It can be seen from Table 1 that the behavior recognition model constructed by the present invention has a high accuracy rate.

[0060] The present invention can also compare the accuracy of the detection results of low-laying chicken identification under the feeding, drinking, and feeding-coupled-drinking detection modes according to the divided thresholds, as shown in Table 2.

[0061] Table 2

[0062] It can be seen that the recognition accuracy of the three discrimination methods is high, and the recognition accuracy of the feeding coupled drinking behavior discrimination method is the highest.

[0063] The low-laying chicken detection system of the present invention is applied in the following scenarios: Model Integration and Automated Detection: Integrate the trained model into the monitoring system of the farm to achieve real-time automated detection. The system can process the behavior data of laying hens collected by high-definition cameras in real time and output detection results through a deep learning model, including the behavior status of each laying hen and the identification and location of low-producing laying hens. This information will help farm management personnel promptly identify low-producing laying hens and conduct targeted management.

[0064] Intelligent Analysis and Report Generation: The system can automatically generate an analysis report on low-producing laying hens based on the detection results. The report content includes key indicators such as the number, distribution, behavior characteristics, and behavior change trends of low-producing laying hens. Through a user-friendly interface, farm personnel can conveniently view and query these analysis results, thereby providing data support for management decisions. In addition, the system can also regularly generate an analysis report on the overall health status of the farm to help farm personnel optimize management strategies.

[0065] System Performance and Scalability: The system design takes into account the stability, compatibility, and scalability of performance. It can support a variety of data collection devices (such as cameras and sensors of different brands) and data interfaces (such as application programming interfaces API (Application Programming Interface), database connections, etc.), and can adapt to the needs of farms of different scales and environmental conditions. The system has a flexible modular architecture and can adjust the system configuration or add more monitoring points according to the scale of the farm. At the same time, the system reserves expansion interfaces to support the subsequent introduction of new data sources, algorithms, or identification methods to continuously improve the detection accuracy and efficiency.

[0066] Real-time Warning and Intelligent Decision Support: Based on the detection results of low-producing laying hens, the system can intelligently issue warning reminders. For example, when the behavior patterns of certain laying hens are abnormal or the number of low-producing laying hens exceeds the threshold, the system will automatically notify farm management personnel. In addition, the system can also adjust farming strategies according to real-time data, such as optimizing the feeding environment and adjusting feed delivery, so as to improve the overall production efficiency of laying hens.

[0067] The present invention deploys the trained laying hen behavior recognition model and laying hen behavior tracking model into the monitoring system of an actual farm to achieve real-time automated detection. It ensures that the system can process the behavior data collected by the high-definition camera 4 in real time and output the detection results. The system generates an analysis report for the detection of low-producing laying hens, including key indicators such as the number, distribution, and behavior characteristics of low-producing laying hens. It provides query and display functions through a user-friendly interface to facilitate decision-making and management by farm management personnel. It ensures that the system has good performance stability, compatibility, and scalability, supports a variety of data collection devices and interfaces, and adapts to the needs of farms of different scales and environments. The system reserves expansion interfaces for subsequent introduction of new data sources and recognition methods to further improve the detection accuracy. Based on the detection results of low-producing laying hens, the system can intelligently issue intelligent warning reminders to provide intelligent decision-making for farmers.

[0068] Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present application can be implemented in various computer languages. The present application is described according to the flowcharts of the methods, systems, and computer program products of the embodiments of the present application.

[0069] Although the preferred embodiments of the present application have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the present invention is intended to be construed as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0070] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the equivalent technology of the present invention, the present application is also intended to include these changes and modifications.

Claims

1. A low-laying chicken detection system based on multi-target behavior tracking, characterized in that: include: A plurality of behavioral action devices are respectively installed in each chicken cage (1) of the chicken house and are used for each laying hen to perform behavioral actions; The behavior detection device can be movably or fixedly installed on the side of each chicken cage (1) in the chicken house, and is used to detect low-laying chickens in a mobile inspection or fixed detection manner according to the behavioral actions of each laying hen at the behavior action device.

2. The low-laying hen detection system based on multi-target behavior tracking according to claim 1 is characterized in that: The behavior action device comprises a feed trough (2) and a waterer (3), wherein the feed trough (2) is installed at the lower outer part of the chicken cage (1) with the notch facing upward, and the waterer (3) is installed at the upper outer part of the chicken cage (1) with the outlet facing downward, and a plurality of laying hens are placed in each chicken cage (1); the behavior detection device comprises a high-definition camera (4), an LED light tube (5), a camera bracket (6), a chassis (7), a display (9), a switch (10), a hard disk recorder (11) and a host, and the chassis (7) can be movably or fixedly arranged at each On the ground between the chicken cages (1), a camera bracket (6) is mounted on a chassis (7), a high-definition camera (4) is mounted on the top of the camera bracket (6) and is positioned toward the feed trough (2) and the waterer (3), and an LED light tube (5) is mounted on the housing of the high-definition camera (4) and is positioned toward the feed trough (2) and the waterer (3); the high-definition camera (4) is electrically connected to a hard disk video recorder (11) via a network cable (8) and a switch (10), and the hard disk video recorder (11), a host computer, and a display (9) are connected to each other.

3. The method for detecting low-laying hens based on multi-target behavior tracking according to any one of claims 1-2, characterized in that: include: Step 1: Build a behavior recognition model and a behavior tracking model and install them in the host; Step 2: using a high-definition camera (4) to collect a number of video frames of laying hens performing behavioral actions in the behavioral action device, and then inputting the frames into a host computer via a network cable (8), a switch (10) and a hard disk recorder (11) for preprocessing and behavior labeling to construct a caged laying hen behavior data set, and using the caged laying hen behavior data set to train a behavior recognition model and a behavior tracking model to obtain a trained laying hen behavior recognition model and a laying hen behavior tracking model; Step 3: using a high-definition camera (4) to collect a number of behavior video frames of each laying hen in the chicken cage (1) to be detected performing a behavior action in the behavior action device, and after the same preprocessing as in step 2, inputting the frames into the laying hen behavior recognition model and the laying hen behavior tracking model in turn, and outputting a laying hen behavior tracking graph after processing, and each laying hen behavior tracking graph is constructed as a laying hen behavior tracking data set; Step 4: After processing the laying hen behavior tracking data set using the laying hen behavior detection method in the host, low-producing laying hens are detected and displayed on the display (9), thereby completing the low-production detection.

4. The low laying hen detection method based on multi-target behavior tracking according to claim 3 is characterized in that: In the step 1, the behavior recognition model adopts the improved target detection YOLOV11n model, and the improved target detection YOLOV11n model connects the backbone network in the target detection YOLOV11n model and each feature extraction C3k2 module in the detection head to the iterative attention feature fusion iAFF module, thereby reconstructing it into a feature extraction fusion C3k2_iAFF composite module; the loss function of the improved target detection YOLOV11n model adopts a loss function using a dynamic non-monotonic focusing mechanism, and the training is performed until the loss function converges to obtain a trained laying hen behavior recognition model; the behavior tracking model adopts a multi-target tracking ByteTrack model.

5. The low-laying chicken detection method based on multi-target behavior tracking according to claim 3 is characterized in that: In the step 2, the behavioral actions performed by the laying hen in the behavioral action device include lowering its head to feed in the feed trough (2) and raising its head to drink water from the waterer (3); for each behavioral video frame, the behavioral video frame is pre-processed, firstly the behavioral video frame containing laying hens is screened out, then adjusted to a preset contrast and a preset size, and then normalized to obtain a normalized image, and then the normalized image is labeled with the behavior of the laying hen, when the head of the laying hen is in the feed trough (2), it is marked as a feeding action; when the distance between the beak of the laying hen and the outlet of the waterer (3) is less than a preset pixel point distance, it is marked as a drinking action; other actions other than feeding and drinking are marked as other actions.

6. The low-laying chicken detection method based on multi-target behavior tracking according to claim 3 is characterized in that: The step 4 is as follows: Step 4.1: For the continuous N Each laying hen in the frame laying hen behavior tracking image is processed using the laying hen behavior detection method. The feeding, drinking, and feeding-coupled drinking behaviors of the laying hen are all used as state detection behaviors. For each state detection behavior, firstly, the continuous N Frame behavior trajectories in the laying hen behavior tracking graph are obtained to obtain the frame ratio of the laying hen's state detection behavior and the average speed of the laying hen's state detection behavior; Step 4.2: Establish a comprehensive behavior intensity model, input the frame number of the laying hen's state detection behavior into the comprehensive behavior intensity model, and output the comprehensive behavior intensity index after processing; Step 4.3: Establish a fusion detection model for low-laying chickens, normalize the frame rate ratio, average speed and comprehensive behavior intensity index of the laying hen's state detection behavior, and input them into the fusion detection model for low-laying chickens. After processing, output the low-laying chicken detection parameters, and detect laying hens based on the low-laying chicken detection parameters.

7. The method for detecting low laying hens based on multi-target behavior tracking according to claim 6 is characterized in that: In the step 4.1, the frame number of the laying hen's state detection behavior is the total number of frames of the laying hen behavior tracking image of the laying hen performing the feeding action in the continuous N The number of frames in the laying hen behavior tracking diagram P e The total number of frames of the laying hen behavior tracking image of the laying hen drinking water is continuous. N The number of frames in the laying hen behavior tracking diagram P d , or the proportion of feeding coupled with drinking P c , P c = P e + P d .

8. The method for detecting low laying hens based on multi-target behavior tracking according to claim 6, characterized in that: In step 4.1, when obtaining the average speed of the state detection behavior of the laying hen, firstly, according to the continuous N The behavior trajectory in the egg-laying chicken behavior tracking diagram is n The instantaneous movement speed of the laying hen's feeding or drinking behavior is obtained with the frame as the unit time window, and the average value of the instantaneous movement speed of each feeding behavior, drinking behavior or feeding-coupled-drinking behavior of the laying hen within the preset total time period is obtained as the average speed.

9. The method for detecting low laying hens based on multi-target behavior tracking according to claim 6, characterized in that: In step 4.2, the comprehensive behavior intensity model is as follows: S = λ h I h + λ m I m + λ l I l I h = N h / N , T > T h I m = N m / N , T l ≤ T ≤ T h I l = N l / N , T < T l in, S is the comprehensive behavior intensity index; I h , I m and I l are high, medium, and low intensity weighting parameters, respectively; I h , I m and I l They are high, medium and low intensity time indexes, respectively; T is the duration of a single state detection behavior, T h and T l are the first and second duration thresholds, respectively; N h , N m and N l The duration of a single state detection behavior within a preset total time period is higher than the first duration threshold T h , at the first duration threshold T l and a second duration threshold T h between and below the second duration threshold T l The total number of frames.

10. The low-laying chicken detection method based on multi-target behavior tracking according to claim 6 is characterized in that: In step 4.3, the low-laying egg-laying fusion detection model is as follows: Q = w 1 P n + w 2 S n + w 3 V n in, Q Testing parameters for low-producing laying hens; w 1. w 2 and w 3 are frame weight, speed weight and intensity weight respectively; P n , S n and V n They are the normalized frame percentage, comprehensive behavior intensity index and average speed of the state detection behavior of laying hens; Under the current state detection behavior, if the detection parameter of a low-laying hen is not lower than the preset threshold, it is judged as a non-low-laying hen, otherwise it is a low-laying hen.

Images