Monitoring, regulating and controlling system and method for cattle pen environment

Abstract

The invention discloses a monitoring, regulating and controlling system and method for a cattle house environment. The system is provided with six units including a multi-dimensional heterogeneous environment parameter real-time acquisition unit and a binocular vision environment image acquisition unit. The acquisition unit acquires data such as temperature, humidity and gas concentration of a cattle house through multiple types of sensors, and the image acquisition unit acquires an environment image by using a binocular camera; after feature extraction and three-dimensional environment reconstruction, the environment state analysis unit comprehensively analyzes data in combination with a preset threshold value, and then the regulation and control execution unit controls ventilation equipment, temperature control equipment and the like to carry out environment regulation. The method comprises the steps of data acquisition, image acquisition, feature processing, environment analysis, equipment regulation and control and the like. According to the invention, comprehensive monitoring and accurate regulation and control of multiple parameters of the environment of the cattle pen are realized, the defects of single monitoring and static regulation and control strategy in the prior art are overcome, a suitable growth environment is created for cattle herds, and the breeding benefit is improved.

Description

Technical Field

[0001] The present invention relates to the field of cattle pen environment monitoring, and in particular to a monitoring and control system and method for cattle pen environment. Background Art

[0002] With the large-scale and intelligent development of animal husbandry, precise monitoring and control of the cattle pen environment is crucial to ensuring cattle health and improving farming efficiency. Cattle growth, reproduction, and production performance are influenced by a variety of factors, including ambient temperature and humidity, air quality, and light intensity. For example, high temperatures and high humidity can easily breed bacteria and cause cattle disease. Excessive concentrations of harmful gases such as ammonia and carbon dioxide can reduce cattle immunity. Traditional farming methods make it difficult to fully and accurately monitor the dynamic changes in the cattle pen environment in real time. Therefore, developing a system and method that can sense and intelligently control the cattle pen environment in real time has become an urgent need for the industry.

[0003] Existing environmental monitoring and control technologies for cattle pens have numerous shortcomings. For one thing, monitoring methods are limited and lack a systematic approach. Most systems rely on only a few sensors to monitor basic parameters such as temperature and humidity, failing to comprehensively monitor light intensity, air pressure, and hazardous gas concentrations. Furthermore, it is difficult to construct a complete data model of the cattle pen environment, resulting in inaccurate assessments of environmental conditions. Furthermore, control strategies lack dynamic adaptability. Traditional systems often rely on preset fixed thresholds for control and are unable to adjust control strategies in a timely manner based on dynamic factors such as real-time changes in the cattle pen environment, cattle activity status, and changes in spatial structure. This makes it prone to control lags or over-control, making it difficult to achieve accurate and efficient control of the cattle pen environment. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides a monitoring and control system and method for a cattle pen environment.

[0005] The technical solution adopted by the present invention is a monitoring and control system for cattle pen environment, comprising:

[0006] A multi-dimensional, heterogeneous environmental parameter real-time acquisition unit, which consists of multiple sensor groups distributed in different areas and at different heights within the cattle pen. The collected analog signals are transmitted to a signal conditioning module via a wired transmission line. The signal conditioning module amplifies and filters the analog signals, and then converts them into digital signals via an analog-to-digital conversion module.

[0007] The binocular vision environment image acquisition unit consists of two sets of binocular vision cameras with different baseline distances. The two sets of binocular vision cameras are installed at diagonal positions on the top of the cattle pen. The collected image raw data is transmitted to the image feature extraction unit through a high-speed data transmission interface;

[0008] The image feature extraction unit receives the original image data transmitted by the binocular visual environment image acquisition unit, uses a scale-invariant feature transformation algorithm based on Gaussian pyramid layering to perform multi-scale feature extraction on the original image data, describes the feature points, forms a feature vector, and transmits the feature vector to the three-dimensional environment reconstruction unit;

[0009] The 3D environment reconstruction unit receives the feature vector transmitted by the image feature extraction unit, matches the feature vectors of the images captured by the two sets of binocular vision cameras based on the feature matching algorithm in the dynamic environment perception algorithm, calculates the 3D coordinates of the feature points using the triangulation principle, constructs the 3D coordinates into a 3D point cloud model using the point cloud data processing algorithm, and transmits the 3D point cloud model to the environmental status analysis unit;

[0010] The environmental status analysis unit receives the three-dimensional point cloud model transmitted by the three-dimensional environmental reconstruction unit and the digital signal transmitted by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit. It conducts a comprehensive analysis of the temperature and humidity, gas concentration, light intensity, and spatial structure of the cattle pen environment based on the preset cattle pen environmental parameter threshold range, and transmits the analysis results to the control execution unit in the form of data instructions;

[0011] The control execution unit receives the data instructions transmitted by the environmental status analysis unit and controls the corresponding control equipment according to the content of the instructions, including controlling the ventilation equipment to adjust the gas circulation in the cowshed, controlling the temperature control equipment to adjust the temperature in the cowshed, controlling the lighting equipment to adjust the light intensity, and controlling the spraying equipment to adjust the humidity in the cowshed.

[0012] Furthermore, the binocular visual environment image acquisition unit adopts the following image exposure adaptive adjustment model formula based on the dynamic environment perception algorithm:

[0013]

[0014] Among them, E adj Represents the adjusted image exposure value; α is the light intensity compensation coefficient, which is obtained through machine learning algorithm training based on the historical light intensity data of the cattle pen and the camera imaging quality data, and the value range is 0.8-1.2; I i It represents the sum of the grayscale values ​​of all pixels in the current collected image, where n is the total number of image pixels; β is the distance compensation coefficient, which is set according to the relationship between the average distance of different monitoring targets in the cattle pen and the current target distance, and the value range is 0.9-1.1; Dist cam-obj Indicates the actual distance from the current camera to the monitored target, which is calculated based on the binocular vision triangulation principle; Dist avg It represents the average distance of all monitoring targets in the cattle pen, obtained through statistical analysis of the historical monitoring data of the cattle pen.

[0015] Furthermore, in the image feature extraction unit, a feature point dynamic screening model formula based on a dynamic environment perception algorithm and a binocular vision environment analysis model is adopted:

[0016]

[0017] Among them, P selected represents the set of filtered feature points; p is the feature point to be filtered; S(p) represents the stability index of feature point p, which is obtained by calculating the difference between the feature descriptors of the feature point at different scales and different viewing angles; γ is the screening threshold coefficient, which is obtained through experimental statistics based on the complexity and noise level of the image in the cattle pen environment, and the value range is 0.6-0.8; It represents the sum of the stability indices of all feature points in the current image, and m is the total number of feature points in the current image.

[0018] Furthermore, in the three-dimensional environment reconstruction unit, the point cloud data denoising model formula based on the dynamic environment perception algorithm is as follows:

[0019]

[0020] in, represents the denoised point cloud data set; Q represents the original point cloud data set; δ is the denoising intensity coefficient, which is determined experimentally based on the influence of equipment vibration and dust interference factors in the cattle pen environment, and the value range is 0.1-0.3; Represents each point q in the original point cloud data set k To point cloud centroid The sum of the distances, l is the total number of original point cloud data points; Represents point q k To point cloud centroid The Euclidean distance of .

[0021] Furthermore, in the environmental state analysis unit, the temperature and humidity comprehensive evaluation model formula based on the binocular visual environment analysis model and the dynamic environment perception algorithm is:

[0022]

[0023] Among them, T eval Represents the comprehensive evaluation value of temperature and humidity; θ T is the temperature weight coefficient, which is set according to the suitable temperature range for cattle growth and the actual temperature fluctuation of the cattle pen, and the value range is 0.6-0.8; T is the current temperature value of the cattle pen collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; θ His the humidity weight coefficient, which is set according to the suitable humidity range for cattle growth and the actual humidity fluctuation in the cowshed, and the value range is 0.2-0.4; H is the current humidity value of the cowshed collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; V is the current air circulation speed in the cowshed, which is collected by the wind speed sensor in the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; V std This is the standard air circulation speed in the cattle pen, which is determined according to the cattle pen design specifications and the growth needs of cattle.

[0024] Furthermore, in the control execution unit, the ventilation equipment control model formula based on the dynamic environment perception algorithm is as follows:

[0025]

[0026] Among them, F vent Indicates the regulated operating power of the ventilation equipment; C NH3 The current ammonia concentration value of the cowshed collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; C NH3-std is the standard value of ammonia concentration in the cattle pen; ΔC NH3 is the ammonia concentration control threshold range; C CO2 The current carbon dioxide concentration value of the cattle pen collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; C CO2-std is the standard value of carbon dioxide concentration in the cattle pen; ΔC CO2 F is the carbon dioxide concentration control threshold range; max The maximum power of the ventilation equipment.

[0027] Furthermore, the image data compression transmission model formula based on the dynamic environment perception algorithm is adopted between the binocular visual environment image acquisition unit and the image feature extraction unit:

[0028]

[0029] Among them, D compressed Indicates the amount of compressed image data; D original represents the amount of original image data; λ is the compression coefficient, which is determined by the adaptive adjustment algorithm according to the network transmission bandwidth and the complexity of the image features, and the value range is 0.1-0.5; Entropy (D original ) represents the information entropy of the original image data and is used to measure the complexity of the image data.

[0030] Furthermore, between the environmental state analysis unit and the control execution unit, the control instruction priority determination model formula based on the binocular visual environment analysis model is:

[0031]

[0032] Among them, P priorityIndicates the priority of the control instruction; Severity (S) indicates the severity of the abnormal situation of the current environmental state, which is obtained by weighted calculation of the degree and duration of deviation of environmental parameters from the standard value; S total Indicates the total number of all current environmental status abnormalities; Severity(s) indicates the severity of the sth environmental status abnormality.

[0033] Furthermore, in the three-dimensional environment reconstruction unit, the formula of the cattle pen spatial structure change detection model based on the dynamic environment perception algorithm and the binocular vision environment analysis model is as follows:

[0034]

[0035] Where ΔS represents the change in the spatial structure of the cattle pen; n points represents the number of feature points used to detect changes in spatial structure; p i-new represents the three-dimensional coordinates of the i-th feature point at the current moment; p i-old Indicates the three-dimensional coordinates of the i-th feature point at the last monitoring moment; Dist(p i-new , p i-old ) represents the Euclidean distance between the three-dimensional coordinates of the i-th feature point at the current moment and the previous monitoring moment.

[0036] A method for monitoring and controlling a cattle pen environment comprises the following steps:

[0037] Step S1: Through various types of sensor groups distributed in different areas and heights of the cattle pen in the multi-dimensional heterogeneous environmental parameter real-time acquisition unit, the temperature and humidity, ammonia concentration, carbon dioxide concentration, light intensity and air pressure analog signals in the cattle pen environment are continuously collected, and after processing by the signal conditioning module and the analog-to-digital conversion module, they are converted into digital signals and transmitted to the data fusion processing unit;

[0038] Step S2, using a binocular vision environment image acquisition unit consisting of two groups of binocular vision cameras with different baseline distances installed at the top diagonal position of the cowshed, collecting the cowshed environment image raw data, and transmitting the data to the image feature extraction unit through a high-speed data transmission interface;

[0039] Step S3: The image feature extraction unit performs multi-scale feature extraction on the received original image data using a scale-invariant feature transformation algorithm based on Gaussian pyramid layering, constructs different scale spaces and image pyramids, detects and describes stable feature points in the image, and forms feature vectors that are transmitted to the three-dimensional environment reconstruction unit.

[0040] Step S4: The three-dimensional environment reconstruction unit matches the feature vectors of the images captured by the two sets of binocular vision cameras based on the feature matching algorithm in the dynamic environment perception algorithm, calculates the three-dimensional coordinates of the feature points using the triangulation principle, constructs a three-dimensional point cloud model using the point cloud data processing algorithm, and transmits it to the environment state analysis unit;

[0041] Step S5: The environmental state analysis unit receives the three-dimensional point cloud model and the digital signal, performs a comprehensive analysis of the temperature and humidity, gas concentration, light intensity, and spatial structure of the cattle pen environment based on a preset cattle pen environmental parameter threshold range, and transmits the analysis results to the control execution unit in the form of data instructions;

[0042] Step S6: The control execution unit receives the data instruction and controls the corresponding control equipment according to the instruction content, including controlling the ventilation equipment, temperature control equipment, lighting equipment and spraying equipment to adjust the cattle pen environment.

[0043] Beneficial effects: The present invention proposes a monitoring and control system and method for cattle pen environment. The system is provided with a multi-dimensional heterogeneous environmental parameter real-time acquisition unit and a binocular vision environment image acquisition unit. The former collects multi-dimensional data such as temperature, humidity, and gas concentration through various types of sensors, and the latter obtains environmental images with the help of a binocular vision camera. The two work together to build a comprehensive and systematic cattle pen environmental data acquisition system to avoid environmental assessment deviations caused by missing monitoring parameters. In terms of the dynamic adaptability of the control strategy, the system integrates the dynamic environment perception algorithm and the binocular vision environment analysis model. The environmental state analysis unit conducts a comprehensive analysis of the collected multi-source data and three-dimensional environmental model based on the preset threshold interval to generate accurate data instructions. The control execution unit, based on the instructions and in combination with the control models of each control device, accurately controls ventilation, temperature control and other equipment for different environmental abnormal conditions, so as to achieve real-time adjustment of the control strategy as the cattle pen environment changes dynamically, avoiding lag or excessive control. At the same time, specific algorithms are used between the units in the system to realize functions such as adaptive adjustment of image exposure, dynamic screening of feature points, and point cloud data denoising, further improving the accuracy of monitoring data and the timeliness and effectiveness of regulation, creating a stable and suitable growth environment for cattle, and effectively improving breeding efficiency and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a diagram of the system unit composition of the present invention;

[0045] Figure 2 The figure is a flow chart of the method steps of the present invention. DETAILED DESCRIPTION

[0046] It should be noted that, unless there is a conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The application is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0047] like Figure 1 As shown, a monitoring and control system for cattle pen environment includes:

[0048] A multi-dimensional, heterogeneous environmental parameter real-time acquisition unit, which includes multiple sensor groups distributed in different areas and at different heights in the cattle pen. Each sensor group includes a temperature and humidity sensor, an ammonia concentration sensor, a carbon dioxide concentration sensor, a light intensity sensor, and an air pressure sensor. The collected analog signals are transmitted to the signal conditioning module via a wired transmission line. The signal conditioning module amplifies and filters the analog signals, and then converts them into digital signals through the analog-to-digital conversion module. Finally, the digital signals are transmitted to the data fusion processing unit.

[0049] Specifically, the multi-dimensional, heterogeneous environmental parameter real-time acquisition unit is the fundamental data source for the entire monitoring and control system. Its core significance lies in the comprehensive and real-time acquisition of various key parameters in the cattle pen environment, providing accurate data support for subsequent environmental analysis and control. Parameters such as temperature and humidity, ammonia concentration, carbon dioxide concentration, light intensity, and air pressure in the cattle pen environment all have a significant impact on the growth, health, and production performance of cattle. By integrating multiple types of sensors, this unit can achieve all-round monitoring of these environmental factors, allowing the system to accurately grasp the real-time status of the cattle pen environment, thereby providing a reliable basis for the formulation of subsequent environmental control strategies. Only by obtaining accurate and comprehensive environmental data can the entire system ensure efficient and precise environmental control.

[0050] During the specific implementation process, the unit distributes multiple sensor groups in different areas and at different heights of the cattle pen according to a specific layout. For example, sensor groups are deployed in the four corners and center of the cattle pen, as well as at different heights such as 0.5 meters, 1.5 meters, and 2.5 meters from the ground. Each sensor group includes a temperature and humidity sensor (measurement accuracy of temperature ±0.3°C, humidity ±2%RH), an ammonia concentration sensor (detection range 0-100ppm, accuracy ±1ppm), a carbon dioxide concentration sensor (detection range 0-5000ppm, accuracy ±10ppm), a light intensity sensor (measurement range 0-20000Lux, accuracy ±50Lux) and an air pressure sensor (measurement range 500-1100hPa, accuracy ±0.1hPa). The analog signal collected by the sensor is transmitted to the signal conditioning module through a wired transmission line composed of shielded twisted pair cables. The signal conditioning module amplifies the signal (the amplification factor is adjustable from 10 to 100 times) and filters it (using a low-pass filter with a cutoff frequency of 100 Hz) to enhance the stability and reliability of the signal. The analog signal is then converted into a digital signal through a 16-bit high-precision analog-to-digital conversion module and finally transmitted to the data fusion processing unit to complete the initial data collection and processing process.

[0051] The binocular vision environment image acquisition unit consists of two sets of binocular vision cameras with different baseline distances. The two sets of binocular vision cameras are installed at diagonal positions on the top of the cattle pen. Each set of binocular vision cameras includes a left camera and a right camera. Each camera is equipped with an optical image stabilization component and an automatic aperture adjustment device. The collected image raw data is transmitted to the image feature extraction unit through a high-speed data transmission interface.

[0052] Specifically, the binocular vision environment image acquisition unit in the system plays a crucial role in acquiring visual information about the cattle pen's spatial environment. Its significance lies in utilizing binocular vision technology to provide the system with rich cattle pen environmental image data, laying the foundation for subsequent environmental feature extraction, three-dimensional environmental reconstruction, and environmental state analysis. The configuration of two sets of binocular vision cameras with different baseline distances can simulate the visual perception principles of the human eye and acquire image information with parallax. This parallax information is a key element in the subsequent calculation of the three-dimensional coordinates of objects and the construction of three-dimensional environmental models. Furthermore, the optical image stabilization components and automatic aperture adjustment devices equipped on each camera effectively ensure that clear and stable images can be obtained even under complex environmental conditions in the cattle pen, such as lighting changes and equipment vibrations. This greatly improves the quality and availability of image data and ensures that the system can accurately perceive the cattle pen's environmental conditions.

[0053] During the specific implementation process, two sets of binocular vision cameras were installed at diagonal positions on the top of the cattle pen. The baseline distance of one set of cameras was set to 30 cm, and the baseline distance of the other set was set to 50 cm. Each set of binocular vision cameras consists of two industrial-grade cameras on the left and right, with a resolution of 4K (3840×2160 pixels) and a frame rate of 30fps. Each camera is equipped with a device with an optical image stabilization level of 4 levels and an automatic aperture adjustment range of F1.4-F16. The camera uses the Gigabit Ethernet interface as a high-speed data transmission interface to transmit the collected uncompressed RAW format image data to the image feature extraction unit using the UDP protocol. The transmission rate can reach 1Gbps, ensuring that the image data can be transmitted quickly and stably, reducing the impact of data delay on system performance, and providing timely and effective data support for subsequent processing.

[0054] The image feature extraction unit receives the original image data transmitted by the binocular visual environment image acquisition unit, uses a scale-invariant feature transformation algorithm based on Gaussian pyramid layering to extract multi-scale features from the original image data, constructs different scale spaces and image pyramids, detects stable feature points in the image, describes the feature points, forms feature vectors, and transmits the feature vectors to the three-dimensional environment reconstruction unit;

[0055] Specifically, the image feature extraction unit is the core module of the system for mining key information from image data. Its significance lies in extracting representative, stable and unique feature information from the original image data transmitted by the binocular vision environment image acquisition unit. This feature information is the core data basis for the subsequent reconstruction of the three-dimensional environment. The cattle pen environment is complex and changeable. The size, position, angle, etc. of the objects in the image may vary. By adopting a scale-invariant feature transformation algorithm based on the Gaussian pyramid layer, the unit can extract multi-scale features from the image and detect and describe stable feature points in the image at different scales. It can effectively deal with scale changes, perspective changes, and lighting changes of objects in the image, ensuring that the extracted feature points have good stability and recognition under various environmental conditions, providing reliable feature data support for the accurate construction of the three-dimensional environment model, and enabling the system to accurately perceive the spatial structure and object distribution of the cattle pen environment.

[0056] In specific implementation, the image feature extraction unit receives raw image data in RAW format with a resolution of 3840×2160 pixels and first constructs a structure consisting of five scale spaces and three image pyramid levels. During this construction process, the image is subjected to varying degrees of Gaussian blur (with standard deviations of 0.5, 1.0, 1.5, 2.0, and 2.5) and downsampling (with a ratio of 1:2) to generate a series of image layers with different resolutions. Within these layers, the DoG (Difference of Gaussian) operator is used to detect extreme points in the scale space as candidate feature points. Feature point locations are then precisely determined by fitting a three-dimensional quadratic function, while edge artifacts and low-contrast points are removed. For the retained feature points, the gradient histogram is calculated in eight directions within a 16×16 neighborhood centered at the feature point to generate a 128-dimensional feature vector. Finally, these feature vectors are packaged in a specific data format and transmitted to the 3D environment reconstruction unit via a high-speed PCI-e interface at a transfer rate of 5GBps, completing the entire image feature extraction process.

[0057] The 3D environment reconstruction unit receives the feature vector transmitted by the image feature extraction unit, matches the feature vectors of the images captured by the two sets of binocular vision cameras based on the feature matching algorithm in the dynamic environment perception algorithm, calculates the 3D coordinates of the feature points using the triangulation principle, constructs the 3D coordinates into a 3D point cloud model using the point cloud data processing algorithm, and transmits the 3D point cloud model to the environmental status analysis unit;

[0058] Specifically, the 3D environment reconstruction unit plays a key role in the system, transforming 2D image information into a 3D spatial model. Its significance lies in transforming the feature vectors transmitted by the image feature extraction unit into a 3D point cloud model of the cattle pen environment through a series of algorithmic processing. This allows for an intuitive and three-dimensional presentation of the cattle pen's spatial structure and object distribution, providing a visual and spatial data foundation for the environmental status analysis unit. Based on the feature matching algorithm in the dynamic environmental perception algorithm, this unit matches the feature vectors of images captured by two sets of binocular vision cameras, calculates the 3D coordinates of the feature points using the principle of triangulation, and finally constructs a 3D point cloud model using a point cloud data processing algorithm, achieving the conversion from 2D images to 3D space. This allows the system to comprehensively and in-depth analyze the cattle pen's environmental conditions from a spatial dimension, such as whether the spatial layout is reasonable and whether there are obstacles affecting cattle activity, providing a more accurate basis for subsequent environmental control decisions.

[0059] In actual implementation, the 3D environment reconstruction unit receives 128-dimensional feature vector data via a high-speed PCI-e interface. It then uses a Hamming distance-based feature matching algorithm to match the feature vectors of the two image sets, setting a matching threshold of 60 (Hamming distance) to select successfully matched feature point pairs. For each set of successfully matched feature points, the X, Y, and Z coordinates of the feature points in 3D space are calculated using triangulation principles based on parameters such as the binocular camera baseline distance (30 cm and 50 cm, respectively), focal length (12 mm), and shooting angle. All calculated 3D coordinate points are assembled into a raw point cloud data set. This raw point cloud data is then denoised and smoothed using a voxel grid filtering algorithm (with a voxel size of 0.05 m × 0.05 m × 0.05 m) to remove outliers and redundant points. Finally, the processed point cloud data is constructed into a complete and smooth three-dimensional point cloud model through the Poisson surface reconstruction algorithm, stored in PLY format, and transmitted to the environmental status analysis unit through the Gigabit Ethernet interface at a transmission rate of 1Gbps for subsequent environmental status analysis.

[0060] The environmental status analysis unit receives the three-dimensional point cloud model transmitted by the three-dimensional environmental reconstruction unit and the digital signal transmitted by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit. It conducts a comprehensive analysis of the temperature and humidity, gas concentration, light intensity, and spatial structure of the cattle pen environment based on the preset cattle pen environmental parameter threshold range, and transmits the analysis results to the control execution unit in the form of data instructions;

[0061] Specifically, the environmental status analysis unit is the core decision-making module of the entire monitoring and control system. Its significance lies in the organic integration and in-depth analysis of the three-dimensional point cloud model transmitted by the three-dimensional environmental reconstruction unit and the digital signal transmitted by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit. Through the preset threshold range of the cattle pen environmental parameters, a comprehensive and integrated assessment of the temperature and humidity, gas concentration, light intensity and spatial structure state of the cattle pen environment is carried out to determine whether the current environment is suitable for the growth and production activities of cattle, and generate corresponding data instructions based on the assessment results, providing a clear control basis for the control execution unit, ensuring that the cattle pen environment can be maintained in the best state in real time and dynamically, and protecting the health and production efficiency of the cattle herd.

[0062] In practice, the environmental status analysis unit receives 3D point cloud model data (PLY format) via a Gigabit Ethernet interface and digital signals from a multi-dimensional, heterogeneous environmental parameter real-time acquisition unit via an RS-485 bus. These data include temperature (accuracy ±0.3°C), humidity (accuracy ±2%RH), ammonia concentration (accuracy ±1ppm), carbon dioxide concentration (accuracy ±10ppm), light intensity (accuracy ±50lux), and air pressure (accuracy ±0.1hPa). The unit has preset threshold ranges for environmental parameters specific to different growth stages of cattle. For example, for calves, the optimal temperature range is 20-25°C and humidity is 60%-70%RH; for mature cattle, ammonia concentrations must be below 20ppm and carbon dioxide concentrations below 1500ppm. During the analysis process, the collected environmental parameters are first compared with the preset threshold ranges to determine whether each parameter is within the normal range. Simultaneously, spatial analysis of the 3D point cloud model is performed to detect structural issues such as objects blocking ventilation openings and areas of uneven lighting. Based on the comprehensive analysis results, data instructions are generated according to the preset decision rules. The instructions include the type of control equipment (ventilation equipment, temperature control equipment, lighting equipment, sprinkler equipment), operation mode (on, off, power adjustment) and specific operating parameters (such as the target temperature of the temperature control equipment, the speed of the ventilation equipment, etc.). Finally, the data instructions are transmitted to the control execution unit through the CAN bus at a transmission rate of 500kbps, driving the control equipment to perform the corresponding operation.

[0063] The control execution unit receives the data instructions transmitted by the environmental status analysis unit and controls the corresponding control equipment according to the content of the instructions, including controlling the ventilation equipment to adjust the gas circulation in the cowshed, controlling the temperature control equipment to adjust the temperature in the cowshed, controlling the lighting equipment to adjust the light intensity, and controlling the spraying equipment to adjust the humidity in the cowshed.

[0064] Specifically, the control execution unit, as the final actuator of the entire monitoring and control system, is responsible for receiving data instructions transmitted by the environmental status analysis unit and accurately and efficiently converting these instructions into actual control operations. By controlling the corresponding control equipment, it achieves precise adjustment of the cattle pen environment, ensuring that various parameters of the cattle pen environment can quickly and stably reach the optimal state for cattle growth and production. Whether it is adjusting the temperature and humidity of the cattle pen or controlling air circulation and light intensity, the control execution unit plays a vital role. It is the key link in ensuring that the entire system can effectively improve the cattle pen environment and enhance farming efficiency, and is directly related to the quality of life and production performance of the cattle herd.

[0065] In specific implementations, the control execution unit receives data instructions via the CAN bus at a transmission rate of 500kbps and then parses the instructions. For temperature control equipment, if the instruction requires temperature adjustment, the unit controls devices such as variable-frequency air conditioners or electric heaters to achieve temperature control by adjusting the operating power of the devices (the power adjustment range for air conditioners is 0.5-5kW, and the power adjustment range for electric heaters is 1-3kW) according to the target temperature in the instruction (accuracy ±0.5°C). For ventilation equipment, the unit controls axial fans or exhaust fans, adjusting the fan speed (the speed adjustment range is 500-2000rpm) according to the instruction to regulate air circulation in the cattle pen. For lighting equipment, the unit controls the LED lighting system, adjusting the light brightness (brightness adjustment range is 0% to 100%) and on / off status according to the instruction. For sprinkler equipment, the unit controls the water pump and nozzles, adjusting the water spray flow rate (the flow adjustment range is 0-5L / min) and spray time (the time adjustment range is 0-60s) according to the instruction to regulate the humidity in the cattle pen. The control execution unit monitors the operating status of the control equipment (such as current, voltage, speed and other parameters) in real time and feeds back to the environmental status analysis unit to form a closed-loop control to ensure the accuracy and stability of the cattle pen environment control.

[0066] The multi-dimensional heterogeneous environmental parameter real-time acquisition unit is connected to the data fusion processing unit, the data fusion processing unit is connected to the environmental state analysis unit, the binocular vision environment image acquisition unit is connected to the image feature extraction unit, the image feature extraction unit is connected to the three-dimensional environment reconstruction unit, the three-dimensional environment reconstruction unit is connected to the environmental state analysis unit, and the environmental state analysis unit is connected to the control execution unit.

[0067] Preferably, the binocular visual environment image acquisition unit adopts the following image exposure adaptive adjustment model formula based on the dynamic environment perception algorithm:

[0068]

[0069] Among them, E adjRepresents the adjusted image exposure value; α is the light intensity compensation coefficient, which is obtained through machine learning algorithm training based on the historical light intensity data of the cattle pen and the camera imaging quality data, and the value range is 0.8-1.2; I i It represents the sum of the grayscale values ​​of all pixels in the current collected image, where n is the total number of image pixels; β is the distance compensation coefficient, which is set according to the relationship between the average distance of different monitoring targets in the cattle pen and the current target distance, and the value range is 0.9-1.1; Dist cam-obj Indicates the actual distance from the current camera to the monitored target, which is calculated based on the binocular vision triangulation principle; Dist avg It represents the average distance of all monitoring targets in the cattle pen, obtained through statistical analysis of the historical monitoring data of the cattle pen.

[0070] Specifically, the lighting conditions in cattle pens are complex and variable. Natural daylight intensity fluctuates significantly with time of day and weather, while artificial lighting at night can also cause uneven illumination. Furthermore, the movement of cattle can obstruct light. If camera exposure parameters are fixed, this can easily lead to over- or underexposure in the image, resulting in loss of detail and blurred features, seriously affecting the accuracy of subsequent image analysis. This technology utilizes an adaptive image exposure adjustment mechanism based on a dynamic environmental perception algorithm. This mechanism dynamically adjusts the camera's exposure based on multiple dimensions, including the real-time light intensity in the cattle pen, the actual distance between the current camera and the monitored target, and the average distance of monitored targets in the pen. The light intensity compensation factor is trained using a machine learning algorithm based on historical light intensity data accumulated over time and corresponding camera image quality data. It corrects for exposure deviations caused by varying light intensity. The distance compensation factor is determined based on the relationship between the average distance of different monitored targets in the pen and the current target distance, adjusting for variations in image brightness caused by varying target distances. During the actual implementation process, the system first obtains ambient lighting data in real time through the light intensity sensor installed in the cattle pen. At the same time, it uses the binocular visual triangulation principle to accurately calculate the actual distance from the camera to the monitoring target, and combines the average distance of the monitoring target obtained from historical data statistics. According to specific calculation rules, the appropriate exposure value is obtained, and then the aperture size, shutter speed and other parameters of the camera are automatically adjusted to ensure that clear, detailed and moderately bright images can be obtained in various complex environments, providing a high-quality data foundation for subsequent image feature extraction, three-dimensional environment reconstruction and other links.

[0071] Preferably, the image feature extraction unit adopts a feature point dynamic screening model formula based on a dynamic environment perception algorithm and a binocular vision environment analysis model:

[0072]

[0073] Among them, P selected represents the set of filtered feature points; p is the feature point to be filtered; S(p) represents the stability index of feature point p, which is obtained by calculating the difference between the feature descriptors of the feature point at different scales and different viewing angles; γ is the screening threshold coefficient, which is obtained through experimental statistics based on the complexity and noise level of the image in the cattle pen environment, and the value range is 0.6-0.8; It represents the sum of the stability indices of all feature points in the current image, and m is the total number of feature points in the current image.

[0074] Specifically, cattle pen environments are complex, and images contain numerous feature points. However, not all of these feature points stably and accurately reflect environmental information. Furthermore, factors such as environmental noise and similar object textures can cause some feature points to become unstable or erroneous. If all of these feature points are used for subsequent processing, the resulting 3D reconstruction will be biased, affecting the accurate assessment of the cattle pen's environmental state. A dynamic feature point screening mechanism, based on a dynamic environmental perception algorithm and a binocular visual environment parsing model, quantitatively evaluates the stability of image feature points to identify truly representative and stable feature points. The stability index of a feature point is calculated by calculating the degree of difference in the feature descriptors of that feature point at different scales and viewpoints. The smaller the difference, the more stable the feature point. The screening threshold coefficient is determined based on the complexity and noise level of the cattle pen images, following extensive experimental statistical analysis. In actual implementation, after receiving the raw image data, the image feature extraction unit first applies a specific algorithm to perform multi-scale feature extraction on the image, obtaining a large number of candidate feature points. It then calculates the stability index for each candidate feature point and compares it to a screening criterion determined by the mean stability index of all feature points and a screening threshold coefficient. Only feature points with stability indices exceeding the screening criterion are retained to form the filtered feature point set. This process effectively removes redundant and unstable feature points, improving their quality and reliability. This provides accurate data support for subsequent three-dimensional environmental reconstruction and environmental status analysis based on these feature points, enabling the system to more accurately perceive changes in the cattle pen environment.

[0075] Preferably, in the three-dimensional environment reconstruction unit, the point cloud data denoising model formula based on the dynamic environment perception algorithm is as follows:

[0076]

[0077] in, represents the denoised point cloud data set; Q represents the original point cloud data set; δ is the denoising intensity coefficient, which is determined experimentally based on the influence of equipment vibration and dust interference factors in the cattle pen environment, and the value range is 0.1-0.3; Represents each point q in the original point cloud data set k To point cloud centroid The sum of the distances, l is the total number of original point cloud data points; Represents point q k To point cloud centroid The Euclidean distance of .

[0078] Specifically, due to interference factors such as equipment vibration, flying dust, and frequent cattle activity in the cattle pen environment, the process of collecting environmental data and reconstructing the 3D environment using binocular vision technology inevitably introduces a large number of noise points into the point cloud data. These noise points interfere with the accurate assessment of the cattle pen's true spatial structure and object distribution. If left unaddressed, they can lead to deviations in the constructed 3D point cloud model, affecting subsequent environmental state analysis and control decisions based on this model. The point cloud data denoising mechanism, based on a dynamic environmental perception algorithm, processes the raw point cloud data by analyzing the distance relationship between each point in the point cloud data and the point cloud's centroid. This denoising intensity coefficient is determined based on the degree of influence of interference factors in the cattle pen environment. The denoising intensity coefficient is determined through multiple experimental tests, simulating different interference scenarios, and combining it with actual application requirements. It determines the degree to which noise points are removed. During implementation, the system first calculates the centroid of the original point cloud data, then calculates the Euclidean distance from each point to the centroid. Based on specific calculation rules, points that deviate from the centroid by more than a certain threshold are identified as noise points and removed. The remaining points are then smoothed to produce a denoised point cloud data set. This denoised point cloud data more realistically and accurately reflects the spatial structure and object distribution of the cattle pen environment, providing reliable three-dimensional data for the environmental status analysis unit. This allows the system to assess the state of the cattle pen environment based on an accurate environmental model and develop effective control strategies.

[0079] Preferably, in the environmental state analysis unit, the temperature and humidity comprehensive evaluation model formula based on the binocular visual environment analysis model and the dynamic environment perception algorithm is:

[0080]

[0081] Among them, T eval Represents the comprehensive evaluation value of temperature and humidity; θ T is the temperature weight coefficient, which is set according to the suitable temperature range for cattle growth and the actual temperature fluctuation of the cattle pen, and the value range is 0.6-0.8; T is the current temperature value of the cattle pen collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; θ His the humidity weight coefficient, which is set according to the suitable humidity range for cattle growth and the actual humidity fluctuation in the cowshed, and the value range is 0.2-0.4; H is the current humidity value of the cowshed collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; V is the current air circulation speed in the cowshed, which is collected by the wind speed sensor in the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; V std This is the standard air circulation speed in the cattle pen, which is determined according to the cattle pen design specifications and the growth needs of cattle.

[0082] Specifically, the growth, development, and health of cattle are closely related to temperature and humidity. Temperature, humidity, and air velocity are mutually influential and interrelated. Therefore, assessing temperature or humidity alone cannot fully reflect the actual suitability of the cattle pen environment. A comprehensive temperature and humidity assessment mechanism, based on a binocular visual environmental analysis model and a dynamic environmental perception algorithm, incorporates multiple key environmental parameters, including temperature, humidity, and air velocity, into the assessment system. By setting temperature and humidity weighting coefficients tailored to cattle growth characteristics and combining them with standard air velocity, a quantitative assessment of the temperature and humidity environment in the cattle pen is performed. The temperature weighting coefficient is determined based on the optimal temperature range for cattle at different growth stages and the actual temperature fluctuations in the cattle pen, reflecting the importance of temperature in environmental assessment. The humidity weighting coefficient is determined based on the optimal humidity range for cattle growth and the actual humidity fluctuations in the cattle pen. In actual implementation, the system continuously and in real time acquires current temperature, humidity, and air velocity data from the cattle pen through a multi-dimensional, heterogeneous real-time environmental parameter acquisition unit. This data is then combined with pre-set weighting coefficients and the standard air velocity according to specific calculation rules to produce a comprehensive temperature and humidity assessment. By comparing this assessment value with the preset optimal temperature and humidity range, the system can accurately determine whether the current temperature and humidity environment in the cattle pen is suitable for cattle growth. If the assessment value exceeds the optimal range, the system will promptly identify the temperature and humidity abnormality and generate corresponding control instructions, driving the control execution unit to adjust the temperature and humidity, thereby ensuring that the temperature and humidity environment in the cattle pen is always optimal for cattle growth.

[0083] Preferably, in the control execution unit, the ventilation equipment control model formula based on the dynamic environment perception algorithm is as follows:

[0084]

[0085] Among them, F vent Indicates the regulated operating power of the ventilation equipment; C NH3 The current ammonia concentration value of the cowshed collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; C NH3-std is the standard value of ammonia concentration in the cattle pen; ΔC NH3 is the ammonia concentration control threshold range; C CO2The current carbon dioxide concentration value of the cattle pen collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; C CO2-std is the standard value of carbon dioxide concentration in the cattle pen; ΔC CO2 F is the carbon dioxide concentration control threshold range; max The maximum power of the ventilation equipment.

[0086] Specifically, the concentration of harmful gases such as ammonia and carbon dioxide in cattle pens directly impacts the health and productivity of cattle. Excessively high concentrations can weaken cattle's immunity and increase the risk of respiratory illness. Therefore, ventilation equipment must promptly and accurately adjust its operating power based on changes in harmful gas concentrations to maintain fresh air in the barn. The ventilation equipment control mechanism, based on a dynamic environmental sensing algorithm, uses real-time ammonia and carbon dioxide concentrations as its core basis. This is combined with pre-set ammonia and carbon dioxide concentration standards, the corresponding concentration control threshold ranges, and the maximum power of the ventilation equipment to determine the controlled operating power of the ventilation equipment through a specific calculation method. The ammonia and carbon dioxide concentration standards are determined based on the health needs of cattle growth and relevant industry standards, while the concentration control threshold ranges are set based on a comprehensive consideration of the barn environment and the equipment's control capabilities. During actual operation, the system continuously monitors the concentrations of ammonia and carbon dioxide in the cattle pen through a real-time, multi-dimensional, heterogeneous environmental parameter acquisition unit. Upon acquiring new concentration data, the system immediately calculates the power parameters for the ventilation equipment according to established calculation rules, substituting the current concentration value with the standard value, the threshold range, and the maximum power of the ventilation equipment. The control execution unit then controls the ventilation equipment based on this power parameter, ensuring that it operates at the appropriate power, efficiently exhausting harmful gases and introducing fresh air. This achieves precise control of the air quality in the cattle pen, creating a healthy and comfortable breathing environment for the cattle, effectively reducing the risk of cattle disease caused by air quality issues, and improving farming efficiency.

[0087] Preferably, the image data compression transmission model formula based on the dynamic environment perception algorithm is adopted between the binocular visual environment image acquisition unit and the image feature extraction unit:

[0088]

[0089] Among them, D compressed Indicates the amount of compressed image data; D original represents the amount of original image data; λ is the compression coefficient, which is determined by the adaptive adjustment algorithm according to the network transmission bandwidth and the complexity of the image features, and the value range is 0.1-0.5; Entropy (D original ) represents the information entropy of the original image data and is used to measure the complexity of the image data.

[0090] Specifically, environmental monitoring of cattle pens requires high-resolution images to capture sufficient environmental details. However, the sheer volume of high-resolution images creates extremely high demands on network transmission bandwidth. However, network conditions at actual cattle pens are often limited. Directly transmitting raw image data can easily lead to transmission delays and data loss, severely impacting the system's real-time performance and stability. This image data compression and transmission mechanism, based on a dynamic environmental perception algorithm, adaptively compresses raw image data by introducing a compression factor that correlates with the cattle pens network bandwidth and the complexity of image features. This factor, combined with information entropy, a measure of image data complexity, is then used to dynamically adjust the compression factor based on the real-time bandwidth of the cattle pens network and the complexity of the image features. The information entropy of the image data reflects the richness and complexity of the information contained in the image data. In practical implementation, after the binocular visual environment image acquisition unit captures the raw image data, the system first calculates its information entropy. Then, based on the available bandwidth of the cattle pens network, an adaptive adjustment algorithm is used to determine an appropriate compression factor. The raw image data is then compressed according to specific calculation rules, and the significantly reduced compressed data is transmitted via a high-speed data transmission interface. After receiving the compressed data, the image feature extraction unit decompresses it according to the corresponding decompression rules, restoring image data with near-original quality. This process significantly improves image data transmission efficiency, reduces network transmission pressure, and ensures that image data can be quickly and stably transmitted to the image feature extraction unit, providing timely data support for subsequent image feature extraction, 3D environment reconstruction, and other processing processes, ensuring the efficient operation of the entire monitoring and control system.

[0091] Preferably, the control instruction priority determination model formula between the environmental state analysis unit and the control execution unit based on the binocular visual environment analysis model is:

[0092]

[0093] Among them, P priority Indicates the priority of the control instruction; Severity (S) indicates the severity of the abnormal situation of the current environmental state, which is obtained by weighted calculation of the degree and duration of deviation of environmental parameters from the standard value; S total Indicates the total number of all current environmental status abnormalities; Severity(s) indicates the severity of the sth environmental status abnormality.

[0094] Specifically, a control instruction priority determination mechanism between the environmental state analysis unit and the control execution unit ensures efficient and orderly environmental control of the cattle pen. The cattle pen environment is complex and ever-changing, and multiple environmental anomalies may occur at the same time, such as excessive temperature, excessive ammonia concentration, and insufficient light intensity. Different environmental anomalies have varying degrees of impact on cattle health and growth. Failure to prioritize these anomalies can result in critical issues not being addressed promptly, impacting the effectiveness of control. The control instruction priority determination mechanism, based on a binocular visual environment analysis model, quantitatively assesses the severity of each environmental anomaly by weighting the degree to which environmental parameters deviate from their standard values ​​and the duration of the anomaly. This weighted calculation then prioritizes each control instruction based on the severity of all anomalies. The degree to which environmental parameters deviate from their standard values ​​reflects the severity of the current environmental anomaly, while the duration of the anomaly reflects the cumulative effect of the anomaly on the cattle. In actual implementation, when the environmental status analysis unit detects an environmental anomaly and generates a control instruction, the system first quantifies the severity of each environmental anomaly according to a specific weighted calculation rule. It then aggregates the severity of all anomalies and calculates the priority value of each control instruction under the current environmental conditions. Upon receiving the control instruction, the control execution unit prioritizes the highest-priority control instructions based on the priority value, ensuring that environmental issues with the greatest impact on cattle health are promptly addressed, achieving precise and efficient control of the cattle pen environment and maximizing the quality of the cattle's growth environment.

[0095] Preferably, in the three-dimensional environment reconstruction unit, the formula of the cattle pen spatial structure change detection model based on the dynamic environment perception algorithm and the binocular vision environment analysis model is as follows:

[0096]

[0097] Where ΔS represents the change in the spatial structure of the cattle pen; n points represents the number of feature points used to detect changes in spatial structure; p i-new represents the three-dimensional coordinates of the i-th feature point at the current moment; p i-old Indicates the three-dimensional coordinates of the i-th feature point at the last monitoring moment; Dist(p i-new , p i-old ) represents the Euclidean distance between the three-dimensional coordinates of the i-th feature point at the current moment and the previous monitoring moment.

[0098] Specifically, during daily cattle pen operations, changes in the pen's spatial structure may occur due to cattle movements, equipment relocation or addition, and facility damage. Failure to promptly detect these changes could impact the cattle's activity space, ventilation, and lighting conditions, negatively impacting their health and productivity. A mechanism for detecting changes in the pen's spatial structure, based on a dynamic environmental perception algorithm and a binocular visual environment parsing model, compares the three-dimensional coordinates of feature points in the pen's environment at different monitoring times to calculate the change in their spatial position, thereby quantitatively assessing the overall change in the pen's spatial structure. Feature points used to detect spatial structure changes are selected during the three-dimensional environment reconstruction process to reliably reflect the characteristics of the pen's spatial structure. In practice, the three-dimensional environment reconstruction unit creates three-dimensional models of the pen's environment at different time points, obtaining the three-dimensional coordinate data for each feature point at each moment. When detecting spatial structure changes, the system compares the three-dimensional coordinates of the feature point at the current moment with the corresponding feature point at the previous monitoring moment. The Euclidean distance between the two points is calculated to determine the position change of each feature point. The position changes of all feature points are then accumulated to quantify the change in spatial structure. Based on this value, the system can determine whether the spatial structure of the cattle pen has changed significantly, and the extent of the change. If the system detects that the spatial structure changes beyond a preset threshold, it will issue a timely warning and transmit the relevant information to the environmental status analysis unit for further analysis of the cause of the change and the development of appropriate control strategies to ensure the stability and suitability of the cattle pen environment.

[0099] like Figure 2 As shown, a method for monitoring and controlling a cattle pen environment comprises the following steps:

[0100] Step S1: Through various types of sensor groups distributed in different areas and heights of the cattle pen in the multi-dimensional heterogeneous environmental parameter real-time acquisition unit, the temperature and humidity, ammonia concentration, carbon dioxide concentration, light intensity and air pressure analog signals in the cattle pen environment are continuously collected, and after processing by the signal conditioning module and the analog-to-digital conversion module, they are converted into digital signals and transmitted to the data fusion processing unit;

[0101] Step S2, using a binocular vision environment image acquisition unit consisting of two groups of binocular vision cameras with different baseline distances installed at the top diagonal position of the cowshed, collecting the cowshed environment image raw data, and transmitting the data to the image feature extraction unit through a high-speed data transmission interface;

[0102] Step S3: The image feature extraction unit performs multi-scale feature extraction on the received original image data using a scale-invariant feature transformation algorithm based on Gaussian pyramid layering, constructs different scale spaces and image pyramids, detects and describes stable feature points in the image, and forms feature vectors that are transmitted to the three-dimensional environment reconstruction unit.

[0103] Step S4: The three-dimensional environment reconstruction unit matches the feature vectors of the images captured by the two sets of binocular vision cameras based on the feature matching algorithm in the dynamic environment perception algorithm, calculates the three-dimensional coordinates of the feature points using the triangulation principle, constructs a three-dimensional point cloud model using the point cloud data processing algorithm, and transmits it to the environment state analysis unit;

[0104] Step S5: The environmental state analysis unit receives the three-dimensional point cloud model and the digital signal, performs a comprehensive analysis of the temperature and humidity, gas concentration, light intensity, and spatial structure of the cattle pen environment based on a preset cattle pen environmental parameter threshold range, and transmits the analysis results to the control execution unit in the form of data instructions;

[0105] Step S6: The control execution unit receives the data instruction and controls the corresponding control equipment according to the instruction content, including controlling the ventilation equipment, temperature control equipment, lighting equipment and spraying equipment to adjust the cattle pen environment.

[0106] A monitoring and control system and method for cattle pen environments utilizes a dual system for real-time acquisition of multi-dimensional, heterogeneous environmental parameters and binocular visual image acquisition. By deploying multiple sensors at different locations and heights within the cattle pen, comprehensive data such as temperature and humidity, hazardous gas concentrations, and light intensity are collected. Simultaneously, two sets of binocular visual cameras, mounted at opposite corners of the top, capture images of the cattle pen environment. These two data collection methods complement each other, enabling the system to comprehensively understand the cattle pen environment. Compared to traditional monitoring methods that rely on only a few sensors, this significantly improves data integrity and accuracy, avoiding environmental assessment biases caused by missing monitoring parameters.

[0107] To address the lack of dynamic adaptability in control strategies, the system deeply integrates dynamic environmental perception algorithms with binocular visual environment analysis models. The environmental state analysis unit, combined with preset environmental parameter thresholds, comprehensively analyzes collected multi-source data and the three-dimensional environmental model to generate precise data instructions. Upon receiving these instructions, the control execution unit, based on a specific control model, precisely controls ventilation, temperature control, lighting, and other equipment according to various environmental anomalies. Whether it's subtle fluctuations in environmental parameters or changes in the spatial structure of the cattle pen, the system can respond quickly and adjust the control strategy in real time, avoiding lags or over-control, thus achieving dynamic and precise control of the cattle pen environment.

[0108] Furthermore, the system has been optimized in data processing and transmission. The image feature extraction unit uses advanced algorithms for feature extraction and screening, while the 3D environment reconstruction unit performs data denoising to ensure data reliability. During data transmission, an image data compression transmission model improves data transmission efficiency. Furthermore, a control command priority determination model rationally schedules control tasks, making the system more efficient and organized. The combined application of these technologies has enabled the system to demonstrate enhanced performance and reliability in cattle pen environmental monitoring and control, providing strong support for the development of modern animal husbandry.

[0109] In the description of the present invention, it should be noted that, unless otherwise expressly specified or limited, the terms "disposed," "installed," "connected," "connected," and "fixed" should be understood in a broad sense. For example, they may refer to fixed connections, detachable connections, or integral connections; they may refer to mechanical connections or electrical connections; they may refer to direct connections or indirect connections through an intermediate medium; and they may refer to internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0110] While embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A monitoring and control system for cattle pen environment, characterized in that: include: The multi-dimensional heterogeneous environmental parameter real-time acquisition unit includes multiple sensor groups distributed in different areas and at different heights of the cattle pen. The collected analog signals are transmitted to the signal conditioning module via a wired transmission line. The signal conditioning module amplifies and filters the analog signals, and then converts them into digital signals through the analog-to-digital conversion module. A binocular visual environment image acquisition unit includes at least two groups of binocular vision cameras with different baseline distances. The binocular vision cameras are respectively installed at diagonal positions on the top of the cattle pen, and the collected image raw data is transmitted to the image feature extraction unit through a high-speed data transmission interface; The image feature extraction unit receives the original image data transmitted by the binocular visual environment image acquisition unit, uses a scale-invariant feature transformation algorithm based on Gaussian pyramid layering to perform multi-scale feature extraction on the original image data, describes the feature points, forms a feature vector, and transmits the feature vector to the three-dimensional environment reconstruction unit; The three-dimensional environment reconstruction unit receives the feature vector transmitted by the image feature extraction unit, matches the feature vectors of the images captured by the two sets of binocular vision cameras based on the feature matching algorithm in the dynamic environment perception algorithm, calculates the three-dimensional coordinates of the feature points using the triangulation principle, constructs the three-dimensional coordinates into a three-dimensional point cloud model through the point cloud data processing algorithm, and transmits the three-dimensional point cloud model to the environmental status analysis unit.

2. A monitoring and control system for cattle pen environment according to claim 1, characterized in that: The system also includes: The environmental status analysis unit receives the three-dimensional point cloud model transmitted by the three-dimensional environmental reconstruction unit and the digital signal transmitted by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit. It conducts a comprehensive analysis of the temperature and humidity, gas concentration, light intensity, and spatial structure of the cattle pen environment based on the preset cattle pen environmental parameter threshold range, and transmits the analysis results to the control execution unit in the form of data instructions; The control execution unit receives the data instructions transmitted by the environmental status analysis unit and controls the corresponding control equipment according to the instructions, including controlling the ventilation equipment to adjust the air circulation in the cattle pen, controlling the temperature control equipment to adjust the temperature of the cattle pen, controlling the lighting equipment to adjust the light intensity, and controlling the spraying equipment to adjust the humidity of the cattle pen; In the binocular visual environment image acquisition unit, the following image exposure adaptive adjustment model formula based on the dynamic environment perception algorithm is adopted: Among them, E adj Represents the adjusted image exposure value; α is the light intensity compensation coefficient, which is obtained through machine learning algorithm training based on the historical light intensity data of the cattle pen and the camera imaging quality data, and the value range is 0.8-1.2; It represents the sum of the grayscale values ​​of all pixels in the current collected image, where n is the total number of image pixels; β is the distance compensation coefficient, which is set according to the relationship between the average distance of different monitoring targets in the cattle pen and the current target distance, and the value range is 0.9-1.1; Dist cam-obk Indicates the actual distance from the current camera to the monitored target, which is calculated based on the binocular vision triangulation principle; Dist avg It represents the average distance of all monitoring targets in the cattle pen, obtained through statistical analysis of the historical monitoring data of the cattle pen.

3. A monitoring and control system for cattle pen environment according to claim 1, characterized in that: In the image feature extraction unit, a dynamic feature point screening model formula based on a dynamic environment perception algorithm and a binocular vision environment analysis model is adopted: Among them, P selected represents the set of filtered feature points; p is the feature point to be filtered; S(p) represents the stability index of feature point p, which is obtained by calculating the difference between the feature descriptors of the feature point at different scales and different viewing angles; γ is the screening threshold coefficient, which is obtained through experimental statistics based on the complexity and noise level of the image in the cattle pen environment, and the value range is 0.6-0.8; It represents the sum of the stability indices of all feature points in the current image, and m is the total number of feature points in the current image.

4. A monitoring and control system for cattle pen environment according to claim 1, characterized in that: In the three-dimensional environment reconstruction unit, the point cloud data denoising model formula based on the dynamic environment perception algorithm is as follows: in, represents the denoised point cloud data set; Q represents the original point cloud data set; δ is the denoising intensity coefficient, which is determined experimentally based on the influence of equipment vibration and dust interference factors in the cattle pen environment, and the value range is 0.1-0.3; Represents each point q in the original point cloud data set k To point cloud centroid The sum of the distances, l is the total number of original point cloud data points; Represents point q k To point cloud centroid The Euclidean distance of .

5. A monitoring and control system for cattle pen environment according to claim 2, characterized in that: In the environmental state analysis unit, the temperature and humidity comprehensive evaluation model formula based on the binocular visual environment analysis model and the dynamic environment perception algorithm is: Among them, T eval Represents the comprehensive evaluation value of temperature and humidity; θ T is the temperature weight coefficient, which is set according to the suitable temperature range for cattle growth and the actual temperature fluctuation of the cattle pen, and the value range is 0.6-0.8; T is the current temperature value of the cattle pen collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; θ H is the humidity weight coefficient, which is set according to the suitable humidity range for cattle growth and the actual humidity fluctuation in the cowshed, and the value range is 0.2-0.4; H is the current humidity value of the cowshed collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; V is the current air circulation speed in the cowshed, which is collected by the wind speed sensor in the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; V std This is the standard air circulation speed in the cattle pen, which is determined according to the cattle pen design specifications and the growth needs of cattle.

6. A monitoring and control system for cattle pen environment according to claim 2, characterized in that: In the control execution unit, the ventilation equipment control model formula based on the dynamic environment perception algorithm is as follows: Among them, F vent Indicates the regulated operating power of the ventilation equipment; C NH3 The current ammonia concentration value of the cowshed collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; C NH3-std is the standard value of ammonia concentration in the cattle pen; ΔC NH3 is the ammonia concentration control threshold range; C CO2 The current carbon dioxide concentration value of the cattle pen collected by the multi-dimensional heterogeneous environmental parameter real-time acquisition unit; C CO2-std is the standard value of carbon dioxide concentration in the cattle pen; ΔC CO2 F is the carbon dioxide concentration control threshold range; max The maximum power of the ventilation equipment.

7. A monitoring and control system for cattle pen environment according to claim 1, characterized in that: The image data compression transmission model formula based on the dynamic environment perception algorithm is adopted between the binocular visual environment image acquisition unit and the image feature extraction unit: Among them, D compressed Indicates the amount of compressed image data; D original represents the amount of original image data; λ is the compression coefficient, which is determined by the adaptive adjustment algorithm according to the network transmission bandwidth and the complexity of the image features, and the value range is 0.1-0.5; Entropy (D original ) represents the information entropy of the original image data and is used to measure the complexity of the image data.

8. A monitoring and control system for cattle pen environment according to claim 2, characterized in that: The control instruction priority determination model formula between the environmental state analysis unit and the control execution unit based on the binocular visual environment analysis model is: Among them, P priority Indicates the priority of the control instruction; Severity (S) indicates the severity of the abnormal situation of the current environmental state, which is obtained by weighted calculation of the degree and duration of deviation of environmental parameters from the standard value; S total Indicates the total number of all current environmental status abnormalities; Severity(s) indicates the severity of the sth environmental status abnormality.

9. A monitoring and control system for cattle pen environment according to claim 1, characterized in that: In the three-dimensional environment reconstruction unit, the formula for detecting changes in the spatial structure of a cattle pen based on the dynamic environment perception algorithm and the binocular vision environment analysis model is as follows: Where ΔS represents the change in the spatial structure of the cattle pen; n points represents the number of feature points used to detect changes in spatial structure; p i-new represents the three-dimensional coordinates of the i-th feature point at the current moment; p i-old Indicates the three-dimensional coordinates of the i-th feature point at the last monitoring moment; Dist(p i-new , p i-old ) represents the Euclidean distance between the three-dimensional coordinates of the i-th feature point at the current moment and the previous monitoring moment.

10. A method for monitoring and controlling a cattle pen environment, characterized in that: The following steps are involved: Step S1: Through various types of sensor groups distributed in different areas and heights of the cattle pen in the multi-dimensional heterogeneous environmental parameter real-time acquisition unit, the temperature and humidity, ammonia concentration, carbon dioxide concentration, light intensity and air pressure analog signals in the cattle pen environment are continuously collected, and after processing by the signal conditioning module and the analog-to-digital conversion module, they are converted into digital signals and transmitted to the data fusion processing unit; Step S2, using a binocular vision environment image acquisition unit consisting of two groups of binocular vision cameras with different baseline distances installed at the top diagonal position of the cowshed, collecting the cowshed environment image raw data, and transmitting the data to the image feature extraction unit through a high-speed data transmission interface; Step S3: The image feature extraction unit performs multi-scale feature extraction on the received original image data using a scale-invariant feature transformation algorithm based on Gaussian pyramid layering, constructs different scale spaces and image pyramids, detects and describes stable feature points in the image, and forms feature vectors that are transmitted to the three-dimensional environment reconstruction unit. Step S4: The three-dimensional environment reconstruction unit matches the feature vectors of the images captured by the two sets of binocular vision cameras based on the feature matching algorithm in the dynamic environment perception algorithm, calculates the three-dimensional coordinates of the feature points using the triangulation principle, constructs a three-dimensional point cloud model using the point cloud data processing algorithm, and transmits it to the environment state analysis unit; Step S5: The environmental state analysis unit receives the three-dimensional point cloud model and the digital signal, performs a comprehensive analysis of the temperature and humidity, gas concentration, light intensity, and spatial structure of the cattle pen environment based on a preset cattle pen environmental parameter threshold range, and transmits the analysis results to the control execution unit in the form of data instructions; Step S6: The control execution unit receives the data instruction and controls the corresponding control equipment according to the instruction content, including controlling the ventilation equipment, temperature control equipment, lighting equipment and spraying equipment to adjust the cattle pen environment.

Images