Difference between revisions of "TB3 ROS OpenCL Caffe"

Revision as of 14:17, 17 July 2018

Reference

• ROS OpenCL Caffe
• YOLO2

Prerequisite

• Done: Ubuntu 16.04
• Done: ROS kinetic
• Done: RGB Camera

Environment Setup

• ROS Kinetic - See TR3_Intel_ROS_Project#Software_Environment
• Catkin Workspace - See TR3_Intel_ROS_Project#Software_Environment
• clCaffe TB3 clCaffe - Maybe Done - This step is misleadingly involved. It is also finicky. It took me several trials to get it right. Once done, it doesn't appear that difficult. It is easy to miss a detail that puts you on a bad path. Furthermore, there is a lot of glue here to stitch together components from diverse environments, for example, ROS catkin, classic Ubuntu, rpm, Debian, and OpenCL. Furthermore, its not clear how many of the support items are need. I found I need some of "Inference Optimization" and "YOLO2 Model Support".

• Create a symbol link in /opt/clCaffe

 $ sudo ln -s /home/eepp/code/clCaffe/ /opt/clCaffe

• Add clCaffe libraries to LD_LIBRARY_PATH to the .bashrc file.

 $ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/opt/clCaffe/build/lib

• Install ROS package for different cameras as needed. See TR3_Intel_ROS_Project#Environment_Setup

Building and Installing

• object_msgs has already been installed. See TR3_Intel_ROS_Project#Package:_Object_Msgs_.281.29

 $ cd ~/catkin_ws/src/

 $ git clone https://github.com/intel/ros_opencl_caffe

 $ cd ~/catkin_ws/

 $ catkin_make
     ...
     [ 98%] Linking CXX executable /home/eepp/catkin_ws/devel/lib/opencl_caffe/unittest_detector
     [ 98%] Built target unittest_detector
     [100%] Linking CXX executable /home/eepp/catkin_ws/devel/lib/movidius_ncs_stream/unittest_stream_detection
     [100%] Built target unittest_stream_detection

 $ catkin_make install

 $ source install/setup.bash

 $ cp ~/catkin_ws/src/ros_opencl_caffe/opencl_caffe/resources/voc.txt /opt/clCaffe/data/yolo/

Running Demo

Demo 1

Passed

 $ roslaunch opencl_caffe_launch usb_cam_viewer.launch

Holly Shit it did something interesting.

This didn't work until I added "Inference Optomization" and "Yolo2 Model Support" items in TB3_clCaffe.

Demo 2

Failed

 $ roslaunch opencl_caffe_launch realsense_viewer.launch

     File "/usr/lib/python2.7/dist-packages/rospkg/rospack.py", line 203, in get_path
       raise ResourceNotFound(name, ros_paths=self._ros_paths)
       ResourceNotFound: realsense_ros_camera

Error: realsense_ros_camera not found. It should be in realsense_samples_ros. See similar issue in TB3_ROS_Moving_Object#Launch_1

Test

TBD

Investigation

 $ cd /home/eepp/ros_packages/roslaunch_to_dot

 $ python roslaunch-to-dot.py --png ~/catkin_ws/install/share/opencl_caffe_launch/launch/opencl_caffe_nodelet.launch caffe_nodelet.dot

 $ gimp caffe_nodelet.png

Also did "usb_cam_viewer.launch"