Sow oestrus judgment method and system based on multi-source data

The sow estrus detection system, which uses multi-source data and combines robots and bionic devices with video analysis, automatically determines the estrus status of sows, solving the problems of high cost and error caused by reliance on manual methods, and achieving efficient and accurate estrus detection.

CN118985480AInactive Publication Date: 2024-11-22INSTITUTE OF ANIMAL SCIENCES OF CHINESE ACADEMY OF AGRICULTURAL SCIENCES

Patent Information

Application Number
CN202411114222.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-11-22
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing methods for determining estrus in sows rely on manual operation, resulting in high labor costs and a shortage of professional breeders, making it difficult to efficiently and accurately identify the estrus cycle of sows.

Method used

A sow estrus detection system based on multi-source data is adopted. The system uses a chassis robot to carry a bionic device that plays the mating calls of boars and releases pheromones. Combined with video acquisition and target detection models, the system analyzes the number of times the sow comes into contact with the bionic device to automatically determine whether the sow is in estrus.

Benefits of technology

It reduces reliance on professional keepers, improves the efficiency and accuracy of estrus detection, reduces human error, and achieves automated and consistent judgment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118985480A_ABST
    Figure CN118985480A_ABST
Patent Text Reader

Abstract

The invention discloses a sow oestrus judgment method and system based on multi-source data, and the method comprises the steps: receiving an action instruction based on a chassis robot, carrying out the path planning, and advancing to a target detection region; a mechanical arm carried on the chassis robot is controlled based on a PID controller, the bionic equipment is moved to a designated position, the boar doll sound is played, and boar pheromones are released; acquiring an interactive video of the sow and the bionic equipment through video acquisition equipment; a target detection model is constructed and trained, the interactive video is analyzed based on the trained target detection model, and the number of times of contact between the sow and the bionic equipment is obtained; and comparing the contact times of the sow and the bionic equipment in unit time with a preset oestrus threshold value so as to judge whether the sow is oestrus or not. According to the sow oestrus detection method provided by the invention, a relatively good oestrus test effect is achieved by releasing boar pheromones and playing boar courtship sound, the needs of manual observation and judgment are reduced, and the oestrus identification efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of breeding pigs, and in particular relates to a sow estrus judgment method and system based on multi-source data. Background Art

[0002] In the process of improving the success rate of sow pregnancy, it is crucial to determine the best time for conception. This requires breeders to closely monitor the estrus cycle of sows so that they can prepare for reproduction in time. Identifying whether sows are in estrus is a key factor in ensuring that they can mate, conceive and give birth normally. Currently, commonly used estrus identification methods include observing the behavior of sows, checking the mucus state of the vaginal mucosa, performing standing reflex tests, and using boars for estrus testing.

[0003] These methods mainly rely on manual operation, usually by experienced breeders to evaluate the estrus behavior of sows, which is a large investment in labor costs. However, with the centralized and large-scale development of the breeding industry, the shortage of professional breeders has become a major problem facing modern pig production. Therefore, it is urgent to propose a sow estrus judgment method and system based on multi-source data to improve efficiency and reduce dependence on professional breeders. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes a sow estrus judgment method and system based on multi-source data to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides a sow estrus judgment system based on multi-source data, comprising a path planning module, a bionic interaction module, a video acquisition module, a data analysis module and an estrus judgment module connected in sequence;

[0006] The path planning module is used to receive action instructions based on the chassis robot and perform path planning based on the action instructions;

[0007] The bionic interaction module is used to control the mechanical arm mounted on the chassis robot, move the bionic device connected to the mechanical arm to a designated position, play the boar's courtship call and release the boar's pheromone;

[0008] The video acquisition module is used to acquire the interactive video between the sow and the bionic device through a video acquisition device;

[0009] The data analysis module is used to build and train a target detection model based on the yolov5 network, analyze the interactive video based on the trained target detection model, and obtain the number of contacts between the sow and the bionic device;

[0010] The estrus judgment module is used to compare the number of contacts between the sow and the bionic device per unit time with a preset estrus threshold value, so as to judge whether the sow is in estrus.

[0011] Optionally, the path planning module includes an instruction receiving unit, a radar scanning unit and a path planning unit;

[0012] The instruction receiving unit is used to receive action instructions from the industrial computer and determine the target detection area; the radar scanning unit is used to collect surrounding environment data based on the radar to build an environment map; the path planning unit is used to plan the travel path of the chassis robot based on the environment map and the target detection area using the A* path planning method.

[0013] Optionally, the bionic interaction module includes a rotation control unit and a bionic interaction unit;

[0014] The rotation control unit is used to control the rotation of the robot arm through a PID controller to move the bionic device to a specified frame;

[0015] The bionic interaction unit is used to control the bionic device in the designated frame to play the boar's courtship call and release the boar's pheromone.

[0016] Optionally, the data analysis module includes a model building unit and a target detection unit;

[0017] The model building unit is used to build a target detection model based on the yolov5 network, and train the target detection model based on the pig nose photo in the interactive video;

[0018] The target detection unit is used to extract single-frame image features in the interactive video based on the trained target detection model, perform target detection on the sow's real pig nose based on the single-frame image features, and obtain an image position frame of the real pig nose; preset the image position frame of the target pig nose, and determine whether the single-frame image is in contact according to the distance between the image position frame of the real pig nose and the image position frame of the target pig nose, and then obtain the number of contacts between the sow and the bionic device based on multiple frames of images.

[0019] The present invention also provides a sow estrus judgment method based on multi-source data, and a sow estrus judgment system based on multi-source data, comprising the following steps:

[0020] The chassis robot receives action instructions, and plans a path based on the action instructions to move to the target detection area;

[0021] The PID controller controls the robotic arm mounted on the chassis robot, moves the bionic device connected to the robotic arm to the designated position, plays the boar's courtship call and releases the boar's pheromone;

[0022] The interactive video between the sow and the bionic device is collected by video acquisition equipment;

[0023] Build and train a target detection model based on the yolov5 network, analyze the interactive video based on the trained target detection model, and obtain the number of contacts between the sow and the bionic device;

[0024] The number of times the sow contacts the bionic device per unit time is compared with the preset estrus threshold to determine whether the sow is in estrus.

[0025] Optionally, the process of path planning for the chassis robot includes: receiving action instructions from an industrial computer and determining a target detection area; collecting surrounding environment data based on radar to build an environment map; and planning the travel path of the chassis robot using an A* path planning method based on the environment map and the target detection area.

[0026] Optionally, the process of obtaining the number of contacts between the sow and the bionic device includes: extracting single-frame image features in the interactive video based on a trained target detection model, performing target detection on the sow's real nose based on the single-frame image features, and obtaining an image position frame of the real nose; presetting an image position frame of the target nose, and determining whether the single-frame image is in contact based on the distance between the image position frame of the real nose and the image position frame of the target nose, and then obtaining the number of contacts between the sow and the bionic device based on multiple frames of images.

[0027] Optionally, the loss function of the target detection model adopts a DIOU function, and the formula of the DIOU function is as follows:

[0028]

[0029] Among them, IoU(A, B) represents the IoU value between image location boxes A and B. represents the center point distance between the image location boxes A and B, and ω represents the length of the image diagonal.

[0030] The present invention also provides an electronic device, comprising: a memory and a processor; the memory is used to store programs; the processor is used to execute the programs to implement various modules of the sow estrus judgment system based on multi-source data.

[0031] The present invention also provides a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, each module of the sow estrus judgment system based on multi-source data is implemented.

[0032] Compared with the prior art, the present invention has the following advantages and technical effects:

[0033] The sow estrus detection method provided by the present invention aims at the characteristics of pigs' poor innate visual ability but sensitive touch, hearing and smell. By releasing boar pheromones and playing boar mating calls, combined with a bionic boar mouth and nose, a better estrus detection effect is achieved. The automated system reduces the need for manual observation and judgment, thereby improving the efficiency of estrus identification.

[0034] The detection system of the present invention includes a chassis robot, an automated robotic arm and a bionic device, which reduces the dependence on professional breeders and the errors caused by human factors, and not only solves the problem of shortage of professional talents, but also improves the consistency and reliability of estrus identification.

[0035] The target detection model constructed by the present invention using the yolov5 network can accurately analyze the video of the interaction between the sow and the bionic device, thereby improving the accuracy of estrus judgment.

[0036] The present invention can make a more objective and accurate estrus judgment based on the data by collecting and analyzing the number of contacts between the sow and the bionic device.

[0037] The system proposed in the present invention can be easily integrated into existing breeding facilities and can be expanded to more detection areas or more sows as needed. It can collect and analyze data in real time, provide timely feedback on the estrus status of sows, and provide instant information for breeding management. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0039] Figure 1 This is a schematic diagram of the structure of a sow estrus judgment system based on multi-source data according to an embodiment of the present invention;

[0040] Figure 2 The present invention is a flowchart of a method for determining sow estrus based on multi-source data according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0042] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0043] Embodiment 1

[0044] like Figure 1 As shown, in this embodiment, a sow estrus judgment system based on multi-source data is provided, including a path planning module, a bionic interaction module, a video acquisition module, a data analysis module and an estrus judgment module connected in sequence;

[0045] The path planning module is used to receive action instructions based on the chassis robot and perform path planning based on the action instructions;

[0046] The bionic interaction module is used to control the mechanical arm mounted on the chassis robot, move the bionic device connected to the mechanical arm to a designated position, play the boar's courtship call and release the boar's pheromone;

[0047] The video acquisition module is used to acquire the interactive video between the sow and the bionic device through a video acquisition device;

[0048] The data analysis module is used to build and train a target detection model based on the yolov5 network, analyze the interactive video based on the trained target detection model, and obtain the number of contacts between the sow and the bionic device;

[0049] The estrus judgment module is used to compare the number of contacts between the sow and the bionic device per unit time with a preset estrus threshold value, so as to judge whether the sow is in estrus.

[0050] It can be implemented that the path planning module includes an instruction receiving unit, a radar scanning unit and a path planning unit; the instruction receiving unit is used to receive action instructions from an industrial computer and determine a target detection area; the radar scanning unit is used to build an environment map based on surrounding environment data collected by radar; the path planning unit is used to plan the travel path of the chassis robot based on the environment map and the target detection area using an A* path planning method.

[0051] It can be implemented that the bionic interaction module includes a rotation control unit and a bionic interaction unit; the rotation control unit is used to control the rotation of the robotic arm through a PID controller to move the bionic device to a designated frame; the bionic interaction unit is used to control the bionic device in the designated frame to play boar courtship calls and release boar pheromones.

[0052] In this embodiment, it should be noted that releasing boar pheromones to sows may include: adjusting the concentration and release duration of the boar pheromones by controlling the odor release device. A central processing device may be used to send a signal to the odor release device to control the odor release device. The odor release device may include an ultrasonic atomization device to better release the gas containing the boar pheromone, and the boar pheromone in the gas may include boar urine, semen, saliva, etc.

[0053] In this embodiment, it should be noted that playing the boar's mating sound to the sow may include: playing the boar's mating sound of various decibels and various ages through a sound playing device. The central processing device may send a signal to the sound playing device to control the sound playing device. Since the sow has a keen hearing, the boar's mating sound may be recorded, and the sow's affection may be tested by using the sound of the boar's mating sound.

[0054] Playing the boar's mating call and releasing the boar's pheromone can be performed simultaneously or in stages. Depending on the breed, one step can be omitted. For example, some breeds of sows will respond only to the boar's mating call, so the step of releasing the boar's pheromone can be omitted.

[0055] The bionic device of this embodiment can use a bionic boar snout. The material of the bionic boar snout can be made of flexible materials such as silicone, so as to be close to the touch of a boar. The height of the bionic boar snout is made consistent with the sow's head through a rotatable mechanical arm. Compared with the entire fake boar model, the volume occupied by the bionic boar snout is much smaller; compared with the need to control the activities of the boar model, it is easier to control the activities of the bionic boar snout; the sow's innate visual ability is poor, but the sense of touch is more developed and more sensitive to touch. In the natural environment, the boar will touch the sow with his snout. By using the bionic boar snout to touch the sow, a better effect of checking and testing the condition is achieved.

[0056] It is feasible that the data analysis module includes a model building unit and a target detection unit; the model building unit is used to build a target detection model based on the yolov5 network, and train the target detection model based on the pig nose photo in the interactive video; the target detection unit is used to extract single-frame image features in the interactive video based on the trained target detection model, perform target detection on the sow's real pig nose based on the single-frame image features, and obtain an image position frame of the real pig nose; preset the image position frame of the target pig nose, determine whether the single-frame image is in contact according to the distance between the image position frame of the real pig nose and the image position frame of the target pig nose, and then obtain the number of contacts between the sow and the bionic device based on multiple frames of images.

[0057] Embodiment 2

[0058] like Figure 2 As shown, the present invention also provides a sow estrus judgment method based on multi-source data, and a sow estrus judgment system based on multi-source data, comprising the following steps:

[0059] The chassis robot receives action instructions, and plans a path based on the action instructions to move to the target detection area;

[0060] The robot arm mounted on the chassis robot is controlled by a PID controller to move the bionic device connected to the robot arm to the designated position, play the boar's courtship call and release the boar's pheromone;

[0061] The interactive video between the sow and the bionic device is collected by video acquisition equipment;

[0062] Build and train a target detection model based on the yolov5 network, analyze the interactive video based on the trained target detection model, and obtain the number of contacts between the sow and the bionic device;

[0063] The number of times the sow contacts the bionic device per unit time is compared with the preset estrus threshold to determine whether the sow is in estrus.

[0064] The feasible process of path planning for the chassis robot includes: receiving action instructions from the industrial computer and determining the target detection area; collecting surrounding environment data based on radar to build an environment map; and planning the travel path of the chassis robot using the A* path planning method based on the environment map and the target detection area.

[0065] The process of obtaining the number of contacts between the sow and the bionic device can be implemented as follows: extracting single-frame image features in the interactive video based on a trained target detection model, performing target detection on the sow's real nose based on the single-frame image features, and obtaining an image position frame of the real nose; presetting an image position frame of the target nose, determining whether the single-frame image is in contact based on the distance between the image position frame of the real nose and the image position frame of the target nose, and then obtaining the number of contacts between the sow and the bionic device based on multiple frames of images.

[0066] The yolov5 network can be implemented to include a backbone network (Backbone): New CSP-Darknet53, a neck network (Neck): SPPF, New CSP-PAN, and a head network (Head): YOLOv3 Head. This embodiment embeds adaptive anchor box calculation in the yolov5 network, and before each network training, it can adaptively calculate the optimal anchor box value in different data sets according to different data sets.

[0067] It is feasible that the loss function of the target detection model adopts an IoU loss function, which is mainly used to measure the difference between the predicted bounding box and the real bounding box.

[0068] IoU loss is a variation of IoU, which is a metric used to measure the degree of overlap between a predicted bounding box and a true bounding box. In object detection, IoU is often used to evaluate the overlap between a predicted box and a true box to determine whether the predicted box is correct. Specifically, for each predicted bounding box, its IoU value with all true bounding boxes needs to be calculated, and then the true bounding box with the largest IoU is selected as its corresponding matching target, thereby calculating its IoU loss.

[0069] It specifically uses the DIOU loss function. DIOU (Distance IoU) introduces the center point distance based on IoU, which can effectively alleviate the impact of positioning error on target detection performance. Specifically, the formula of the DIOU function is as follows:

[0070]

[0071] Among them, IoU(A, B) represents the IoU value between image location boxes A and B. represents the center point distance between the image location boxes A and B, and ω represents the length of the image diagonal.

[0072] Embodiment 3

[0073] The present invention also provides an electronic device, comprising: a memory and a processor; the memory is used to store programs; the processor is used to execute the programs, and when the program instructions are executed by the computer, the computer can execute the various modules of the sow estrus judgment system based on multi-source data.

[0074] Embodiment 4

[0075] The present invention also provides a readable storage medium on which a computer program is stored. When the computer program is executed by a processor, each module of the sow estrus judgment system based on multi-source data is implemented.

[0076] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A sow estrus judgment system based on multi-source data, characterized in that: It includes a path planning module, a bionic interaction module, a video acquisition module, a data analysis module and an estrus judgment module which are connected in sequence; The path planning module is used to receive action instructions based on the chassis robot and perform path planning based on the action instructions; The bionic interaction module is used to control the mechanical arm mounted on the chassis robot, move the bionic device connected to the mechanical arm to a designated position, play the boar's courtship call and release the boar's pheromone; The video acquisition module is used to acquire the interactive video between the sow and the bionic device through a video acquisition device; The data analysis module is used to build and train a target detection model based on the yolov5 network, analyze the interactive video based on the trained target detection model, and obtain the number of contacts between the sow and the bionic device; The estrus judgment module is used to compare the number of contacts between the sow and the bionic device per unit time with a preset estrus threshold value, so as to judge whether the sow is in estrus.

2. The sow estrus judgment system based on multi-source data according to claim 1, characterized in that: The path planning module includes an instruction receiving unit, a radar scanning unit and a path planning unit; The instruction receiving unit is used to receive the action instruction from the industrial computer and determine the target detection area; the radar scanning unit is used to collect the surrounding environment data based on the radar to build an environment map; the path planning unit is used to use A based on the environment map and the target detection area. * The path planning method plans the travel path of the chassis robot.

3. The sow estrus judgment system based on multi-source data according to claim 1, characterized in that: The bionic interaction module includes a rotation control unit and a bionic interaction unit; The rotation control unit is used to control the rotation of the robot arm through a PID controller to move the bionic device to a specified frame; The bionic interaction unit is used to control the bionic device in the designated frame to play the boar's courtship call and release the boar's pheromone.

4. The sow estrus judgment system based on multi-source data according to claim 1, characterized in that: The data analysis module includes a model building unit and a target detection unit; The model building unit is used to build a target detection model based on the yolov5 network, and train the target detection model based on the pig nose photo in the interactive video; The target detection unit is used to extract single-frame image features in the interactive video based on the trained target detection model, perform target detection on the real pig nose of the sow based on the single-frame image features, and obtain an image position frame of the real pig nose; An image position frame of the target pig nose is preset, and whether a single frame image is in contact is determined according to the distance between the image position frame of the real pig nose and the image position frame of the target pig nose, and then the number of contacts between the sow and the bionic device is obtained based on multiple frames of images.

5. A method for determining sow estrus based on multi-source data, characterized in that: The sow estrus judgment system based on multi-source data according to any one of claims 1 to 4 comprises the following steps: The chassis robot receives action instructions, and plans a path based on the action instructions to move to the target detection area; The PID controller controls the robotic arm mounted on the chassis robot, moves the bionic device connected to the robotic arm to the designated position, plays the boar's courtship call and releases the boar's pheromone; The interactive video between the sow and the bionic device is collected by video acquisition equipment; Build and train a target detection model based on the yolov5 network, analyze the interactive video based on the trained target detection model, and obtain the number of contacts between the sow and the bionic device; The number of times the sow contacts the bionic device per unit time is compared with the preset estrus threshold to determine whether the sow is in estrus.

6. The sow estrus judgment method based on multi-source data according to claim 5, characterized in that: The process of path planning for the chassis robot includes: receiving action instructions from the industrial computer and determining the target detection area; collecting surrounding environment data based on radar to build an environment map; and planning the chassis robot's travel path based on the environment map and the target detection area using the A* path planning method.

7. The sow estrus judgment method based on multi-source data according to claim 5, characterized in that: The process of obtaining the number of contacts between the sow and the bionic device includes: extracting single-frame image features in the interactive video based on the trained target detection model, performing target detection on the sow's real pig nose based on the single-frame image features, and obtaining an image position frame of the real pig nose; presetting an image position frame of the target pig nose, determining whether the single-frame image is in contact according to the distance between the image position frame of the real pig nose and the image position frame of the target pig nose, and then obtaining the number of contacts between the sow and the bionic device based on multiple frames of images.

8. The sow estrus judgment method based on multi-source data according to claim 7, characterized in that: The loss function of the target detection model adopts the DIOU function, and the formula of the DIOU function is as follows: Among them, IoU(A, B) represents the IoU value between image location boxes A and B. represents the center point distance between the image location boxes A and B, and ω represents the length of the image diagonal.

9. An electronic device, characterized in that: include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each module of the sow estrus judgment system based on multi-source data as described in any one of claims 1-4.

10. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each module of the sow estrus judgment system based on multi-source data as described in any one of claims 1 to 4 is implemented.

Citation Information

Patent Citations

  • Sow estrus detection method and device

    CN112640809A

  • Sow estrus state automatic detection system based on interaction characteristics

    CN115226650A

  • Sow oestrus detection method based on adaptive navigation and related equipment

    CN115943908A

Cited By

  • Local boar remote semen collection training method based on simulation human interaction

    CN120525659A