tutorial_openvino.md

Tutorial Enable OpenVINO™ (Open Visual Inference & Neural Network Optimization)

In this tutorial, we introduce how to enable OpenVINO™ option for grasp detection. Based on convolutional neural networks (CNN), the toolkit extends workloads across Intel® hardware and maximizes performance.

1. Install OpenVINO™ toolkit

OpenVINO supports multiple host OS. We verified with 64-bit Ubuntu 16.04.

For gpd execution at Intel CPU, please follow the Installing the OpenVINO™ Toolkit for Linux.

For gpd execution at Intel GPU, please follow additional steps to install OpenCL NEO driver Additional Installation Steps for Processor Graphics (GPU).

For gpd execution at Intel Movidius NCS, please follow additional steps to configure UDEV rules. Additional Installation Steps for the Intel® Movidius™ Neural Compute Stick.

2. Verify OpenVINO™ installation

Try to run the demo scripts: Verify the Demo Scripts

Then try to run OpenVINO inference example applications: Build the Samples, Running the Samples

3. Build GPD with OpenVINO

Setup OpenVINO environment variables, replacing <OPENVINO_INSTALL_DIR> with the specific location.

source <OPENVINO_INSTALL_DIR>/bin/setupvars.sh

Once OpenVINO installed, build GPD with option "USE_OPENVINO" ([OFF]|ON)

catkin_make -DCMAKE_BUILD_TYPE=Release -DUSE_OPENVINO=ON --pkg gpd

4. Launch GPD with OpenVINO

The launch process is similar to Detect Grasps With an RGBD camera, just with an additional param "device" ([0:CPU]|1:GPU|2:VPU|3:FPGA) to specify the target device to execute the grasp detection.

# launch the openni camera for pointcloud2
roslaunch openni2_launch openni2.launch
# start rviz
rosrun rviz rviz
# setup OpenVINO environment variables, replacing <OPENVINO_INSTALL_DIR> with the specific location
source <OPENVINO_INSTALL_DIR>/bin/setupvars.sh
# launch the grasp detection
roslaunch gpd tutorial1.launch device:=0