Description

Implementing a DeepStream pipeline to detect objects of interest is only one of the many components that compose an AI powered system. Using these predictions to perform business decisions is yet another challenge. RidgeRun is creating a set of Python reference designs to be used as robust, modular and extensible starting points for applications using deep learning in some form. This allows you to quickly build a prototype for by focusing on what makes your demo unique instead of wasting time in infrastructure work.

The Deepstream Reference Design is composed by several components:

RTSP camera capture
Event logging to a web dashboard
Snapshot taking of every event
Comprehensive design and customization documentation
Custom media processing interfaces
Business rules interface for custom processing implementation

One of the reference designs is the Automatic Parking Lot Vehicle Registration, which implements a system capable of identifying the time, date and location of some predefined events. More specifically:

License plate enters the premise
License plate parks on/leaves a parking stall
License plate leaves the premise

Learn more in our developer's wiki.

This product includes 5 hours of engineering support for any task related to integration and testing of the Deepstream Reference Design.

Boards Supported

NVIDIA Jetson Orin
NVIDIA Jetson Orin Nano
NVIDIA Jetson Orin NX
NVIDIA Jetson Xavier
NVIDIA Jetson Xavier NX
NVIDIA Jetson TX2
NVIDIA Jetson Nano

DeepStream Reference Design