The perception on the edge is just one part of streaming analytics IoT application. The other side is analytics, where you take the metadata generated by the perception pipeline to create actionable insights. In this complete end-to-end solution, we are showing a smart parking garage application where DeepStream is used to build the perception graph on the edge and Apache Spark is used to build the analytics engine. The entire application with perception and analytics pipeline is open sourced and available on GitHub and NGC.
This example is designed for a particular parking garage with a very specific geo-coordinate information and should only serve as a reference for system architects and developers who are building and deploying a complete end-to-end solution.
GitHub: https://github.com/NVIDIA-AI-IOT/deepstream_360_d_smart_parking_application
Analytics server: https://github.com/NVIDIA-AI-]IOT/deepstream_360_d_smart_parking_application/tree/master/analytics_server_docker
NGC: https://ngc.nvidia.com/catalog/containers/nvidia:deepstream_360d