Automatic detection of dead birds on poultry farms using artificial vision

On poultry farms, early detection of dead birds is key to ensuring animal welfare, maintaining biosecurity on the premises and avoiding health risks. Traditionally, this task is performed manually, which is a time-consuming process that depends on human intervention and is difficult to scale up on large farms.

The end customer was looking for an innovative solution that would automate the detection of dead birds, reducing the operational burden, improving process efficiency and moving towards a more digitised and intelligent production model.

NAIR Center carried out the selection, adaptation, optimisation and integration of Artificial Intelligence algorithms, including computer vision and image processing techniques, ensuring their efficient execution in a mobile robotic system, under real-time constraints, computational resources and integration with the rest of the system modules. As a result, the system:

• Analyses images captured in real time in complex and unstructured environments.

• Correctly detects and classifies dead birds as opposed to other elements in the environment.

• Improvement of animal welfare and biosecurity on farms.
• Reduction of the time and operating costs associated with manual tasks.
• Increase in the efficiency and reliability of the detection process.

Phenomenon Robotics

Navarre (Spain)