本次安装版本为：Tensorflow == 2.2.0， numpy == 1.21.5，OpenCV == 4.5.5，其版本兼容
建议通读全文，然后根据需要进行安装。
[En]
It is recommended to read through the full text and then install it as needed.
（本次安装的路径均为：/home/pi/.local/lib/python3.7/site-packages）
（Tensorflow == 2.2.0，Numpy == 1.21.5，OpenCV == 4.5.5的树莓派.img镜像
链接：https://pan.baidu.com/s/17sbbeK59d92LFEA7v2-vDQ
提取码：abrf
无线账号3-124
密码sanjiao124
树莓派账号pi
密名123456
如果系统中没有重要的环境，建议备份重要的文件，然后格式化和刻录我的图像。你不必经历这个过程[大脑疼痛]。
[En]
If there is no important environment in the system, it is recommended to back up the important files, then format and burn my image. You don’t have to go through this process [brain pain].
更改热点(无线网的账号密码)：https://blog.csdn.net/lizhen222222/article/details/122935984（未验证））
先更新和升级软件包：
sudo apt-get update&&sudo apt-get upgrade
接下来正式安装
一，Tensorflow2
1.这些玩意都得安
1.sudo apt-get install -y libhdf5-dev libc-ares-dev libeigen3-dev gcc gfortran python-dev libgfortran5
2.sudo apt-get install -y libatlas3-base libatlas-base-dev libopenblas-dev libopenblas-base libblas-dev
3.sudo apt-get install -y liblapack-dev cython libatlas-base-dev openmpi-bin libopenmpi-dev python3-dev
4.sudo pip3 install keras_applications==1.0.8 --no-deps
5.sudo pip3 install keras_preprocessing==1.1.0 --no-deps
6.sudo pip3 install h5py==2.9.0
7.sudo pip3 install pybind11
8.pip3 install -U --user six wheel mock
9.wget "https://raw.githubusercontent.com/PINTO0309/Tensorflow-bin/master/tensorflow-2.2.0-cp37-cp37m-linux_armv7l_download.sh"
10../tensorflow-2.2.0-cp37-cp37m-linux_armv7l_download.sh
11.sudo pip3 uninstall tensorflow
12.sudo -H pip3 install tensorflow-2.2.0-cp37-cp37m-linux_armv7l.whl
如果”9.”下载失败了，去下载安装包，注意下载的包要与树莓派对应
下载地址(传送门)：github：(https://github.com/lhelontra/tensorflow-on-arm/releases)
（如果下载慢，可用我的另一篇教程加速github加速）
1）查看树莓派位数
getconf LONG_BIT
2）查看树莓派版本
uname -a
所以我下载的是
下载好之后拉到树莓派/home/pi/.local/lib/python3.7/site-packages目录下，解压，
终端cd /home/pi/.local/lib/python3.7/site-packages，执行
sudo pip3 install tensorflow-2.2.0-cp37-none-linux_armv7l.whl
（版本号对应修改，即sudo pip3 install+文件名）
2.验证是否安装成功
python3
>>>import tensorflow as tf
>>>tf.__version__
▲PS：查询tensorflow安装路径为:
tf.__path__
二，numpy
pip3 install --target=/home/pi/.local/lib/python3.7/site-packages numpy==1.19.5 -i https://pypi.douban.com/simple --trust -host=pypi.douban.com
2.验证是否安装成功
python3
>>>import numpy
>>>numpy.__version__
或者
pip3 show numpy
三，OpenCV
先扩展容量，然后在安装后重新启动
[En]
First expand the capacity, then restart, after the installation
sudo raspi-config
Advanced Options—-Expand Filesysterm
1.安装OpenCV相关依赖包
（官方安装步骤：Install OpenCV 4 on your Raspberry Pi（没必要按着官方的做，按接下来的步骤完全可以安装成功））
（安装包下载地址Releases–链接没问题，但同样没必要下载之后再安装）
1）更新和升级软件包：
sudo apt-get update&&sudo apt-get upgrade
2）安装图像I/O包：
sudo apt-get install libjpeg-dev libtiff5-dev libjasper-dev libpng-dev
3）安装视频I/O包，方便获取 视频流：
sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libv4l-dev
sudo apt-get install libxvidcore-dev libx264-dev
4）为编译 highgui模块，我们需要安装GTK开发库：
sudo apt-get install libfontconfig1-dev libcairo2-dev
sudo apt-get install libgdk-pixbuf2.0-dev libpango1.0-dev
sudo apt-get install libgtk2.0-dev libgtk-3-dev
5）安装一些额外项来优化opencv内部操作：
sudo apt-get install libatlas-base-dev gfortran
6）安装适用于HDF5数据集和Qt GUI：
sudo apt-get install libhdf5-dev libhdf5-serial-dev libhdf5-103
sudo apt-get install libqtgui4 libqtwebkit4 libqt4-test python3-pyqt5
▲PS：豆瓣安装
pip3 install opencv-python -i https://pypi.douban.com/simple --trust -host=pypi.douban.com
指定安装路径
pip3 install --target=/home/pi/.local/lib/python3.7/site-packages opencv-python
7）所以指定路径下豆瓣安装：
pip3 install --target=/home/pi/.local/lib/python3.7/site-packages opencv-python -i https://pypi.douban.com/simple --trust -host=pypi.douban.com
2.验证是否安装成功
python3
>>>import cv2
>>>cv2.__version__
第四，这之后是安装过程中遇到问题时重新安装的步骤。我希望你不会用它。
[En]
Fourth, after this is the step of reinstalling when you encounter problems in the installation process. I hope you won’t use it.
1）a.卸载numpy
sudo pip3 uninstall --target=/home/pi/.local/lib/python3.7/site-packages numpy==1.19.5
b.(这条好用)
pip3 install -U numpy
2）
删除文件夹
sudo rm -rf /
3）用cpu （predict.py）
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"
▲PS.换源
sudo nano /etc/apt/sources.list
将原有内容用#注释掉
阿里源
deb http://mirrors.aliyun.com/raspbian/raspbian/ buster main contrib non-free rpi
deb-src http://mirrors.aliyun.com/raspbian/raspbian/ buster main contrib non-free rpi
清华源
deb http://mirrors.tuna.tsinghua.edu.cn/raspbian/raspbian/ buster main non-free contrib rpi
deb-src http://mirrors.tuna.tsinghua.edu.cn/raspbian/raspbian/ buster main non-free contrib rpi
sudo nano /etc/apt/sources.list.d/raspi.list
将原有内容用#注释掉
阿里源
deb http://mirrors.aliyun.com/raspbian/raspbian/ buster main
deb-src http://mirrors.aliyun.com/raspbian/raspbian/ buster main
清华源
deb http://mirrors.tuna.tsinghua.edu.cn/raspberrypi/ buster main ui
Ctrl+X，然后会提示输入Y保存修改、输入N放弃修改
3.更新源、软件
sudo apt-get update&&sudo apt-get upgrade
4.清理垃圾
df -hl 命令可查询当前硬盘使用情况
sudo apt-get autoclean –清理旧版本的软件缓存
sudo apt-get clean–清理所有软件缓存
sudo apt-get autoremove–删除系统不再使用的孤立软件
▲ubuntu安装opencv
sudo apt install cmake
pip install scikit-build
sudo apt-get install build-essential libgtk2.0-dev libavcodec-dev libavformat-dev libjpeg-dev libswscale-dev libtiff5-dev
sudo apt-get install libgtk2.0-dev
sudo apt-get install pkg-config
安装
pip3 install opencv-python -i https://pypi.douban.com/simple --trust -host=pypi.douban.com
Original: https://blog.csdn.net/DDonggggg/article/details/122222424
Author: DD᭄ꦿng
Title: 树莓派安装Tensorflow2、OpenCV、numpy
原创文章受到原创版权保护。转载请注明出处：https://www.johngo689.com/509222/
转载文章受原作者版权保护。转载请注明原作者出处！