Animal behavior monitoring: Datasets & Models
Computer vision workflow that detects and tracks livestock in barns, feedlots, and pastures to flag behaviors like feeding, lying, and standing. Runs on standard farm camera footage to surface behavioral changes automatically.
Animal behavior monitoring
This workflow performs animal behavior detection on standard farm camera footage, locating individual livestock in each frame and classifying their posture and activity. It flags behaviors such as feeding, lying, standing, and moving across barns, feedlots, and pastures, producing per animal detections that downstream logic can use to track patterns and surface behavioral changes.
Try it out
Powered by general detection model
Run on custom image
Drop an image here or click to upload
Why automate animal behavior monitoring
Manually watching herds for signs of illness, heat stress, lameness, or estrus is labor-intensive and easy to miss, especially overnight or across large operations. A trained detection model watches every animal continuously, catching subtle shifts in activity that often precede disease or welfare problems, so intervention happens earlier. For agriculture teams, this means fewer missed cases, less reliance on scarce skilled labor, and objective behavioral data instead of periodic spot checks. The result is faster decisions on treatment, feeding, and breeding.
Key benefits
Behavior monitoring works with the cameras farms already have, so deployment does not require new hardware in most barns. Detection scales from a single pen to thousands of animals across multiple sites without adding staff, and the RF-DETR architecture delivers strong accuracy on crowded scenes with occlusion. Outputs integrate cleanly into herd-management software and alerting pipelines for real-time response.
How to use it
- Upload an image: Drop any photo into the widget above to try animal behavior monitoring for free, right on this page.
- Preview instant results: The model detects cow, pig, chicken, sheep, feed trough within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
- Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train RF-DETR on your own images for production accuracy.
- Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.
Recommended model for animal behavior monitoring
RF-DETR is state of the art for real-time object detection, instance segmentation, and keypoint detection, with higher accuracy and lower latency than comparable models, and it's Apache 2.0 licensed. When you're ready to build a production model for animal behavior monitoring, fine-tune it on your own labeled images, then deploy to a Serverless Cloud API endpoint.
Not sure where to start?
Describe your Animal Behavior Monitoring project to Roboflow Agent.