Computer vision with neural networks: cameras that see and understand in real time

Computer vision lets a computer understand images and video the way a person would, but without getting tired and in real time. With it, cameras that only record today can count, recognize and send alerts.

How does a convolutional neural network (CNN) work?

The image enters the network and passes through convolution and pooling layers that detect edges, shapes and increasingly complex objects. At the end, the network classifies what it sees and assigns a confidence level, for example «person 0.97» or «vehicle 0.95».

Applications for businesses and cities

• People counting: occupancy in stores, events and public spaces.
• Vehicle control: access to parking lots and residential complexes.
• License plate recognition: automatic logging of entries and exits.
• Vessel monitoring: docks, marinas and coastal areas.

Computer vision in Cartagena

In a city with a bay, docks, tourism and heavy traffic, computer vision has very concrete uses. At GETS we combine these models with our experience in access control and electronic security so the cameras you already have work better.

Frequently asked questions

Do I need to replace my cameras?

In many cases, no. We assess your current cameras and tell you whether they work or what's missing.

Does it work at night?

It depends on the camera and the lighting. During the assessment we review the conditions at each point.

What about privacy?

We design every solution in compliance with Law 1581 of 2012, Colombia's personal data protection law.