Precision Livestock Farming for Behavior Measurement and Welfare Management in Broiler Chickens

Rapid development of broiler production brings up multiple challenges for animal welfare. In recent years, Precision Livestock Farming (PLF) has been widely used for addressing these challenges. PLF aims to improve production efficiency and animal welfare via developing a real-time, continuous and automatic system. The objective of the dissertation is to explore and develop PLF techniques for broiler behavior measurement and welfare management based on welfare enrichment, smart sensors (accelerometer, microphone) and cameras.

The effects of Elevated Platform (EP) and Robotic Vehicle (RV on litter moisture content (LMC), NH₃ concentration, paw quality, plumage cleanliness, and bird activity index (AI) were investigated and described in Chapter II. The results show that the application of EP and RV could significantly improve housing environment and bird welfare.

The performance of using three-dimensional accelerometer and two machine learning models (K-Nearest Neighbor and Support Vector Machine) to classify broiler walking, resting, feeding and drinking behaviors was described in Chapter III. Both models show high sensitivities in identifying broiler resting (> 85%) and walking (99% for both).

The frequency ranges of six common sounds, including bird vocalization, fan, feed system, heater, wing flapping, and dustbathing in a commercial farm were explored and described in Chapter IV. The frequency ranges of bird vocalization continuously decreased as birds grew.

The effects of six image sampling time intervals (0.04, 0.2, 1, 10, 60, and 300 s) on the accuracy of broiler AI at different bird ages, locations and times of day were investigated and described in Chapter V. The results show time interval of 0.04 s yielded the highest broiler AI, followed by 0.2 s with an acceptable accuracy and 80% less computational resources.

The correlation of broiler gait score (GS) with several flock metrics (age, body weight, activity and distribution) were investigated and described in Chapter VI. Broiler GS positively correlated with body weight (R² = 0.97 for Cobb and Ross), while negatively correlated with activity (R² > 0.76 for Cobb and Ross).

In conclusion, PLF technologies for broilers are promising approaches to enhance farm management efficiency, monitor broiler behaviors, and improve bird welfare conditions.