Software Application Overview: AI Object Detection Inference for Video

Object detection has many different uses in Command and Control (C2) systems such as autonomous control, target tracking, threat detection, and general surveillance. Graphics Processing Units (GPUs) are the de-facto standard hardware for these types of workloads in datacenter environments. Still, when deploying to an edge environment many considerations are required to ensure an optimized deployment. This paper provides a general overview of how to utilize GPUs for AI inference for object detection at the edge using NVIDIA^® HoloScan as well as an overview of the many considerations to account for when selecting the most optimal GPU for any specific ground vehicle solution.