79 / 2025-04-15 22:34:15

Robotic monitoring of heat stress in caged hens: thermal panting-based warning modeling for temperature-humidity-wind environments

Laying hens,inspection robot,THI,environment monitoring,vision system

3. 智慧牧场与动物福利 > （5）智能设备和工具：环境控制、饲喂、供水、挤奶，及行为监测

As global warming leads to more frequent heatwaves, heat stress (HS) has become a pivotal constraint on poultry welfare and productivity. While the temperature and humidity index (THI) is commonly used to evaluate thermal conditions, it does not account for the varying levels of HS experienced by individual poultry in intensive farming settings. It is still a challenge to intelligently monitor the HS of caged hens. This study introduces an innovative early-warning method using YOLOv8, deployed on an automatic inspection robot system, to monitor thermal panting behavior in caged hens. The system also monitors air temperature (T), relative humidity (RH), and wind velocity (V) to assess the environmental conditions when panting behavior is triggered. Data collected from a 50,000-hen (Jing Ting 6) stacked-cage layer house in a subtropical climate between June and August 2023 demonstrated that panting behavior was strongly influenced by T, RH, V, and the position of hens inside the layer house (P < 0.01). Air temperature showed the strongest association (r = 0.767, P < 0.01), followed by the position of hens inside the layer house (r = 0.582, P < 0.01). During the monitoring period, air temperature within the cage ranged from 22.2 °C to 34.3 °C, with critical thermal thresholds for panting escalation occurring at 25.5 °C (front), 27.1 °C (middle) and 27.4 °C (rear). In addition, wind velocity within the cage ranged from 0.1 m/s to 3.7 m/s during the test period, was found to alleviate heat stress, though its effect diminished as air temperature increased. Based upon the practical data of three consecutive months, a simple linear regression model was developed to predict panting scores from the air temperature, relative humidity, and wind velocity (R² = 0.647). This non-invasive, hen-level monitoring approach enables early HS warnings and provides decision support for precision environmental control, advancing intelligent agricultural practices in poultry farming.