Instructions to Generate Disparity Maps for Stereo Images

These are instructions for generating a disparity map for a given stereo image pair.

The isaac_ros_ess_visualizer.py script supports:

• rosbag input type

• Raw input images

• Camera info files

To generate a disparity estimation from raw inputs:

1. Complete the Quickstart guide.

2. Launch the ESS Disparity Node using one of the following options:

   To run ESS at a threshold of 0.3 run:

   ros2 launch isaac_ros_ess isaac_ros_ess.launch.py engine_file_path:=${ISAAC_ROS_WS}/isaac_ros_assets/models/dnn_stereo_disparity/dnn_stereo_disparity_v4.1.0_onnx/ess.engine threshold:=0.3
   

   To run ESS at a threshold of 0.0 run:

   ros2 launch isaac_ros_ess isaac_ros_ess.launch.py engine_file_path:=${ISAAC_ROS_WS}/isaac_ros_assets/models/dnn_stereo_disparity/dnn_stereo_disparity_v4.1.0_onnx/ess.engine threshold:=0.0
   

3. Visualize and validate the output of the package:

   ros2 run isaac_ros_ess isaac_ros_ess_visualizer.py --raw_inputs
   

   With threshold set to 0.4, you should see:

   

   With threshold set to 0.0, you should see:

   

4. Try your own examples:

   ros2 run isaac_ros_ess isaac_ros_ess_visualizer.py --raw_inputs \
           --left_image_path '<Absolute path to your left image>' \
           --right_image_path '<Absolute path to your right image>' \
           --camera_info_path '<Absolute path your camera info json file>'