About
Designed and implemented advanced computer vision systems for the forestry industry to detect unsafe operator access to harvesters and forwarders, and to accurately estimate trailer load levels (as a percentage) of cut logs, improving both safety compliance and operational efficiency.
Challenge, approach, and impact
Enhancing Safety Through Video-Based Behavior Monitoring
Ensuring operator safety during machinery access is a critical challenge in forestry operations. This project addresses the difficulty of identifying unsafe mounting and dismounting practices by analyzing existing onboard video footage, enabling data-driven safety improvements and reducing accident risk.
Monitoring Load Limits and Machine Activity Through Video Analytics
Managing trailers' load limits and understanding machine usage are key operational challenges. This project addresses the need to prevent trailer overloading, estimate log load levels, break down machine activity into specific phases, and count grapple actions—using video-based analysis to improve safety, efficiency, and decision-making in the field.
How we built
Testimonials
Joaquín Campo, Machine Learning Engineer @ Pento.ai
Pento.ai
“Working on this project was both technically challenging and deeply rewarding. It was magnificent to be part of such a large operation, and the collaboration with the client was fantastic. The structure of the project matched really well with our way of working, which made it easy to make progress, showcase results, and iterate effectively. Props to the whole team!“
Team structure
Client team
Sofía C
Specialist, Own Harvesting Production Support
Daily point of contact
Federico G
Manager, Data and Analytics
Project stakeholder
The client stakeholders were working closely with the team at Pento.ai
Agency team
2 x Machine Learning Engineer
Production
1 x Governance
Governance
1 x Tech Lead
Production
