Tutorial for DNN Object Detection with Isaac Sim

Overview

This tutorial walks you through a pipeline for object(people) detection using DetectNet consuming images from Isaac Sim

Tutorial Walkthrough

(1)Complete the quickstart here
=> Already Done
(2)Launch the Docker container using the run_dev.sh script:

cd ${ISAAC_ROS_WS}/src/isaac_ros_common && \
  ./scripts/run_dev.sh

(3)Run the setup script to download the PeopleNet Model from NVIDIA GPU Cloud(NGC) and convert it to a .etlt file

cd /workspaces/isaac_ros-dev/src/isaac_ros_object_detection/isaac_ros_detectnet && \
  ./scripts/setup_model.sh --height 720 --width 1280 --config-file resources/isaac_sim_config.pbtxt

(4)Launch the pre-composed pipeline launch file:

cd /workspaces/isaac_ros-dev && \
  ros2 launch isaac_ros_detectnet isaac_ros_detectnet_isaac_sim.launch.py

(5)Install and launch Isaac Sim following the steps in the Isaac ROS Isaac Sim Setup Guide

(6)Press Play to start publishing data from the Isaac Sim
=> Memory Error
(7)You should see the image from Isaac Sim with the rectangles overlaid over detected people in the frame

Reference:
https://nvidia-isaac-ros.github.io/concepts/object_detection/detectnet/tutorial_isaac_sim.html
https://nvidia-isaac-ros.github.io/concepts/object_detection/detectnet/tutorial_custom_model.html