Performance of an image analysis processing system for hen tracking in an environmental preference chamber

Area of Science:

• Animal Science
• Agricultural Engineering
• Computer Vision

Background:

• Image processing is crucial for livestock monitoring, including identification, tracking, and behavior analysis.
• Automated systems offer efficiency gains over manual observation in animal welfare studies.

Purpose of the Study:

• To quantify the performance of an image processing system for monitoring laying hen navigation and occupancy in an environmental preference chamber.
• To compare automated detection rates with human observations for validating system accuracy.

Main Methods:

• An image processing system utilizing an ellipse-fitting model was implemented in a four-compartment environmental chamber.
• Top-view images from cameras in each compartment were analyzed to detect hen presence.
• System performance was evaluated by comparing detected compartment occupancy duration with human observations.

Main Results:

• The image processing system achieved a high success detection rate of 95.9 ± 2.6% for total compartment occupancy.
• The system demonstrated suitability for assessing environmental choices and significantly reduced data processing time.
• A preliminary study using the system indicated a negative trend between ammonia levels and hen compartment occupancy.

Conclusions:

• The developed image processing system is accurate and efficient for monitoring laying hen behavior and environmental preferences.
• This automated approach provides a reliable alternative to manual video analysis, saving considerable time.
• Further research is warranted to explore the system's application in detailed environmental aversion studies, such as ammonia exposure.