Automatic identification and recording system for abnormal respiratory behaviors of rabbits

Abstract

The invention relates to the technical field of rabbit behavior recognition, in particular to an automatic recognition and recording system for abnormal respiratory behaviors of rabbits. According to the technical scheme, the system comprises an image acquisition module, a data transmission module, a data analysis module and a data storage and interaction module. The image acquisition module is used for acquiring image data and environment data of a rabbit, the data transmission module is responsible for transmitting the data acquired by the image acquisition module to the data analysis module, and the data analysis module processes and analyzes the received data to identify abnormal breathing behaviors of the rabbit. And the data storage and interaction module is used for storing data and realizing interaction operation with a user. According to the method, image acquisition, data transmission, analysis, storage and interaction are comprehensively optimized, the accuracy, reliability and convenience of recognition of abnormal respiratory behaviors of rabbits are remarkably improved, fine management and health guarantee of rabbit breeding are effectively assisted, and considerable economic benefits and efficient management experience are brought to farmers.

Description

Technical Field

[0001] The invention relates to the technical field of rabbit behavior recognition, and in particular to an automatic recognition and recording system for abnormal rabbit breathing behavior. Background Art

[0002] In the process of rabbit breeding, the health status of rabbits plays a vital role in the breeding benefits. Abnormal breathing is often one of the important signs of rabbit health problems. Timely and accurate identification of abnormal breathing behavior of rabbits can provide key basis for early disease diagnosis and treatment, thereby effectively reducing the mortality rate of rabbits and reducing the economic losses of breeders.

[0003] Traditional rabbit health monitoring mainly relies on manual observation by breeders. However, this method has many limitations. First, manual observation is difficult to achieve continuous and uninterrupted, and breeders cannot focus on monitoring the respiratory status of a single rabbit for a long time, which easily leads to missed detection of abnormal respiratory behavior. Secondly, there are errors in people's subjective judgment. Different breeders may have different judgment criteria for abnormal breathing, and they are prone to fatigue and negligence after long-term observation, making the judgment results inaccurate and unreliable. Furthermore, manual observation can only be carried out when the breeder is present, and remote monitoring of the rabbit's respiratory status cannot be achieved, which is not conducive to refined management in large-scale breeding scenarios.

[0004] With the continuous development of science and technology, some automated monitoring technologies have begun to be applied to the field of animal health monitoring. For example, some existing technologies use a single ordinary optical camera to collect animal images and infer breathing conditions by analyzing the movement of the animal's body in the image. However, this method is insufficient in obtaining the depth information of the rabbit's body, and it is difficult to accurately locate the key areas related to breathing, such as the exact position and movement amplitude of the chest cavity, which affects the accuracy of respiratory feature extraction. In addition, in terms of data transmission, some traditional monitoring systems have low transmission rates or poor stability, and are prone to data loss or transmission delays, resulting in a lag in the identification of abnormal respiratory behavior. In terms of data analysis algorithms, the algorithms used in some existing technologies are relatively simple, and the comprehensive judgment of respiratory frequency, respiratory depth and respiratory rhythm is not accurate enough. There is a lack of effective compensation and adaptation mechanisms for complex environmental factors (such as light changes, temperature and humidity fluctuations, etc.), which cannot meet the needs of high-precision rabbit abnormal respiratory behavior identification.

[0005] In summary, there is an urgent need for an automatic recognition and recording system for abnormal rabbit breathing behavior that can overcome the defects of existing technologies and realize high-precision, automation, remoteness, stability and reliability, so as to improve the health management level and breeding efficiency of rabbit breeding. Summary of the invention

[0006] The purpose of the present invention is to address the problem that there is a lack of high-precision, stable and reliable automatic recognition and recording of abnormal breathing behavior in rabbits in the background technology, and to propose an automatic recognition and recording system for abnormal breathing behavior in rabbits.

[0007] The technical solution of the present invention is a system for automatically identifying and recording abnormal breathing behavior of rabbits, comprising: An image acquisition module is used to obtain the image data and environmental data of the rabbit; The data transmission module is responsible for transmitting the data collected by the image acquisition module to the data analysis module; The data analysis module processes and analyzes the received data to identify abnormal breathing behavior of rabbits; The data storage and interaction module is used to store data and implement interactive operations with users.

[0008] Optionally, the image acquisition module includes: The image acquisition unit uses a combination of a high-resolution, high-frame-rate depth camera and an ordinary optical camera. The depth camera has a resolution of 1280×720 pixels and a frame rate of 60fps. The ordinary optical camera has a resolution of 1920×1080 pixels and a frame rate of 30fps. The camera is equipped with automatic exposure, automatic white balance functions and an 8-24mm variable-focus lens, and is installed on a movable robotic arm. The range of movement of the robotic arm in the X-axis direction is 0-2 meters, the Y-axis direction is 0-1.5 meters, and the Z-axis direction is 0.5-1.2 meters. The environmental monitoring unit includes a temperature and humidity sensor with an accuracy of ±0.3°C, a measurement range of −10°C to 40°C, and a humidity measurement range of 20% to 90%, and a light sensor with a measurement range of 0 to 20,000 lux.

[0009] Optionally, the data transmission module includes: Wired transmission unit uses high-speed Ethernet cable for data transmission, follows TCP / IP protocol, and has a transmission rate of 1Gbps or above; The wireless transmission unit is equipped with a Wi-Fi 6 wireless transmission module, complies with the IEEE 802.11ax standard, has a theoretical maximum transmission rate of 9.6Gbps, and a dual-band transmission band of 2.4GHz and 5GHz. The wired transmission unit and the wireless transmission unit can switch automatically, and the switching time is 8-10 seconds.

[0010] Optionally, the data analysis module includes an image preprocessing unit, and the image preprocessing unit specifically includes: The image registration subunit registers the images captured by the depth camera and the optical camera using a registration algorithm based on feature points; The noise removal subunit uses a bilateral filtering algorithm to remove noise from the registered image.

[0011] Optionally, in the image registration subunit, feature points in the two images are extracted by a scale-invariant feature transformation algorithm, and feature points in the depth image are and feature points in the optical image , calculate the matching relationship between them, and set the error threshold of feature point matching Pixels, the matching error is optimized by the least squares method, so that in, is the depth image The three-dimensional coordinates of the feature points are expressed as ,in is the depth value, The optical image The two-dimensional coordinates of the feature points, is the error threshold of feature point matching, in pixels. is the number of matching feature point pairs, thus obtaining accurate image registration results.

[0012] Optionally, in the noise removal subunit, the bilateral filtering formula is: in, It's pixels The intensity value after filtering, is a pixel The original intensity value, It's pixels Neighborhood of is a spatial domain Gaussian function, which is used to calculate the spatial distance between pixels. Calculate weights, standard deviation , is a range Gaussian function that is used to calculate the pixel intensity difference Calculate weights, standard deviation , is the normalization coefficient, .

[0013] Optionally, the data analysis module further includes a rabbit breathing feature extraction unit, and the rabbit breathing feature extraction unit includes: Breathing area positioning subunit, seed point , set the growth threshold Then, starting from the seed point, the adjacent pixels whose depth values ​​differ from the seed point depth values ​​within the growth threshold range are merged into the growth area, and the iteration is continued until no new pixel points meet the conditions. The area threshold of the growth area is set to Square centimeters, by comparing with the preset rabbit body model, the accurate breathing area is further determined; The respiratory frequency calculation subunit analyzes the image sequence of the respiratory area and uses the optical flow method to calculate the movement speed of the respiratory area. Frame and Pixels within the frame breathing area The optical flow vector is , solve the optical flow equation using the Horn-Schunck optical flow algorithm: in, is the image intensity function, which indicates the brightness value of a point in the image over time ( ) and space changes, , are optical flow vector components, representing the pixel points at and Movement speed in the direction, setting the optical flow speed threshold m / s. When the optical flow velocity of more than 30% of the pixels in the breathing area exceeds the optical flow velocity threshold, it is recorded as a breathing action. The breathing frequency ,in For in time The number of breathing movements detected within is the total monitoring time. The normal rabbit respiratory rate range is 30 to 60 times per minute. If it exceeds this range, it is considered abnormal. The breathing depth analysis subunit calculates the maximum change in the depth of the chest area during breathing based on the depth image sequence of the breathing area and sets the depth change threshold If the chest depth change during breathing exceeds the depth change threshold, the breathing depth is judged to be abnormal; Optionally, the data analysis module further includes an abnormal behavior recognition unit, which is provided with a respiratory rhythm judgment subunit to analyze the time interval sequence between consecutive respiratory actions. , calculate the autocorrelation function of the time interval: in, is the autocorrelation function at the lag order The value at is used to measure the periodicity of the breathing time interval sequence. Time interval series The average value of is the total number of time interval sequences, is the lag order. Set the autocorrelation function threshold , when the autocorrelation function is in the lag order range When both values ​​are less than the threshold, the respiratory rhythm is judged to be abnormal.

[0014] Optionally, the data storage and interaction module includes: The data storage unit uses a distributed file system to store raw image data and stores them by date, rabbit number and camera type. It uses a relational database to store analysis data, including individual rabbit information, respiratory characteristic data and abnormal behavior records. The data storage period is no less than 1 year. The user interaction unit develops a Web-based user interface. Users can log in through a browser to view the rabbit's respiratory monitoring data and abnormal behavior reports. The interface provides real-time data display function, showing the changes in the rabbit's respiratory frequency and breathing depth in the form of charts, and also displays the ambient temperature, humidity, and light intensity information. Users set alarm threshold parameters on the interface. When abnormal rabbit respiratory behavior occurs, the system sends a text message alarm to the administrator through the SMS gateway.

[0015] Optionally, the SMS content includes the rabbit number, the type of anomaly, and detailed information about the time of occurrence, and the alarm information is highlighted on the user interface so that the administrator can take timely measures.

[0016] Compared with the prior art, the present invention has at least one of the following beneficial technical effects: 1. Through the combination of depth and optical cameras and environmental perception functions, it provides accurate images for respiratory analysis and adapts to various environments to reduce environmental interference.

[0017] 2. Data transmission is efficient and stable. Wired high-speed transmission and wireless backup switching ensure timely and accurate data transmission, avoiding monitoring errors caused by transmission interruptions.

[0018] 3. Accurate data analysis and advanced algorithms enable accurate image preprocessing, accurate extraction of respiratory features, and accurate identification of multi-dimensional abnormal behaviors, improving the timeliness and accuracy of health problem detection.

[0019] 4. Data storage is reliable and user interaction is convenient, which is conducive to data management traceability and long-term health tracking, allowing breeders to monitor the health status of rabbits in real time and flexibly set alarms, thereby improving breeding management efficiency and rabbit survival rate.

[0020] The present invention optimizes image acquisition, data transmission, analysis, storage and interaction in all aspects, significantly improving the accuracy, reliability and convenience of identifying abnormal respiratory behavior in rabbits, effectively facilitating the refined management and health protection of rabbit breeding, and bringing considerable economic benefits and efficient management experience to breeders. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a functional block diagram of a system for automatically identifying and recording abnormal breathing behavior in rabbits. DETAILED DESCRIPTION

[0022] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0023] Embodiment: The present application discloses an automatic recognition and recording system for abnormal breathing behavior of rabbits, which is composed of an image acquisition module, a data transmission module, a data analysis module, and a data storage and interaction module. Each module is described in detail below.

[0024] 1. Image acquisition module Image acquisition unit The system uses a combination of a high-resolution, high-frame-rate depth camera and an ordinary optical camera. The depth camera has a resolution of 1280×720 pixels and a frame rate of 60fps, which can accurately obtain the three-dimensional information and depth data of the rabbit's body; the ordinary optical camera has a resolution of 1920×1080 pixels and a frame rate of 30fps, which is used to capture the texture and detailed features of the rabbit's body surface.

[0025] The camera is equipped with automatic exposure and automatic white balance functions, which can automatically adjust the exposure time and color balance according to the ambient light intensity to ensure stable image quality. The camera's lens focal length is 8-24mm zoom lens, which can adapt to shooting needs at different distances.

[0026] It is installed on a movable robotic arm, and the range of movement of the robotic arm is 0 to 2 meters in the X-axis direction, 0 to 1.5 meters in the Y-axis direction, and 0.5 to 1.2 meters in the Z-axis direction. The camera position can be automatically adjusted through a preset program to fully cover the rabbit's activity area. A high-resolution, high-frame-rate depth camera is combined with an ordinary optical camera. The depth camera can accurately obtain the three-dimensional information and depth data of the rabbit's body, and the ordinary optical camera can capture the texture and detail features of the rabbit's body surface. The two work together to provide a rich and accurate image data foundation for subsequent precise respiratory feature analysis. For example, when locating the rabbit's chest breathing area, depth information helps to more accurately determine its position and range, while texture details can assist in judging subtle changes in the body surface caused by breathing movements.

[0027] The camera is equipped with automatic exposure and automatic white balance functions and a variable-focus lens, which can automatically adjust the exposure time and color balance according to the ambient light intensity, adapt to the shooting needs of different distances, and ensure stable image quality. No matter how the lighting conditions of the breeding environment change, clear and accurate rabbit images can be obtained, reducing image errors caused by environmental factors and improving the accuracy of abnormal breathing behavior recognition.

[0028] Environmental Monitoring Unit Temperature and humidity sensor: A temperature and humidity sensor with an accuracy of ±0.3°C is selected to monitor the temperature and humidity of the rabbit breeding environment in real time. The measurement range is −10°C to 40°C and the humidity is 20% to 90%.

[0029] Light sensor: A light intensity sensor with a measurement range of 0 to 20,000 lux is used to sense the ambient light intensity so that the light factor can be compensated during image acquisition. The temperature and humidity sensor and light sensor in the environmental monitoring unit monitor the temperature, humidity and light intensity of the rabbit breeding environment in real time. On the one hand, these environmental data can be used to compensate for the light factor during image acquisition and reduce the interference of ambient light changes on image analysis; on the other hand, the combination of environmental data and rabbit breathing data helps to more comprehensively analyze the potential correlation between abnormal rabbit breathing behavior and environmental factors, and provide breeders with a more in-depth reference for breeding environment optimization and disease prevention.

[0030] 2. Data transmission module Wired transmission unit High-speed Ethernet cables (CAT6 and above) are used for data transmission, with a transmission rate of 1Gbps or more, to stably transmit the data collected by the image acquisition module to the server or local computer where the data analysis module is located. The TCP / IP protocol is used for data encapsulation and transmission to ensure the integrity and accuracy of the data. The wired transmission unit uses high-speed Ethernet cables for data transmission, follows the TCP / IP protocol, and has a transmission rate of 1Gbps or more. It can quickly and stably transmit a large amount of data collected by the image acquisition module to the server or local computer where the data analysis module is located, ensuring the integrity and accuracy of the data. In large-scale breeding scenarios, the image data of multiple rabbits can be transmitted in a timely manner to avoid data backlogs and transmission delays, providing strong support for real-time abnormal respiratory behavior recognition.

[0031] Wireless transmission unit It is also equipped with a Wi-Fi 6 wireless transmission module as a backup transmission mode, which automatically switches when the wired network fails. The Wi-Fi 6 module complies with the IEEE 802.11ax standard, with a theoretical transmission rate of up to 9.6Gbps. The transmission frequency band is 2.4GHz and 5GHz dual-band, and it can automatically switch the frequency band according to the signal strength and interference to ensure the stability of data transmission.

[0032] It is worth mentioning that the wired transmission unit and the wireless transmission unit can switch automatically, and the switching time is 8-10 seconds. The design of combining wired and wireless with automatic switching function ensures the stability and reliability of data transmission. Even if the wired network fails, the wireless transmission module can quickly take over the data transmission task, and the switching time does not exceed 10 seconds, ensuring that data transmission is not interrupted and avoiding missed detection or misjudgment of abnormal breathing behavior due to transmission interruption.

[0033] 3. Data Analysis Module Image preprocessing unit Image registration: The images captured by the depth camera and the optical camera are registered using a registration algorithm based on feature points. First, the SIFT (Scale Invariant Feature Transform) algorithm is used to extract the feature points in the two images. and feature points in the optical image , calculate the matching relationship between them. Set the error threshold of feature point matching Pixels, the matching error is optimized by the least squares method, so that in, is the depth image The three-dimensional coordinates of the feature points are expressed as ,in is the depth value, The optical image The two-dimensional coordinates of the feature points, is the error threshold of feature point matching, in pixels. is the number of matching feature point pairs, thus obtaining accurate image registration results.

[0034] Noise removal: For the registered image, the bilateral filtering algorithm is used to remove noise. The bilateral filtering formula is: in, It's pixels The intensity value after filtering, is a pixel The original intensity value, It's pixels Neighborhood of is a spatial domain Gaussian function, which is used to calculate the spatial distance between pixels. Calculate weights, standard deviation , is a range Gaussian function that is used to calculate the pixel intensity difference Calculate weights, standard deviation , is the normalization coefficient, The image registration subunit uses a registration algorithm based on feature points, extracts feature points through the SIFT algorithm, and optimizes the matching error using the least squares method. The error threshold is only 0.1 pixels, which can achieve accurate registration of depth camera and optical camera images, so that subsequent respiratory feature analysis is based on more accurate image fusion data. The noise removal subunit uses a bilateral filtering algorithm, which effectively removes noise from the image by reasonably setting the standard deviation of the spatial domain Gaussian function to 0.5 and the standard deviation of the value domain Gaussian function to 10, improves image quality, reduces the interference of noise on respiratory feature extraction, and further improves the accuracy of the analysis results.

[0035] Rabbit breathing feature extraction unit Respiratory region positioning: Using the depth image information, the rabbit's chest area is positioned as the main respiratory monitoring area through the region growing algorithm. First, a seed point in the chest area is selected. , set the growth threshold Meters (indicates the depth difference threshold), and then starting from the seed point, the adjacent pixels whose depth values ​​differ from the seed point depth value within the growth threshold range are merged into the growth area, and iterate continuously until no new pixels meet the conditions. The area threshold of the growth area is set to Square centimeters, by comparing with the preset rabbit body model, the accurate breathing area is further determined.

[0036] Respiratory frequency calculation: Analyze the image sequence of the respiratory area and use the optical flow method to calculate the movement speed of the respiratory area. Frame and +1 frame of pixels within the breathing area The optical flow vector is , solve the optical flow equation using the Horn-Schunck optical flow algorithm: in, is the image intensity function, which indicates the brightness value of a point in the image over time ( ) and space changes, , are optical flow vector components, representing the pixel points at and The speed of movement in the direction. Set the optical flow speed threshold Meters per second. When the optical flow speed of more than 30% of the pixels in the breathing area exceeds this threshold, it is recorded as a breathing action. ,in For in time (in seconds) the number of breathing movements detected, The total monitoring time is the normal rabbit respiratory rate range of 30 to 60 times per minute. Respiratory rate outside this range is considered abnormal.

[0037] Respiratory depth analysis: Based on the depth image sequence of the respiratory area, calculate the maximum change in the depth of the chest area during the breathing process. Set the depth change threshold If the chest cavity depth changes beyond this range during breathing, the breathing depth is judged to be abnormal.

[0038] The respiratory region positioning subunit uses the depth image information and the regional growing algorithm to locate the rabbit chest region as the main respiratory monitoring region. By setting the seed point depth difference growth threshold to 0.2 meters and the growth region area threshold to 100-300 square centimeters, and combining the preset rabbit body model for secondary screening, it can accurately determine the respiratory region and provide an accurate monitoring range for respiratory frequency and depth calculation. The respiratory frequency calculation subunit uses the optical flow method to calculate the movement speed of the respiratory region. According to the Horn-Schunck optical flow algorithm, the optical flow velocity threshold is set to 0.05 meters / second. Combined with the standard that the normal rabbit respiratory frequency range is 30-60 times per minute, it can accurately calculate the respiratory frequency and accurately judge the abnormal respiratory frequency. The respiratory depth analysis subunit sets the depth change threshold to 0.03-0.1 meters according to the depth image sequence of the respiratory region. It can accurately calculate the maximum change in the depth of the chest region during breathing and judge the abnormal respiratory depth. The comprehensive and accurate analysis of the respiratory frequency and depth can achieve comprehensive and accurate monitoring of the rabbit's respiratory status.

[0039] Abnormal Behavior Recognition Unit Respiratory rhythm judgment: Analyze the time interval sequence between consecutive breathing actions , calculate the autocorrelation function of the time interval:

[0040] in, is the autocorrelation function at the lag order The value at is used to measure the periodicity of the breathing time interval sequence. Time interval series The average value of is the total number of time interval sequences, is the lag order. Set the autocorrelation function threshold , when the autocorrelation function is in the lag order range When the values ​​of the above two parameters are less than the threshold, the respiratory rhythm is judged to be abnormal. The respiratory rhythm judgment subunit of the abnormal behavior recognition unit calculates the autocorrelation function of the time interval sequence between consecutive breathing actions, sets the autocorrelation function threshold to 0.3, and the lag order to 3-10, which can accurately judge the abnormal respiratory rhythm. Compared with the traditional single indicator judgment or simple algorithm analysis, this multi-dimensional respiratory feature analysis and abnormal behavior recognition method greatly improves the accuracy and reliability of abnormal respiratory behavior recognition, and can detect potential health problems of rabbits earlier, providing the possibility for timely treatment.

[0041] 4. Data Storage and Interaction Module Data storage unit A distributed file system (such as Ceph) is used to store raw image data and processed analysis data. The raw image data is stored by date, rabbit number and camera type, and the analysis data is stored in a relational database (such as PostgreSQL), including individual information, respiratory characteristic data, and abnormal behavior records of rabbits. The database table structure is reasonably designed to ensure efficient storage and query of data. The data storage time is not less than 1 year to meet the needs of long-term health tracking. The data storage unit uses a distributed file system to store raw image data, which is stored by date, rabbit number and camera type to facilitate data management and traceability. At the same time, a relational database is used to store analysis data, including individual information, respiratory characteristic data, abnormal behavior records of rabbits, etc. The database table structure is reasonably designed to ensure efficient storage and query of data. The data storage time is not less than 1 year, which meets the needs of long-term health tracking and provides breeders with historical data references on the respiratory health status of rabbits, which helps to analyze the development trends and laws of abnormal respiratory behavior of rabbits, thereby formulating more scientific breeding management strategies.

[0042] User interaction unit Develop a web-based user interface, where users can log in to the system through a browser to view the rabbit's respiratory monitoring data and abnormal behavior reports. The interface provides real-time data display function, which displays the rabbit's respiratory frequency and breathing depth changes in the form of charts (such as bar charts and line charts), and also displays environmental temperature, humidity, and light intensity information. Users can also set alarm threshold parameters on the interface. When abnormal rabbit breathing behavior occurs, the system sends an SMS alarm to the administrator through the SMS gateway. The SMS content includes the rabbit number, abnormal type, and detailed information on the time of occurrence, and the alarm information is highlighted on the user interface so that the administrator can take timely measures. User Interaction Unit Develop a web-based user interface, where users can log in to the system through a browser to view the rabbit's respiratory monitoring data and abnormal behavior reports. The interface provides real-time data display function, which displays the rabbit's respiratory frequency and breathing depth changes in the form of intuitive charts, and also displays environmental temperature, humidity, light intensity and other information, so that breeders can understand the real-time health status of rabbits and breeding environment parameters at a glance. Users can also flexibly set alarm threshold parameters on the interface according to actual needs. When abnormal breathing behavior occurs in rabbits, the system sends a detailed SMS alarm to the administrator through the SMS gateway, including information such as the rabbit number, abnormal type, and occurrence time, and highlights the alarm information on the user interface to facilitate administrators to take timely measures. This convenient user interaction method, whether locally or remotely, allows breeders to control the health dynamics of rabbits in real time, achieve refined breeding management, and improve breeding efficiency and rabbit survival rate.

[0043] The above specific embodiments are only several optional embodiments of the present invention. Based on the technical solutions of the present invention and the relevant inspirations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. A system for automatically identifying and recording abnormal breathing behavior of rabbits, characterized in that: include: An image acquisition module is used to obtain the image data and environmental data of the rabbit; A data transmission module, transmitting the data collected by the image acquisition module to the data analysis module; The data analysis module processes and analyzes the received data to identify abnormal breathing behavior of rabbits; The data storage and interaction module is used to store data and implement interactive operations with users.

2. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 1, characterized in that: The image acquisition module comprises: The image acquisition unit uses a combination of a high-resolution, high-frame-rate depth camera and an ordinary optical camera. The depth camera has a resolution of 1280×720 pixels and a frame rate of 60fps. The ordinary optical camera has a resolution of 1920×1080 pixels and a frame rate of 30fps. The camera is equipped with automatic exposure, automatic white balance functions and an 8-24mm variable-focus lens, and is installed on a movable robotic arm. The range of movement of the robotic arm in the X-axis direction is 0-2 meters, the Y-axis direction is 0-1.5 meters, and the Z-axis direction is 0.5-1.2 meters. The environmental monitoring unit includes a temperature and humidity sensor with an accuracy of ±0.3°C, a measurement range of −10°C to 40°C, and a humidity measurement range of 20% to 90%, and a light sensor with a measurement range of 0 to 20,000 lux.

3. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 1, characterized in that: The data transmission module comprises: Wired transmission unit uses high-speed Ethernet cable for data transmission, follows TCP / IP protocol, and has a transmission rate of 1Gbps or above; The wireless transmission unit is equipped with a Wi-Fi 6 wireless transmission module, complies with the IEEE 802.11ax standard, has a theoretical maximum transmission rate of 9.6Gbps, and a dual-band transmission band of 2.4GHz and 5GHz. The wired transmission unit and the wireless transmission unit can switch automatically, and the switching time is 8-10 seconds.

4. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 1, characterized in that: The data analysis module includes an image preprocessing unit, and the image preprocessing unit specifically includes: The image registration subunit registers the images captured by the depth camera and the optical camera using a registration algorithm based on feature points; The noise removal subunit uses a bilateral filtering algorithm to remove noise from the registered image.

5. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 4, characterized in that: In the image registration subunit, feature points in the two images are extracted by a scale-invariant feature transformation algorithm. and feature points in the optical image , calculate the matching relationship between them, and set the error threshold of feature point matching Pixels, the matching error is optimized by the least squares method, so that in, is the depth image The three-dimensional coordinates of the feature points are expressed as ,in is the depth value, The optical image The two-dimensional coordinates of the feature points, is the error threshold of feature point matching, in pixels. is the number of matched feature point pairs.

6. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 5, characterized in that: In the noise removal subunit, the bilateral filtering formula is: in, It's pixels The intensity value after filtering, is a pixel The original intensity value, It's pixels Neighborhood of is a spatial domain Gaussian function, which is used to calculate the spatial distance between pixels. Calculate weights, standard deviation , is a range Gaussian function that is used to calculate the pixel intensity difference Calculate weights, standard deviation , is the normalization coefficient, 。 7. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 1, characterized in that: The data analysis module also includes a rabbit breathing feature extraction unit, which includes: Breathing area positioning subunit, seed point , set the growth threshold Then, starting from the seed point, the adjacent pixels whose depth values ​​differ from the seed point depth values ​​within the growth threshold range are merged into the growth area, and the iteration is continued until no new pixel points meet the conditions. The area threshold of the growth area is set to Square centimeters, and the breathing area is further determined by comparing it with the preset rabbit body model; The respiratory frequency calculation subunit analyzes the image sequence of the respiratory area and uses the optical flow method to calculate the movement speed of the respiratory area. Frame and Pixels within the frame breathing area The optical flow vector is , solve the optical flow equation using the Horn-Schunck optical flow algorithm: in, is the image intensity function, which indicates the brightness value of a point in the image over time ( ) and space changes, , are optical flow vector components, representing the pixel points at and Movement speed in the direction, setting the optical flow speed threshold m / s. When the optical flow velocity of more than 30% of the pixels in the breathing area exceeds the optical flow velocity threshold, it is recorded as a breathing action. The breathing frequency ,in For in time The number of breathing movements detected within is the total monitoring time. The normal rabbit respiratory rate range is 30 to 60 times per minute. If it exceeds this range, it is considered abnormal. The breathing depth analysis subunit calculates the maximum change in the depth of the chest area during breathing based on the depth image sequence of the breathing area and sets the depth change threshold If the chest cavity depth change during breathing exceeds the depth change threshold, the breathing depth is judged to be abnormal.

8. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 1, characterized in that: The data analysis module also includes an abnormal behavior recognition unit, which is provided with a respiratory rhythm judgment subunit to analyze the time interval sequence between continuous respiratory actions. , calculate the autocorrelation function of the time interval: in, is the autocorrelation function at the lag order The value at is used to measure the periodicity of the breathing time interval sequence. Time interval series The average value of is the total number of time interval sequences, is the lag order, set the autocorrelation function threshold , when the autocorrelation function is in the lag order range When both values ​​are less than the threshold, the respiratory rhythm is judged to be abnormal.

9. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 1, characterized in that: The data storage and interaction module includes: The data storage unit uses a distributed file system to store raw image data and stores them by date, rabbit number and camera type, and uses a relational database to store analysis data, including individual rabbit information, respiratory characteristic data and abnormal behavior records; The user interaction unit develops a Web-based user interface. Users can log in through a browser to view the rabbit's respiratory monitoring data and abnormal behavior reports. The interface provides real-time data display function, showing the changes in the rabbit's respiratory frequency and breathing depth in the form of charts, and also displays the ambient temperature, humidity, and light intensity information. Users set alarm threshold parameters on the interface. When abnormal rabbit respiratory behavior occurs, the system sends a text message alarm to the administrator through the SMS gateway.

10. The system for automatically identifying and recording abnormal breathing behavior of rabbits according to claim 9, characterized in that: The content of the SMS includes the rabbit number, the abnormality type, and detailed information of the occurrence time, and the alarm information is highlighted on the user interface.

Images