Method, device, equipment and storage medium for identifying health status of chicken

Abstract

The present invention provides a method, device, equipment and storage medium for identifying the health status of a chicken. The method comprises: obtaining a motion video containing a target chicken collected by a camera; inputting the motion video into a trained target tracking network to obtain a picture of the target chicken's area and the activity characteristics of the target chicken; inputting the picture of the target chicken's area into a trained chicken head detection network to obtain the color characteristics of the target chicken's comb and the depth characteristics of the area where the chicken's head is located; obtaining the temperature of the chicken's head in the area where the target chicken is located by an infrared thermometer matched with the camera as the temperature characteristics of the chicken's head; inputting the activity characteristics, comb color characteristics, depth characteristics of the area where the chicken's head is located and the temperature characteristics of the chicken's head into a trained classification network to obtain the health status of the target chicken. The present invention can improve the detection accuracy of sick chickens.

Description

Technical Field

[0001] The present invention relates to the field of computer recognition technology, and in particular to a method, device, equipment and storage medium for identifying the health status of chickens. Background Art

[0002] With the continuous development of the breeding industry and the increase in the density of livestock and poultry breeding, the number of caged chickens in a single cage is also increasing rapidly. In this case, high requirements are put forward for the disease prevention and control of livestock and poultry. Once livestock and poultry are sick or in a sub-healthy state, if they are not discovered and treated in time, it will cause great safety hazards and economic losses.

[0003] Take sick chickens as an example. Sick chickens have a group effect. If the sick chickens in cages are not identified and detected in time, they will quickly spread throughout the entire cage, causing large-scale illness of chickens, which will not only cause serious economic losses to the breeders, but also bring hidden dangers to public health such as food safety.

[0004] However, livestock and poultry farming mainly relies on manual visual inspections, which increases the workload of inspectors on the one hand, and on the other hand, sick chickens cannot be found in time due to negligence of personnel or too many livestock and poultry. At present, although there are intelligent recognition methods based on machine vision, the recognition accuracy is low and cannot meet the needs of livestock and poultry farming. Therefore, a method for identifying sick chickens with higher accuracy is urgently needed. Summary of the invention

[0005] The embodiments of the present invention provide a method, device, equipment and storage medium for identifying the health status of chickens, so as to solve the problem of low accuracy in identifying sick chickens.

[0006] In a first aspect, an embodiment of the present invention provides a method for identifying the health status of a chicken, comprising:

[0007] Obtaining a motion video containing a target chicken captured by a camera;

[0008] The motion video is input into the trained target tracking network to obtain the image of the target chicken area and the activity characteristics of the target chicken;

[0009] Input the image of the target chicken area into the trained chicken head detection network to obtain the color features of the target chicken's comb and the depth features of the chicken head area;

[0010] The temperature of the chicken head in the target chicken area is obtained by using an infrared thermometer matched with the camera as the temperature feature of the chicken head;

[0011] The activity characteristics, comb color characteristics, depth characteristics of the chicken head area, and temperature characteristics of the chicken head of the target chicken are input into the trained classification network to obtain the health status of the target chicken.

[0012] In a possible implementation, the image of the target chicken area is input into the trained chicken head detection network to obtain the color features of the target chicken's comb and the depth features of the chicken head area, including:

[0013] Input the image of the target chicken area into the trained chicken head detection network to obtain the color score of the target chicken's comb, and use the color score of the target chicken's comb as the target chicken's comb color feature;

[0014] Through the backbone feature extraction network of the chicken head detection network, the deep features of the image of the target chicken area are obtained, and the deep features are reduced in dimension to obtain the deep features of the target chicken's head area.

[0015] In a possible implementation, the color score of the target chicken's comb is:

[0016] score=color_probability×confidence;

[0017] Among them, color_probability is the possibility of the comb color of the target chicken, and the dimension is 4 dimensions; confidence is the confidence of whether there is an object in the area where the comb of the target chicken is located, and the dimension is 1 dimension.

[0018] In a possible implementation, the method further includes:

[0019] Acquire a first training sample, wherein the first training sample includes a motion video sample containing a chicken, each picture containing a chicken in the motion video sample corresponds to first annotation information, wherein the first annotation information indicates an area where the chicken is located in the picture;

[0020] The pre-constructed target tracking network is trained based on the first training sample and the first annotation information corresponding thereto to obtain a trained target tracking network.

[0021] In a possible implementation, the method further includes:

[0022] Obtain a second training sample, and input the second training sample into the trained target tracking network to obtain a picture sample, wherein the second training sample includes a motion video sample containing a chicken, and the picture sample is a picture of the area where the chicken is located cropped from the second training sample;

[0023] Obtaining second annotation information corresponding to the image sample through annotation, wherein the second annotation information indicates the color of the cockscomb in the image sample;

[0024] The pre-constructed chicken head detection network is trained based on the second training sample and its corresponding second annotation information to obtain a trained chicken head detection network.

[0025] In a possible implementation, the activity characteristic of the target chicken is the displacement change of the center point of the target chicken within a fixed time.

[0026] In a possible implementation, the temperature characteristic t of the target chicken's head is i for:

[0027]

[0028] Among them, T i is the head temperature of the ith chicken measured by infrared thermometer, T max is the maximum value of the chicken head temperature collected, T min It is the minimum value of the chicken head temperature collected.

[0029] In a second aspect, an embodiment of the present invention provides a device for identifying the health status of a chicken, characterized in that it includes:

[0030] An acquisition module is used to acquire a motion video containing a target chicken captured by a camera;

[0031] Obtain a first feature module, which is used to input the motion video into the trained target tracking network to obtain a picture of the area where the target chicken is located and the activity features of the target chicken;

[0032] Obtain the second and third feature modules, which are used to input the image of the target chicken area into the trained chicken head detection network to obtain the comb color feature of the target chicken and the depth feature of the chicken head area;

[0033] The fourth feature acquisition module is used to acquire the temperature of the chicken head in the area where the target chicken is located by using an infrared thermometer matched with the camera as the temperature feature of the chicken head;

[0034] The health status determination module is used to input the activity characteristics, comb color characteristics, depth characteristics of the chicken head area and temperature characteristics of the chicken head of the target chicken into the trained classification network to obtain the health status of the target chicken.

[0035] In a possible implementation, obtaining the second and third characteristic modules is specifically used for:

[0036] Input the image of the target chicken area into the trained chicken head detection network to obtain the color score of the target chicken's comb, and use the color score of the target chicken's comb as the target chicken's comb color feature;

[0037] Through the backbone feature extraction network of the chicken head detection network, the deep features of the image of the target chicken area are obtained, and the deep features are reduced in dimension to obtain the deep features of the target chicken's head area.

[0038] In a possible implementation, the color score of the target chicken's comb is:

[0039] score=color_probability×confidence;

[0040] Among them, color_probability is the possibility of the comb color of the target chicken, and the dimension is 4 dimensions; confidence is the confidence of whether there is an object in the area where the comb of the target chicken is located, and the dimension is 1 dimension.

[0041] In a possible implementation manner, obtaining the first feature module is further used for:

[0042] Acquire a first training sample, wherein the first training sample includes a motion video sample containing a chicken, each picture containing a chicken in the motion video sample corresponds to first annotation information, wherein the first annotation information indicates an area where the chicken is located in the picture;

[0043] The pre-constructed target tracking network is trained based on the first training sample and the first annotation information corresponding thereto to obtain a trained target tracking network.

[0044] In a possible implementation, obtaining the second and third feature modules is further used for:

[0045] Obtain a second training sample, and input the second training sample into the trained target tracking network to obtain a picture sample, wherein the second training sample includes a motion video sample containing a chicken, and the picture sample is a picture of the area where the chicken is located cropped from the second training sample;

[0046] Obtaining second annotation information corresponding to the image sample through annotation, wherein the second annotation information indicates the color of the cockscomb in the image sample;

[0047] The pre-constructed chicken head detection network is trained based on the second training sample and its corresponding second annotation information to obtain a trained chicken head detection network.

[0048] In a possible implementation, the activity characteristic of the target chicken is the displacement change of the center point of the target chicken within a fixed time.

[0049] In a possible implementation, the temperature characteristic t of the target chicken's head is i for:

[0050]

[0051] Among them, T i is the head temperature of the ith chicken measured by infrared thermometer, T max is the maximum value of the chicken head temperature collected, T min It is the minimum value of the chicken head temperature collected.

[0052] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the processor implements the steps of the method described in the first aspect or any possible implementation manner of the first aspect.

[0053] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described in the first aspect or any possible implementation of the first aspect are implemented.

[0054] The embodiment of the present invention provides a method, device, equipment and storage medium for identifying the health status of a chicken. First, a motion video containing a target chicken captured by a camera is obtained, and then the motion video is input into a trained target tracking network to obtain a picture of the area where the target chicken is located and the activity characteristics of the target chicken. After that, the picture of the area where the target chicken is located is input into a trained chicken head detection network to obtain the color characteristics of the target chicken's comb and the depth characteristics of the area where the chicken head is located. Next, the temperature of the chicken head in the area where the target chicken is located is obtained by an infrared thermometer matched with the camera as the temperature characteristics of the chicken head. Finally, the activity characteristics, comb color characteristics, depth characteristics of the area where the chicken head is located and temperature characteristics of the chicken head are input into a trained classification network to obtain the health status of the target chicken. In this way, by identifying the captured motion video containing the chicken, the health status of the chicken can be quickly and accurately identified. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0056] Figure 1 is a flow chart of an implementation of a method for identifying the health status of chickens provided in an embodiment of the present invention;

[0057] Figure 2 is a schematic diagram of the structure of a device for identifying the health status of chickens provided by an embodiment of the present invention;

[0058] Figure 3 is a schematic diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0059] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0060] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0061] As described in the background technology, since sick chickens have a group effect, if sick chickens are not identified and detected in time, they will quickly spread among the chickens, leading to a large-scale disease phenomenon, which will not only cause serious economic losses, but also bring public health problems such as food safety. However, at present, caged chicken farms still mainly rely on manual inspections, with a low recognition rate for sick chickens and require a lot of manpower.

[0062] Therefore, using deep learning-based computer vision technology to replace manual monitoring to achieve real-time identification of sick chickens and chicken condition monitoring in chicken farms is of great significance for alleviating labor shortages in chicken farms, strengthening epidemic control, improving the level of automation in the breeding industry, and realizing smart breeding.

[0063] The comb of a healthy chicken is ruddy in color, uniform in texture, and has round and bright eyes, while the comb of a sick chicken is abnormal in color, dry and uneven, and the eyes are half-closed or fully closed. Research on sick chicken identification based on traditional machine vision mainly includes using support vector machines to judge dead chickens through the key points of the comb and their movement. Sick chickens are identified by changing the color of the comb image, extracting the changes and features of the eye state, and counting the number of chicken feet and their behavior. However, the recognition accuracy is low and cannot meet the needs of livestock and poultry farming. Therefore, a sick chicken identification method with higher accuracy is urgently needed.

[0064] In order to solve the problems in the prior art, the embodiments of the present invention provide a method, device, equipment and storage medium for identifying the health status of a chicken. The following first introduces the method for identifying the health status of a chicken provided by the embodiments of the present invention.

[0065] The execution subject of the chicken health status identification method may be a chicken health status identification device, and the chicken health status identification device may be an electronic device having a processor and a memory, such as a mobile electronic device or a non-mobile electronic device. The embodiment of the present invention is not specifically limited.

[0066] See also Figure 1 , which shows a flowchart of the implementation of the method for identifying the health status of chickens provided by an embodiment of the present invention, and is described in detail as follows:

[0067] Step S110: Obtain a motion video containing a target chicken captured by a camera.

[0068] The target chicken may be one chicken, multiple chickens, chickens with special requirements, or all chickens in a certain area. Users can make a selection based on actual conditions and there is no limitation here.

[0069] The motion video may be a motion video of multiple time periods at multiple monitoring points, or a motion video of multiple time periods at a certain monitoring point, which is not limited here.

[0070] Step S120: input the motion video into the trained target tracking network to obtain a picture of the area where the target chicken is located and the activity characteristics of the target chicken.

[0071] Among them, the activity characteristic of the target chicken is the displacement change of the center point of the target chicken within a fixed time.

[0072] In some embodiments, the training process of the target tracking network may be:

[0073] Step S1210: Obtain a first training sample.

[0074] The first training sample includes a motion video sample containing a chicken, and each picture containing a chicken in the motion video sample corresponds to first annotation information, wherein the first annotation information indicates the area where the chicken is located in the picture.

[0075] Step S1220: training the pre-constructed target tracking network based on the first training sample and its corresponding first annotation information to obtain a trained target tracking network.

[0076] The first training sample and the first annotation information corresponding thereto are input into a pre-built target tracking network, and the target tracking network is trained to obtain a trained target tracking network.

[0077] In order to test the trained target tracking network, a motion video containing a chicken is selected as a test sample and input into the trained target tracking network to detect the detection effect of the trained target tracking network.

[0078] Specifically, a rectangular frame marking method may be used to mark the area where the chicken is located in each picture containing the chicken. Then the first training sample and the label corresponding thereto are input into the target tracking network for training.

[0079] After the target tracking network training is completed, detection can be performed. The motion video is input into the trained target tracking network to obtain the image of the target chicken area and the displacement change of the target chicken's center point in a fixed time. This displacement change is used as the activity feature of the target chicken.

[0080] Step S130: input the image of the target chicken area into the trained chicken head detection network to obtain the color features of the target chicken's comb and the depth features of the chicken head area.

[0081] In some embodiments, a picture of the target chicken area is input into a trained chicken head detection network to obtain a color score of the target chicken's comb, and the color score of the target chicken's comb is used as a color feature of the target chicken's comb.

[0082] Specifically, the color score of the target chicken's comb is:

[0083] score=color_probability×confidence;

[0084] Among them, color_probability is the possibility of the color of the target chicken's comb, and the dimension is 4 dimensions; confidence is the confidence of whether there is an object in the area where the target chicken's comb is located, and the dimension is 1 dimension.

[0085] In some embodiments, the training process of the chicken head detection network can be:

[0086] Step S1310: Obtain a second training sample, and input the second training sample into the trained target tracking network to obtain an image sample.

[0087] The second training sample may be the same as or different from the first training sample, which is not limited here. The second training sample includes a motion video sample containing a chicken, and the picture sample is a picture of the area where the chicken is located cut out from the second training sample.

[0088] Step S1320: Obtain second annotation information corresponding to the image sample by annotation. The second annotation information indicates the color of the rooster's comb in the image sample. The color of the rooster's comb may include bright red, white, dark blue, and yellow.

[0089] Step S1330: training the pre-constructed chicken head detection network based on the second training sample and its corresponding second annotation information to obtain a trained chicken head detection network.

[0090] The second training sample and the second annotation information corresponding thereto are input into a pre-built chicken head detection network for training, thereby obtaining a trained chicken head detection network.

[0091] Through the backbone feature extraction network of the trained chicken head detection network, the deep features of the image of the target chicken area can be obtained, and the deep features are reduced in dimension using the dimensionality reduction method to obtain the deep features of the target chicken's head area.

[0092] Step S140: Obtain the temperature of the chicken head in the area where the target chicken is located by using an infrared thermometer matched with the camera as the temperature feature of the chicken head.

[0093] After testing the temperature of the chicken head using an infrared thermometer matched with the camera, the collected temperature is normalized and used as the temperature feature of the chicken head.

[0094] Among them, the temperature characteristic of the target chicken’s head is t i for:

[0095]

[0096] Among them, T i is the head temperature of the ith chicken measured by infrared thermometer, T max is the maximum value of the chicken head temperature collected, T min It is the minimum value of the chicken head temperature collected.

[0097] Step S150: input the activity characteristics of the target chicken, the color characteristics of the comb, the depth characteristics of the area where the chicken head is located, and the temperature characteristics of the chicken head into the trained classification network to obtain the health status of the target chicken.

[0098] First, the activity features, comb color features, depth features of the chicken head area, and temperature features of the chicken head of the target chicken are fused into sick chicken identification features. Then, the sick chicken identification features are input into the trained classification network to identify the health status of the target chicken.

[0099] The method for identifying the health status of a chicken provided by the present invention first obtains a motion video containing a target chicken captured by a camera, and then inputs the motion video into a trained target tracking network to obtain a picture of the area where the target chicken is located and the activity characteristics of the target chicken. After that, the picture of the area where the target chicken is located is input into a trained chicken head detection network to obtain the color characteristics of the comb of the target chicken and the depth characteristics of the area where the chicken head is located. Next, the temperature of the chicken head in the area where the target chicken is located is obtained by an infrared thermometer matched with the camera as the temperature characteristics of the chicken head. Finally, the activity characteristics, comb color characteristics, depth characteristics of the area where the chicken head is located, and temperature characteristics of the chicken head of the target chicken are input into a trained classification network to obtain the health status of the target chicken. In this way, by identifying the collected motion video containing the chicken, the health status of the chicken can be quickly and accurately identified.

[0100] The following is a detailed description of the method for identifying the health status of chickens using a specific embodiment:

[0101] Before recognition, the target tracking network, chicken head detection network and classification network need to be trained. The specific training process is as follows:

[0102] A total of 10 surveillance videos of chicken movements at different monitoring points in a chicken farm were collected using a camera. Each video lasted 5 minutes. Among them, 7 of the surveillance videos were divided into the first training samples, and the remaining 3 were divided into the second training samples.

[0103] Among them, the target tracking network is trained, and the area where the chicken is located in each picture containing a chicken in the first training sample is marked with a rectangular box, and then the first training sample and the corresponding label marked with a rectangular box are input into the pre-built deepsort target tracking network for training.

[0104] Among them, to train the chicken head detection network, first use the annotation software to annotate the color of the comb in the picture containing the chicken area, wherein the color of the comb mainly includes four categories: bright red, white, dark blue, and yellow. Input the first training sample into the trained target tracking network to obtain a picture sample, which is a picture of the chicken area cut out from the second training sample, and then use the annotation software to annotate the color of the comb in the picture sample to obtain the second annotation information. Among them, the first training sample and the second annotation information are input, and the coordinates of the comb area and the comb color annotation are output, and then the pre-constructed Yolov3 chicken head detection network is trained.

[0105] The classification network includes three fully connected layers, the number of neurons in the first fully connected layer is 1000, the number of neurons in the second fully connected layer is 500, and the number of neurons in the third fully connected layer is 100. An output layer is added after the third fully connected layer, and the number of neurons is 1. Finally, a sigmoid activation function is used. The classification network is trained by inputting the fusion features of the activity features, comb color features, depth features of the chicken head area, and temperature features of the chicken head output in the above steps into the classification network.

[0106] After the three networks are trained, the health status of the chicken can be identified, and the second training sample is selected as the test set for identification.

[0107] First, the second training sample is input into the trained deepsort target tracking network to obtain the position of the center point of the target chicken and the image of the area where the target chicken is located. Then, the relative displacement of the target chicken within 60 seconds is calculated as the activity feature F of the target chicken. vel .

[0108] Then, the image of the target chicken area is input into the trained yolov3 chicken head detection network for testing, and the four-dimensional color feature F of the target chicken head is obtained. col , the color score of the target chicken's comb is:

[0109] score=color_probability×confidence;

[0110] Among them, color_probability is the possibility of the color of the target chicken's comb, and the dimension is 4 dimensions; confidence is the confidence of whether there is an object in the area where the target chicken's comb is located, and the dimension is 1 dimension.

[0111] At the same time, the picture of the target chicken area is input into the trained yolov3 chicken head detection network, and the deep features of the chicken head area are extracted through the yolov3 backbone feature extraction network darknet53. The input image size of darknet53 is 416×416×3, and the final output feature map size is 13×13×1024. The 1×1×1024-dimensional features of the grid where the chicken head is located are retained. The PCA principal component analysis method is used to reduce the 1024-dimensional deep features to 50 dimensions, and finally the deep features F of the chicken head area are obtained. deep .

[0112] Secondly, the temperature of the chicken head in the target chicken area is obtained by using an infrared thermometer matched with the camera, and the collected chicken head temperature is normalized and used as the temperature feature F of the chicken head. temp .

[0113] Temperature characteristic t of the target chicken head after normalization i for:

[0114]

[0115] Among them, T i is the head temperature of the ith chicken measured by infrared thermometer, T max is the maximum value of the chicken head temperature collected, T min It is the minimum value of the chicken head temperature collected.

[0116] Finally, the 1-dimensional activity feature F of the target chicken obtained in the above process is vel , 4-dimensional cockscomb color feature F col , 1D temperature characteristic of chicken head F temp And the depth feature F of the area where the 50-dimensional chicken head is located deep Fusion into sick chicken identification feature F final , the 56-dimensional F fonal By inputting it into the trained classification network, the health status of the target chicken can be obtained.

[0117] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0118] Based on the chicken health status identification method provided in the above embodiment, the present invention also provides a specific implementation of a chicken health status identification device applied to the chicken health status identification method. Please refer to the following embodiment.

[0119] like Figure 2 As shown, a device 200 for identifying the health status of chickens is provided, characterized in that it includes:

[0120] An acquisition module 210 is used to acquire a motion video containing a target chicken captured by a camera;

[0121] The first feature acquisition module 220 is used to input the motion video into the trained target tracking network to obtain a picture of the target chicken area and the activity features of the target chicken;

[0122] The second and third feature acquisition module 230 is used to input the image of the target chicken area into the trained chicken head detection network to obtain the color feature of the target chicken's comb and the depth feature of the chicken head area;

[0123] The fourth feature acquisition module 240 is used to acquire the temperature of the chicken head in the area where the target chicken is located by using an infrared thermometer matched with the camera as the temperature feature of the chicken head;

[0124] The health status determination module 250 is used to input the activity characteristics, comb color characteristics, depth characteristics of the chicken head area and temperature characteristics of the chicken head of the target chicken into the trained classification network to obtain the health status of the target chicken.

[0125] In a possible implementation, obtaining the second and third feature modules 230 is specifically used for:

[0126] Input the image of the target chicken area into the trained chicken head detection network to obtain the color score of the target chicken's comb, and use the color score of the target chicken's comb as the target chicken's comb color feature;

[0127] Through the backbone feature extraction network of the chicken head detection network, the deep features of the image of the target chicken area are obtained, and the deep features are reduced in dimension to obtain the deep features of the target chicken's head area.

[0128] In a possible implementation, the color score of the target chicken's comb is:

[0129] score=color_probability×confidence;

[0130] Among them, color_probability is the possibility of the color of the target chicken's comb, and the dimension is 4 dimensions; confidence is the confidence of whether there is an object in the area where the target chicken's comb is located, and the dimension is 1 dimension.

[0131] In a possible implementation, the first feature acquisition module 220 is further used to:

[0132] Acquire a first training sample, wherein the first training sample includes a motion video sample containing a chicken, each picture containing a chicken in the motion video sample corresponds to first annotation information, wherein the first annotation information indicates a region where the chicken is located in the picture;

[0133] The pre-constructed target tracking network is trained based on the first training sample and the first annotation information corresponding thereto to obtain a trained target tracking network.

[0134] In a possible implementation, the second and third feature acquisition module 230 is further used to:

[0135] Obtain a second training sample, and input the second training sample into the trained target tracking network to obtain a picture sample, wherein the second training sample includes a motion video sample containing a chicken, and the picture sample is a picture of the area where the chicken is located cropped from the second training sample;

[0136] Obtaining second annotation information corresponding to the image sample through annotation, wherein the second annotation information indicates the color of the comb in the image sample;

[0137] The pre-constructed chicken head detection network is trained based on the second training sample and its corresponding second annotation information to obtain a trained chicken head detection network.

[0138] In a possible implementation, the activity characteristic of the target chicken is the displacement change of the center point of the target chicken within a fixed time.

[0139] In a possible implementation, the temperature characteristic t of the target chicken's head is i for:

[0140]

[0141] Among them, T i is the head temperature of the ith chicken measured by infrared thermometer, T max is the maximum value of the chicken head temperature collected, T min It is the minimum value of the chicken head temperature collected.

[0142] Figure 3 Schematic diagram of an electronic device provided by an embodiment of the present invention. Figure 3 As shown, the electronic device 3 of this embodiment includes: a processor 30, a memory 31, and a computer program 32 stored in the memory 31 and executable on the processor 30. When the processor 30 executes the computer program 32, the steps of the above-mentioned method for identifying the health status of each chicken are implemented, for example Figure 1 Alternatively, when the processor 30 executes the computer program 32, the functions of each module in the above-mentioned device embodiments are realized, for example, Figure 2 The functions of the modules 210 to 250 are shown.

[0143] Exemplarily, the computer program 32 may be divided into one or more modules, which are stored in the memory 31 and executed by the processor 30 to implement the present invention. The one or more modules may be a series of computer program instruction segments capable of implementing specific functions, which are used to describe the execution process of the computer program 32 in the electronic device 3. For example, the computer program 32 may be divided into Figure 2 Modules 210 to 250 are shown.

[0144] The electronic device 3 may include, but is not limited to, a processor 30 and a memory 31. Those skilled in the art will appreciate that Figure 3It is only an example of the electronic device 3 and does not constitute a limitation of the electronic device 3. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, the electronic device may also include input and output devices, network access devices, buses, etc.

[0145] The processor 30 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0146] The memory 31 may be an internal storage unit of the electronic device 3, such as a hard disk or memory of the electronic device 3. The memory 31 may also be an external storage device of the electronic device 3, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the electronic device 3. Further, the memory 31 may also include both an internal storage unit and an external storage device of the electronic device 3. The memory 31 is used to store the computer program and other programs and data required by the electronic device. The memory 31 may also be used to temporarily store data that has been output or is to be output.

[0147] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0148] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0149] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0150] In the embodiments provided by the present invention, it should be understood that the disclosed devices / electronic devices and methods can be implemented in other ways. For example, the device / electronic device embodiments described above are only schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0151] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0152] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0153] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned identification method of each chicken health status can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practices in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practices, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0154] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for identifying the health status of a chicken, characterized in that: include: Obtaining a motion video containing a target chicken captured by a camera; Input the motion video into a trained target tracking network to obtain a picture of the area where the target chicken is located and activity characteristics of the target chicken; Input the image of the target chicken area into the trained chicken head detection network to obtain the color score of the target chicken's comb, and use the color score of the target chicken's comb as the target chicken's comb color feature; The deep features of the image of the target chicken area are obtained through the backbone feature extraction network of the chicken head detection network, and the deep features are subjected to dimensionality reduction processing to obtain the deep features of the target chicken area; The temperature of the chicken head in the area where the target chicken is located is obtained by using an infrared thermometer matched with the camera as a temperature feature of the chicken head; Inputting the activity characteristics, comb color characteristics, depth characteristics of the chicken head area and temperature characteristics of the chicken head of the target chicken into the trained classification network to obtain the health status of the target chicken; The color score of the comb of the target chicken is for: ； is the probability of the comb color of the target chicken, with a dimension of 4; The confidence level of whether there is an object in the area where the comb of the target chicken is located, with a dimension of 1; Temperature characteristics of the target chicken's head for: ； The infrared thermometer is used to test the i The temperature of the chicken head, is the maximum value of the chicken head temperature collected, It is the minimum value of the chicken head temperature collected.

2. The identification method according to claim 1, characterized in that: The method further comprises: Acquire a first training sample, wherein the first training sample includes a motion video sample containing a chicken, each picture containing a chicken in the motion video sample corresponds to first annotation information, wherein the first annotation information indicates an area where the chicken is located in the picture; The pre-constructed target tracking network is trained based on the first training sample and the first annotation information corresponding thereto to obtain a trained target tracking network.

3. The identification method according to claim 2, characterized in that: The method further comprises: Obtain a second training sample, and input the second training sample into the trained target tracking network to obtain a picture sample, wherein the second training sample includes a motion video sample containing a chicken, and the picture sample is a picture of the chicken area cropped from the second training sample; Obtaining second annotation information corresponding to the image sample through annotation, wherein the second annotation information indicates the color of the rooster comb in the image sample; The pre-constructed chicken head detection network is trained based on the second training sample and its corresponding second annotation information to obtain a trained chicken head detection network.

4. The identification method according to claim 1, characterized in that: The activity characteristic of the target chicken is the displacement change of the center point of the target chicken within a fixed time.

5. A device for identifying the health status of chickens, characterized in that: include: An acquisition module is used to acquire a motion video containing a target chicken captured by a camera; Obtaining a first feature module, which is used to input the motion video into a trained target tracking network to obtain a picture of the area where the target chicken is located and an activity feature of the target chicken; Obtain a second and a third feature module, which is used to input the image of the target chicken area into the trained chicken head detection network to obtain the color score of the target chicken's comb, and use the color score of the target chicken's comb as the target chicken's comb color feature; The deep features of the image of the target chicken area are obtained through the backbone feature extraction network of the chicken head detection network, and the deep features are subjected to dimensionality reduction processing to obtain the deep features of the target chicken area; A fourth feature acquisition module is used to acquire the temperature of the chicken head in the area where the target chicken is located by using an infrared thermometer matched with the camera as a temperature feature of the chicken head; A health status determination module is used to input the activity characteristics, comb color characteristics, depth characteristics of the chicken head area and temperature characteristics of the chicken head of the target chicken into a trained classification network to obtain the health status of the target chicken; The color score of the comb of the target chicken is for: ； is the probability of the comb color of the target chicken, with a dimension of 4; The confidence level of whether there is an object in the area where the comb of the target chicken is located, with a dimension of 1; Temperature characteristics of the target chicken's head for: ； The infrared thermometer is used to test the i The temperature of the chicken head, is the maximum value of the chicken head temperature collected, It is the minimum value of the chicken head temperature collected.

6. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 4.

7. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images