Computer vision

Unifying all data your cameras and sensors are gathering into a single source of truth is difficult due to hardware and software incompatibilities among many manufacturers and platforms.

Viam solves this issue through unique architecture based in gRPC and WebRTC that makes it easy for machines to communicate with each other. Enable any new or existing camera on any machine to leverage any CV model across your entire fleet.

Security outcomes that flexible CV pipelines can enable

Motion detection

Turns on the camera and saves footage only whenever any object or person moves within the field of vision.

Object tracking

Keeps track of where multiple objects in a scene are located.

Object identification

Turns on and saves footage when people and animals are detected by camera.

Facial recognition

Recognizes and verifies whether people are authorized to be in a certain area.

All-in-one platform for adding intelligence to camera hardware

Use with any camera feed

Set up your cameras as sensor components to programmatically access them in any of Viam's open-source SDKs. Viam’s highly modular architecture allows it to support any type of camera feed - use built-in cameras supported by Viam out of the box, or leverage our open marketplace, the Modular Registry, to use or retrofit any camera you want with little to no additional investment needed.

Collect data from the edge

Use data capture and cloud sync services to reliably collect data from your camera regardless of network conditions. Data can be collected and stored locally when connectivity is not available, and machines running Viam can seamlessly update data feeds once a connection is reestablished.

Build and deploy CV applications with ease

Viam empowers developers to seamlessly use software to manage hardware, leverage cloud services, and develop custom applications in their preferred programming language through Viam SDKs.

Customize CV to your needs

Enhance recognition capabilities by adding more pictures of the people you want to accurately identify, view detailed analytics on identified people and objects over periods of time, and personalize when certain people or objects should trigger alerts depending on the circumstances when they were recorded on camera.

Quickly test and deploy updated CV models to all or part of your fleet

Viam’s fleet management system allows for a computer vision model time-to-market measured in minutes, and iteration times measured in seconds, on either a single location or for multiple cameras, locations, and organizations. Instant notifications are also available through webhooks for incidents and anomalies, and include image capture.

CV offloading

Lighten the load for your smart machine by sending data-intensive object recognition workloads for processing to the cloud. Running CV in the cloud saves money, resources, and battery power without compromising detection precision.