Automated Broiler Phenotyping project presented at the Benelux region meeting of the International Society for Applied Ethology

The PPP Automated Broiler Phenotyping is a collaboration between Cobb Europe (Cobb-Vantress), Dorset Identification and FarmResult, and started in 2022. The project focuses on identifying, upscaling and implementing technologies for automated individual broiler phenotyping, to improve the currently recorded and novel phenotypic traits related to health, welfare and performance. In this project, there is extensive collaboration with broiler breeding practice and technological partners, linking the research to full practical implementation. Some of the ongoing work in this project was presented at the Benelux region meeting of the ISAE.

Pose estimation to characterize walking in broilers

István Fodor presented his work on automated pose estimation to characterize walking in broilers in relation to lameness. Using video images recorded from behind while broilers were walking through a corridor, 11 pose features, including limb angles and step height, were analysed. In addition, each bird’s gait was scored manually. In his presentation, István showed that limb angles were sharper and the maximum leg lift during steps was lower for lame broilers than for broilers without impaired gait. This observation can aid in the development of more sophisticated gait prediction models based on video data.

Automated activity tracking in broilers

Malou van der Sluis presented her work on using sensor technologies, such as radio frequency identification (RFID), to automatically track activity levels in broilers at the individual level. She discussed how the tracking works, as well as what insights were gained from the resulting data. For example, she showed that broilers with an impaired gait walk shorter distances in a day than birds with a good gait. Furthermore, she presented different options for upscaling the RFID recordings to larger groups of broilers, using different antenna configurations. Overall, there appears to be potential for lower-level-of-detail activity tracking in broilers, potentially in combination with computer vision.

Stay tuned for regular updates about our work!