PPE and procedure checks with computer vision.
Deep-learning video analysis to support safety checks in electrical substations: protective equipment and observable steps in operating procedures.
Make situations needing review visible.
Work in electrical substations requires attention to personal protective equipment and safety procedures. Relying solely on manual observation makes it difficult to apply consistent criteria across activities and sites. The project introduces visual support to identify situations that need review.
- Computer vision
- Deep learning
- Operational safety
What we designed.
The project requires specialised visual models, knowledge of field operations and integration with monitoring systems. Its value lies in making checks more systematic while keeping verification and decisions with safety professionals.
Recognise protective equipment
Deep-learning computer vision models analyse video streams to recognise equipment such as helmets, gloves and safety footwear. Real-time processing supports checks during field operations.
Connect images and procedures
The system relates recognised elements to the checks required by the operational process, supporting review of the steps visible in the footage. The scope must distinguish what video can show from conditions requiring dedicated technical checks.
Integrate checks into monitoring
Analysis results are integrated with company monitoring systems. Operators and managers gain an additional reference for reviewing detected situations and handling them through the required procedures.
Video streams and checking criteria
Deep-learning visual recognition
Items for review in monitoring systems
What it makespossible.
- Real-time recognition of PPE visible in the images.
- Support for more consistent checks on observable activities.
- Information integrated into monitoring to guide field verification.
Want to apply computer vision to checks in your operational processes?
Let’s start with your context, available data and the first useful outcome.
Tell us about your project
Document data extraction and reconciliation.
Neural networks and NLP to extract information from varied delivery notes, transport documents and invoices, then reconcile it across automotive accounting and management systems.