Yolov3, Openpose, Tensorflow2, ROS, multi-thread
It also support for remote GPU server. You can grap a frame from D435 in local and then send the data to remote server to process the data and return the result.
This is my final year project "3D Action Recognition based on Openpose and YOLO".
Following the openpose homepage instruction to install openpose and compile the python api.
conda create -n tensorflow2 python=3.6
pip install -r requirements.txt
At first, following the ros wiki instruction to install ROS and create a ROS workspace
Then, create a ROS package which names act_recognizer.
cd catkin_ws/src
mkdir -p act_recognizer
Therefore, in your ROS workspace file folder, the structure will be as the following,
->~/catkin_ws
----->build
----->devel
----->src
--------->act_recognizer
Cloen the repo and copy all files to ROS package act_recognizer
Modify the act_talker.py. Likes,
sys.path.append('{$your root path}/catkin_ws/src/act_recognizer')
Modify the config.py. Change the path about YOLO data, such as YOLO.CLASSES and YOLO.ANCHORS.
Change your openpose python api path in Module/poser.py, so that your code can import pyopenpose correctly.
Additionally, you also have to change the openpose model path.
Module/poser.py
----->class PoseLoader()
--------->self._params["model_folder"] = your openpose model folder
Download checkpoints from BaiduYun the extract code is cxj6. Then move yolov3.weights into checkpoints folder checkpoints/YOLO and mlp.h5 to checkpoints. You should create a checkpoints folder first probably.
cd act_recognizer/src
mkdir -p checkpoints/YOLO
The ROS package whic written by Python do not need to compile.
We have to specify the python interpreter in all .py files in the first line to use the conda env tensorflow2's python interpreter.
Likes this,
#!/home/dongjai/anaconda3/envs/tensorflow2/bin/python
Then, run the code following,
roscore
cd catkin_ws
conda activate tensorflow2
source ./devel/setup.bash
roslaunch act_recognizer run.launch