Jetson Nano 配置 Mediapipe(无编译)

Jetson Nano 下载 Mediapipe

增加交换以获得更多交换内存

===============================

git clone https://github.com/JetsonHacksNano/installSwapfile.git  
cd installSwapfile

./installSwapfile.sh

setup - 预安装包括 Python 和 pip

===================

sudo apt update
sudo apt-get update
sudo apt-get install python3-pip
sudo pip3 install -U pip testresources setuptools==49.6.0

TensorFlow 的 python 库

================================

sudo apt-get install libhdf5-serial-dev hdf5-tools libhdf5-dev zlib1g-dev zip libjpeg8-dev liblapack-dev libblas-dev gfortran

下载Python库文件

=================================

sudo pip3 install -U --no-deps numpy==1.19.4 future==0.18.2 mock==3.0.5 keras_preprocessing==1.1.2 keras_applications==1.0.8 gast==0.4.0 protobuf pybind11 cython pkgconfig

# Numpy 的一部分 -> 以二进制格式存储大量数据
sudo env H5PY_SETUP_REQUIRES=0 pip3 install -U h5py==3.1.0

sudo apt-get install python3-opencv

从源代码中构建Mediapipe

=======================================

git clone https://github.com/google/mediapipe.git
cd mediapipe

为Mediapipe设置安装更多库

===============================================

sudo apt-get install -y libopencv-core-dev  libopencv-highgui-dev libopencv-calib3d-dev libopencv-features2d-dev libopencv-imgproc-dev libopencv-video-dev

改动权限

sudo chmod 744 setup_opencv.sh

从源代码运行安装：大约 30 分钟

====================================================

./setup_opencv.sh

最后一步

sudo pip3 install opencv_contrib_python
git clone https://github.com/PINTO0309/mediapipe-bin
cd mediapipe-bin

sudo apt install curl

./v0.8.5/numpy119x/mediapipe-0.8.5_cuda102-cp36-cp36m-linux_aarch64_numpy119x_jetsonnano_L4T32.5.1_download.sh

sudo pip3 install numpy-1.19.4-cp36-none-manylinux2014_aarch64.whl
sudo pip3 install mediapipe-0.8.5_cuda102-cp36-none-linux_aarch64.whl
pip3 install dataclasses 


#download examples from here : https://github.com/feitgemel/BodyPos/tree/master/MediaPipe/Demo

git clone https://github.com/feitgemel/BodyPos.git
cd BodyPos
cd MediaPipe
cd Demo
python3 MediaPipe-Holistic.py