Multi-modal data fusion intelligent weighing equipment for shepherds

Abstract

The invention discloses intelligent weighing equipment with multi-modal data fusion for shepherds. The intelligent weighing equipment comprises a slope-type cage-shaped weighing platform, a physical sign monitoring module, an image recognition module, a communication and positioning module and a low-power-consumption energy management module, wherein the physical sign monitoring module, the image recognition module, the communication and positioning module and the low-power-consumption energy management module are arranged in the slope-type cage-shaped weighing platform and are in communication connection with one another; the physical sign monitoring module comprises a non-contact infrared temperature measurement sensor, a PPG heart rate sensor and a triaxial accelerometer; the image recognition module comprises a wide-angle camera, an infrared light supplement lamp and an edge calculation unit; the communication and positioning module comprises a dual-mode communication unit and a dual-mode positioning unit; the power consumption energy management module comprises a flexible solar panel, a vibration energy recovery unit and an intelligent power distribution controller.

Description

Technical Field

[0001] The present invention relates to the field of smart animal husbandry technology, and in particular to an intelligent weighing device for sheep herding that integrates weighing, vital sign monitoring, image recognition, and low-power communication, and uses multimodal data fusion to achieve sheep growth trend prediction and health management through multimodal data fusion and edge computing. Background Art

[0002] With rising living standards and a growing demand for meat products, the livestock industry has developed rapidly. Lamb, in particular, has become increasingly popular due to its unique texture and rich nutritional profile. However, traditional sheep farming management methods still suffer from numerous shortcomings, such as limited functionality, complex deployment, limited real-time performance, and low energy efficiency. Currently, sheep farms rely primarily on manual weighing to obtain sheep weight data. This method is not only time-consuming and labor-intensive, but can also easily cause stress reactions in the sheep, affecting the accuracy of the weighing results. Furthermore, existing weighing equipment typically only records weight and cannot simultaneously collect vital signs (such as temperature and heart rate) and image data, resulting in fragmented data and an incomplete reflection of the sheep's health. Furthermore, traditional weighing equipment often requires the coordinated operation of multiple devices, such as weighing platforms, cameras, and gateways. This not only increases installation and maintenance costs but also reduces system reliability. With the development of intelligent technology, the livestock farming industry is also constantly innovating. However, existing intelligent weighing systems still have some challenges, such as requiring relevant technical and professional knowledge to operate, which is difficult for ordinary people. Furthermore, sensors are susceptible to environmental influences, which limits the accuracy of the weighing system. Furthermore, current weighing equipment often relies on cloud-based data processing. Due to network latency, early warning information often lags, hindering the timeliness of farming decisions. Therefore, there is an urgent need for intelligent weighing equipment that can integrate multimodal data and perform real-time analysis and prediction to improve the efficiency and accuracy of sheep farming. Summary of the Invention

[0003] The present invention aims to overcome the deficiencies of the prior art and provide an intelligent weighing device with multimodal data fusion for sheep herding.

[0004] In order to achieve the above object, the technical solution provided by the present invention is: The multimodal data fusion intelligent weighing equipment for shepherding comprises a sloped cage-shaped weighing platform (1) and a vital sign monitoring module, an image recognition module, a communication and positioning module, and a low-power energy management module arranged in the sloped cage-shaped weighing platform (1) and communicatively connected to each other; the vital sign monitoring module comprises a non-contact infrared temperature sensor (4), a PPG heart rate sensor (5), and a three-axis accelerometer (6); the image recognition module comprises a wide-angle camera (7), an infrared fill light (8), and an edge computing unit (9); the communication and positioning module comprises a dual-mode communication unit (10) and a dual-mode positioning unit (11); and the power consumption energy management module comprises a flexible solar panel (12), a vibration energy recovery unit (13), and an intelligent power distribution controller (14).

[0005] Preferably, the sloped cage-shaped weighing platform (1) comprises a cage-shaped frame (101) with an open end and a sloped weighing plate (102) arranged at the bottom of the cage-shaped frame (101), wherein the sloped weighing plate (102) has a built-in strain gauge sensor (2) and a dynamic weighing controller (3).

[0006] More preferably, the inclination angle of the sloped weighing plate (102) is 15°; and a Kalman filter algorithm is integrated into the dynamic weighing controller (3).

[0007] Preferably, the non-contact infrared temperature sensor (4), PPG heart rate sensor (5) and three-axis accelerometer (6) respectively collect the sheep's body temperature, heart rate and activity intensity data in real time.

[0008] Preferably, the wide-angle camera (7) has a resolution of 1080P@30fps; and the edge computing unit (9) is preloaded with a lightweight YOLOv5 model for real-time analysis of the waist-to-hip ratio and backfat thickness of sheep.

[0009] Preferably, the dual-mode communication unit (10) supports LoRa and NB-IoT protocols, and the dual-mode positioning unit (11) supports Beidou and GPS positioning.

[0010] Preferably, the multimodal data fusion intelligent weighing equipment for shepherding further includes an alloy electromagnetic shielding layer wrapping each component.

[0011] The present invention will be further described below The multimodal data fusion intelligent weighing equipment for sheep herding described in this invention includes a weighing module, a vital sign monitoring module, an image recognition module, a communication and positioning module, and a low-power energy management module. The weighing module uses a strain gauge sensor embedded in a sloped weighing platform to guide sheep into the equipment in a one-way manner. The module also includes a dynamic weighing algorithm that uses Kalman filtering to eliminate motion interference and achieve high-precision weighing. The weighing controller controls the operating state of the weighing platform, effectively preventing the impact of invalid data on data collection during the weight increase phase when beef cattle enter the weighing platform and the weight decrease phase when beef cattle leave the weighing platform. The vital sign monitoring module integrates a non-contact infrared temperature sensor and a PPG heart rate sensor to monitor the sheep's body temperature and heart rate in real time. The module also includes a three-axis accelerometer to monitor the activity intensity of sheep as they pass. The image recognition module is equipped with a wide-angle camera and infrared fill light. It runs a lightweight YOLOv5 model through an edge computing unit to extract body condition scores such as waist-to-hip ratio and backfat thickness in real time, and supports day and night shooting. The communication and positioning module utilizes dual-mode communication and positioning technologies to ensure reliable and real-time data transmission. LoRa communication technology boasts a range of up to 10 km, while NB-IoT communication technology offers low power consumption. The combined use of Beidou and GPS dual-mode positioning achieves accuracy within 5 meters. The low-power energy management module incorporates a dual-mode power supply system, including flexible solar panels and a vibration energy recovery module, achieving low-power operation through intelligent power distribution strategies. The device enters deep sleep mode when no sheep are present and wakes only upon detecting proximity. The device also incorporates electromagnetic shielding, using μ-metal alloy to encase the weighing sensor and wireless module, effectively reducing EMI interference. The device has an IP68 protection rating, operates in temperatures ranging from -30°C to 60°C, and features a wind and sand repellent coating.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. By organically integrating multiple modal data, including weighing, vital sign monitoring, and image recognition, the system overcomes the limitations of traditional weighing equipment, which only records weight, and achieves comprehensive monitoring of sheep health, improving data integrity and accuracy. 2. The system adopts an integrated hardware architecture design, integrating functional modules such as the weighing platform, vital sign sensors, and cameras, reducing the number of devices, lowering installation and maintenance costs, and improving system reliability and stability. 3. Localized data processing based on edge computing units eliminates reliance on cloud servers, overcomes the problem of delayed warnings caused by network latency, enables real-time analysis and decision-making, and improves the timeliness of livestock farming management. 4. The use of dual-mode communication (LoRa + NB-IoT) and dual-mode positioning (Beidou + GPS) technologies enhances the device's communication and positioning capabilities, improves reliable data transmission and real-time performance, and expands the scope of unmanned applications. 5. Through the effective integration and intelligent analysis of multi-source data, the system accurately predicts sheep growth trends, providing a scientific basis for sheep farming, supporting personalized health management and precise feeding decisions, and significantly improving farming efficiency and economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 : Schematic diagram of the structure of the present invention.

[0014] In the figure: 1. Sloped cage-shaped weighing platform; 2. Strain gauge sensor; 3. Dynamic weighing controller; 4. Non-contact infrared temperature sensor; 5. PPG heart rate sensor; 6. Three-axis accelerometer; 7. Wide-angle camera; 8. Infrared fill light; 9. Edge computing unit; 10. Dual-mode communication unit; 11. Dual-mode positioning unit; 12. Flexible solar panel; 13. Vibration energy recovery unit; 14. Intelligent power distribution controller; 101. Cage rack; 102. Sloped weighing plate. DETAILED DESCRIPTION

[0015] The present invention will be further described below with reference to the accompanying drawings and examples. Example 1

[0016] See also Figure 1The intelligent weighing equipment for multimodal data fusion for shepherding includes a sloped cage-shaped weighing platform 1 (weighing module) and a vital sign monitoring module, an image recognition module, a communication and positioning module, and a low-power energy management module arranged in the sloped cage-shaped weighing platform 1 and communicatively connected to each other; the vital sign monitoring module includes a non-contact infrared temperature sensor 4, a PPG heart rate sensor 5 and a three-axis accelerometer 6; the image recognition module includes a wide-angle camera 7, an infrared fill light 8 and an edge computing unit 9; the communication and positioning module includes a dual-mode communication unit 10 and a dual-mode positioning unit 11; the power consumption energy management module includes a flexible solar panel 12 (the solar panel is tilted 30° to the south, the conversion efficiency is ≥23%, and the average daily power generation is ≥180 watt-hours), a vibration energy recovery unit 13 (using the LTC3588 chip to convert the vibration kinetic energy of the weighing platform into electrical energy, with a conversion efficiency of ≥35% and an average daily power generation of ≥150 mAh) and an intelligent power distribution controller 14.

[0017] The sloped cage-shaped weighing platform 1 includes a cage frame 101 with an open end and a sloped weighing plate 102 arranged at the bottom of the cage frame 101 (inclination angle of 15°, length 1.5 meters, width 1.2 meters, height 1.5 meters; the surface is provided with anti-slip rubber texture, the texture depth is 2mm, the spacing is 10mm, and the inclined section length is 1.5m). The sloped weighing plate 102 has a built-in strain gauge sensor 2 (0-200kg, accuracy ±0.1kg) and a dynamic weighing controller 3 (sampling frequency is 100Hz). The inclination angle of the sloped weighing plate 102 is 15°; the dynamic weighing controller 3 integrates a Kalman filter algorithm. The non-contact infrared temperature sensor 4 and the PPG heart rate sensor Sensor 5 and triaxial accelerometer 6 respectively collect real-time data on sheep's body temperature, heart rate, and activity intensity. The wide-angle camera 7 has a resolution of 1080P at 30fps. The edge computing unit 9 is preloaded with a lightweight YOLOv5 model for real-time analysis of sheep waist-to-hip ratio and backfat thickness. The dual-mode communication unit 10 supports LoRa and NB-IoT protocols (communication range ≥ 10km), and the dual-mode positioning unit 11 supports Beidou and GPS positioning (positioning accuracy ≤ 5 meters). The multimodal data fusion intelligent weighing equipment for sheep herding also includes an alloy electromagnetic shielding layer (a 0.5mm thick μ-metal alloy is used to encase the weighing sensor and wireless communication circuit).

[0018] All the above modules are connected to the data center for communication.

[0019] Software Configuration: Edge Computing Unit Pre-trained Model: YOLOv5s trained on 5,000 sheep images (mAP@0.5 accuracy = 0.93). Cloud Synchronization: Compressed data packets uploaded via LoRa at dawn every day (average traffic ≤ 10 megabytes / day).

[0020] The workflow of the dynamic weighing controller includes: using a three-axis accelerometer to detect the acceleration threshold (>0.5g) when the sheep enters the weighing platform; triggering the Kalman filter algorithm to eliminate motion noise, and the valid data judgment condition is that the weight stabilization time is ≥0.3 seconds; outputting the final weight data with an error of ≤±0.2%.

[0021] In the vital sign monitoring module: the non-contact infrared temperature sensor has a measurement accuracy of ±0.2°C and is installed at a height of 30 cm from the surface of the weighing platform; the PPG heart rate sensor has a sampling frequency of 100 Hz; the three-axis accelerometer model is LIS3DH, with a range of ±16g.

[0022] In the image recognition module: the wide-angle camera model is OV5640, with a field of view of 120°, and is equipped with an infrared fill light with a wavelength of 850nm; the YOLOv5 model is trained based on 5,000 sheep images, with an accuracy of mAP@0.5=0.93.

[0023] The working mode of the dual-mode communication unit is: LoRa is used to transmit real-time warning data with a transmission delay of ≤500ms; compressed data packets are uploaded via NB-IoT every morning with an average daily traffic of ≤10MB.

[0024] The dual-mode positioning unit uses BeiDou-3 and GPS L5 dual-band signal fusion positioning, and the positioning data is updated every 10 seconds.

[0025] The intelligent power distribution strategy of the low-power energy management module includes: when there are no sheep passing by, the device enters deep sleep mode with standby power consumption ≤ 0.1W; when sheep are detected approaching (distance ≤ 2m), each module is awakened by the infrared sensor.

[0026] The communication data signal flow is shown as follows: Sheep enter ↓ Acceleration detection (three-axis accelerometer) → wake up the device ↓ Weighing data (strain gauge sensor) → Kalman filter (dynamic weighing controller) ↓ ↓ Vital sign data (body temperature, heart rate) → Edge computing unit ← Image data (wide-angle camera) ↓ ↓ Data fusion (health score) → real-time warning (LoRa) / cloud synchronization (NB-IoT) ↓ Data Center (Storage and Analysis).

Claims

1. An intelligent weighing device with multimodal data fusion for sheep herding, characterized in that: The multimodal data fusion intelligent weighing equipment for shepherding comprises a sloped cage-shaped weighing platform (1) and a vital sign monitoring module, an image recognition module, a communication and positioning module, and a low-power energy management module arranged in the sloped cage-shaped weighing platform (1) and communicatively connected to each other; the vital sign monitoring module comprises a non-contact infrared temperature sensor (4), a PPG heart rate sensor (5), and a three-axis accelerometer (6); the image recognition module comprises a wide-angle camera (7), an infrared fill light (8), and an edge computing unit (9); the communication and positioning module comprises a dual-mode communication unit (10) and a dual-mode positioning unit (11); and the power consumption energy management module comprises a flexible solar panel (12), a vibration energy recovery unit (13), and an intelligent power distribution controller (14).

2. The multimodal data fusion intelligent weighing device for sheep herding according to claim 1, characterized in that: The sloped cage-shaped weighing platform (1) comprises a cage-shaped frame (101) with an open end and a sloped weighing plate (102) arranged at the bottom of the cage-shaped frame (101); the sloped weighing plate (102) has a built-in strain gauge sensor (2) and a dynamic weighing controller (3).

3. The multimodal data fusion intelligent weighing device for sheep herding according to claim 2, characterized in that: The sloped weighing plate (102) has an inclination angle of 15°; and a Kalman filter algorithm is integrated into the dynamic weighing controller (3).

4. The multimodal data fusion intelligent weighing device for sheep herding according to claim 1, characterized in that: The non-contact infrared temperature sensor (4), PPG heart rate sensor (5) and three-axis accelerometer (6) respectively collect the sheep's body temperature, heart rate and activity intensity data in real time.

5. The multimodal data fusion intelligent weighing device for sheep herding according to claim 1, characterized in that: The wide-angle camera (7) has a resolution of 1080P@30fps; the edge computing unit (9) is preloaded with a lightweight YOLOv5 model for real-time analysis of the waist-to-hip ratio and backfat thickness of sheep.

6. The multimodal data fusion intelligent weighing device for sheep herding according to claim 1, characterized in that: The dual-mode communication unit (10) supports LoRa and NB-IoT protocols, and the dual-mode positioning unit (11) supports Beidou and GPS positioning.

7. The multimodal data fusion intelligent weighing device for sheep herding according to claim 1, characterized in that: The multimodal data fusion intelligent weighing equipment for sheep herding also includes an alloy electromagnetic shielding layer that wraps each component.

Images