Detection of Aggressive Behaviour of Cattle Operators

Client

Aggressive handling of cattle can lead to problems like stressed or frightened cattle, injuries, and poor overall welfare of the animals. This can result in reduced quality of their end products, decreased productivity, safety risks, and violations of animal welfare regulations.

InData Labs has developed an AI solution that uses computer vision technologies like object detection and pose estimation for automated aggression detection of cattle operators directed against cows. This solution has been developed for a US-based client from the cattle industry to enable a safe environment for cattle keeping them healthy and providing sustainable livestock production for farmers.

Challenge: identify key gestures and behavior of operators around cows

To enable aggression detection, our team of data scientists had to identify the object classes:

• Person
• Cows
• Milking booth
• Equipment used
• Parts of cow: head tail legs udder.

Solution: AI video analytics to track the occurrence of events of aggression

Today, AI is employed in animal health and welfare observation. Object detection and pose estimation technologies allow automated aggression detection cases:

• punches with fists, sticks, kicks.
• pushes and pulls
• hand spanking, spanking with a towel
• aggression with water directed at the cow’s head, etc.

To train the model, the InData Labs’ team used the pose-based approach with person detection and tracking algorithms. Identifying a person’s key points and extracting important features based on 60-70 frames per person the AI-based model estimates whether the behaviour is aggressive or not.

There are several ways to solve this challenge, though we went for pose estimation because it was a good fit in terms of tech capabilities. The scheme demonstrates our approach to pose estimation:

Result: AI solution to preserve the health and welfare of cows

Preserving the health and welfare of cows is a crucial responsibility for individuals, communities, and organizations. The upkeep and management of cattle can have a significant impact on their development and procreation, that’s why it’s essential to improve care using both traditional methods like sanitary dwelling, healthy nutrition, vaccination, and illness prevention, and innovative ones like implementing AI for a safe environment and good welfare of the animals.

InData Labs and the Client have collaborated on an AI-driven solution for cattle operators’ aggression recognition. The solution enables prompt detection of hostile behavior directed against cows with the help of technology.

Watch the video to learn how the technology works: