版权声明:本文为博主原创文章,未经博主允许不得转载。 https://blog.csdn.net/qq_28637193/article/details/80835873
创建热点 开源项目https://github.com/oblique/create_ap
将上面代码克隆到本地在终端执行命令: git clone https://github.com/oblique/create_ap
执行:cd create_ap
:sudo make install
安装依赖库在终端执行:sudo apt-get install util-linux procps hostapd iproute2 iw haveged dnsmasq
创建wifi热点(github项目上有多种安装方式可以自己查找安装方式)
开机启动 终端执行: sudo vim /etc/rc.local
打开文件后,在开头位置加入 sudo create_ap wlan0 eth0 热点名,密码
超频:arm_freq=1400
:over_voltage=4
远程桌面:sudo apt-get install tightvncserver
:sudo apt-get install xrdp
:sudo /etc/init.d/xrdp start
:sudo update-rc.d xrdp defaults
树莓派换源 :sudo vim /etc/apt/sources.list
注释掉原来的,加入中科大源 deb http://mirrors.ustc.edu.cn/raspbian/raspbian/ jessie main contrib non-free rpi
: sudo vim /etc/apt/sources.list.d/raspi.list
注释掉原来的
deb http://mirrors.ustc.edu.cn/archive.raspberrypi.org/debian/ jessie main ui
下载opencv3.3.1版本源码: wget https://github.com/opencv/opencv/archive/3.3.1.tar.gz -O opencv3.3.1.tar.gz
下载opencv_contrib3.3.1版本源码: wget https://github.com/opencv/opencv_contrib/archive/3.3.1.tar.gz -O opencv_contrib.tar.gz
备注: 下载opencv和 opencv_contrib 版本号一致, wget 后-C 是断点续传, -O是更改名称
将opencv-3.3.1 和openv_contrib-3.3.1 文件解压
:tar xvzf opencv3.3.1.tar.gz
:tar xvzf opencv_contrib.tar.gz
编译之前更新所有软件包:sudo apt update & sudo apt -y upgrade
安装编译工具:sudo apt-get install build-essential cmake pkg-config
安装各种图像格式包:sudo apt-get install libjpeg-dev libtiff5-dev libjasper-dev libpng12-dev
安装视频格式所需要的包:sudo apt-get install libavcodec-dev libavformat-dev libswcale-dev libv4l-dev
:sudo apt-get install libxvidcore-dev libx264-dev
安装gtk 库 : sudo apt-get install libgtk3.0
进一步优化:sudo apt-get install libatlas-base-dev gfortran
进入opencv目录:mkdir release && cd release
: sudo cmake -D CMAKE_BUILD_TYPE=RELEASE \
-D CMAKE_INSTALL_PREFIX=/usr/local \
-D OPENCV_EXTRA_MODULES_PATH=/home/pi/Desktop/opencv/opencv_contrib-3.3.1/modules \
-D INSTALL_PYTHON_EXAMPLES=ON \
-D BUILD_EXAMPLES=ON ..
OPENCV_EXTRA_MODULES_PATH后面是自己opencv_contrib文件夹中modules文件路径
网树莓派多核编译会导致报错,使用单核编译,编译时间比较长的四五个小时
编译:sudo make
安装:sudo make install
链接动态库:sudo ldconfig
备注:libswcale-dev 可能安装不上去,可以不安装,不影响编译
下载protobuf : wget https://github.com/google/protobuf/archive/v3.4.1.tar.gz -O protobuf3.4.1.tar.gz
解压 : sudo tar xvzf protobuf3.4.1.tar.gz
编译时保证链接网络编译时会下载东西,这个可以多核编译
安装必要工具:sudo apt-get install autoconf
:sudo apt-get install automake
:sudo apt-get install libtool
编译: sudo ./autogen.sh
: sudo ./configure
: sudo make
:sudo make check
:sudo make install
wiring pi下载: git clone git://git.drogon.net/wiringPi
安装 : cd wiringPi
:./bulid
检查安装是否成功 : gpio -v
tensorflow安装:
下载:wget https://github.com/lhelontra/tensorflow-on-arm/releases/download/v1.4.1/tensorflow-1.4.1-cp27-none-linux_armv7l.whl
安装:sudo pip install tensorflow-1.4.1-cp27-none-linux_armv7l.whl -i https://pypi.tuna.tsinghua.edu.cn/simple
tensorflow models 下载 :git clone https://github.com/tensorflow/models.git
配置视觉使用的环境;https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md
安装:sudo apt-get install python-matplotlib
: sudo pip install pillow
PC机安装lxml方式: sudo pip install lxml
树莓派lxml方式: sudo apt-get install python3-lxml
# From tensorflow/models/research
:protoc object_detection/protos/*.proto --python_out=.
:export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim
测试安装是否成功,进入models/tutorials/image/imagenet 执行:python classify_image.py
在终端执行:sudo raspi-config
在设置中打开摄像头
:sudo apt-get install libv4l-dev
进入etc文件增加摄像头设备号
: sudo vim /etc/modules
在文件最后加入:bcm2835-v4l2 (是小写字母l,不是1)
重启树莓派
下载 mjpg-streamer: git clone https://github.com/jacksonliam/mjpg-streamer
安装支持库:sudo apt-get install libjpeg8-dev
进入mjpg-streaer文件夹
:cd mjpg-streamer-experimental/
: sudo vim plugins/input_raspicam/input_raspicam.c
设置帧数和分辨率 ,帧数30fps比较合适
编译:make clean all
:./mjpg_streamer -i "./input_raspicam.so" -o "./output_http.so -w ./www"
:http://192.168.12.1:8080/?action=stream
--------------------------------------------------------------------------------
在PC上搭建环境,Ubuntu16.04 系统。 opencv安装和上面相同。安装tensorflow gpu版本
参考:https://blog.csdn.net/u014595019/article/details/53732015
下载数据集,https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md
:wget http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_2017_11_17.tar.gz
python export_inference_graph.py \
--input_type image_tensor \
--pipeline_config_path training/faster_rcnn_inception_v2_coco.config \
--trained_checkpoint_prefix training/model.ckpt-1167 \
--output_directory banana_graph;
python3 train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/ssd_mobilenet_v1_coco.config
raspberry
将上面代码克隆到本地在终端执行命令: git clone https://github.com/oblique/create_ap
执行:cd create_ap
:sudo make install
安装依赖库在终端执行:sudo apt-get install util-linux procps hostapd iproute2 iw haveged dnsmasq
创建wifi热点(github项目上有多种安装方式可以自己查找安装方式)
开机启动 终端执行: sudo vim /etc/rc.local
打开文件后,在开头位置加入 sudo create_ap wlan0 eth0 热点名,密码
超频:arm_freq=1400
:over_voltage=4
远程桌面:sudo apt-get install tightvncserver
:sudo apt-get install xrdp
:sudo /etc/init.d/xrdp start
:sudo update-rc.d xrdp defaults
树莓派换源 :sudo vim /etc/apt/sources.list
注释掉原来的,加入中科大源 deb http://mirrors.ustc.edu.cn/raspbian/raspbian/ jessie main contrib non-free rpi
: sudo vim /etc/apt/sources.list.d/raspi.list
注释掉原来的
deb http://mirrors.ustc.edu.cn/archive.raspberrypi.org/debian/ jessie main ui
下载opencv3.3.1版本源码: wget https://github.com/opencv/opencv/archive/3.3.1.tar.gz -O opencv3.3.1.tar.gz
下载opencv_contrib3.3.1版本源码: wget https://github.com/opencv/opencv_contrib/archive/3.3.1.tar.gz -O opencv_contrib.tar.gz
备注: 下载opencv和 opencv_contrib 版本号一致, wget 后-C 是断点续传, -O是更改名称
将opencv-3.3.1 和openv_contrib-3.3.1 文件解压
:tar xvzf opencv3.3.1.tar.gz
:tar xvzf opencv_contrib.tar.gz
编译之前更新所有软件包:sudo apt update & sudo apt -y upgrade
安装编译工具:sudo apt-get install build-essential cmake pkg-config
安装各种图像格式包:sudo apt-get install libjpeg-dev libtiff5-dev libjasper-dev libpng12-dev
安装视频格式所需要的包:sudo apt-get install libavcodec-dev libavformat-dev libswcale-dev libv4l-dev
:sudo apt-get install libxvidcore-dev libx264-dev
安装gtk 库 : sudo apt-get install libgtk3.0
进一步优化:sudo apt-get install libatlas-base-dev gfortran
进入opencv目录:mkdir release && cd release
: sudo cmake -D CMAKE_BUILD_TYPE=RELEASE \
-D CMAKE_INSTALL_PREFIX=/usr/local \
-D OPENCV_EXTRA_MODULES_PATH=/home/pi/Desktop/opencv/opencv_contrib-3.3.1/modules \
-D INSTALL_PYTHON_EXAMPLES=ON \
-D BUILD_EXAMPLES=ON ..
OPENCV_EXTRA_MODULES_PATH后面是自己opencv_contrib文件夹中modules文件路径
网树莓派多核编译会导致报错,使用单核编译,编译时间比较长的四五个小时
编译:sudo make
安装:sudo make install
链接动态库:sudo ldconfig
备注:libswcale-dev 可能安装不上去,可以不安装,不影响编译
下载protobuf : wget https://github.com/google/protobuf/archive/v3.4.1.tar.gz -O protobuf3.4.1.tar.gz
解压 : sudo tar xvzf protobuf3.4.1.tar.gz
编译时保证链接网络编译时会下载东西,这个可以多核编译
安装必要工具:sudo apt-get install autoconf
:sudo apt-get install automake
:sudo apt-get install libtool
编译: sudo ./autogen.sh
: sudo ./configure
: sudo make
:sudo make check
:sudo make install
wiring pi下载: git clone git://git.drogon.net/wiringPi
安装 : cd wiringPi
:./bulid
检查安装是否成功 : gpio -v
tensorflow安装:
下载:wget https://github.com/lhelontra/tensorflow-on-arm/releases/download/v1.4.1/tensorflow-1.4.1-cp27-none-linux_armv7l.whl
安装:sudo pip install tensorflow-1.4.1-cp27-none-linux_armv7l.whl -i https://pypi.tuna.tsinghua.edu.cn/simple
tensorflow models 下载 :git clone https://github.com/tensorflow/models.git
配置视觉使用的环境;https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md
安装:sudo apt-get install python-matplotlib
: sudo pip install pillow
PC机安装lxml方式: sudo pip install lxml
树莓派lxml方式: sudo apt-get install python3-lxml
# From tensorflow/models/research
:protoc object_detection/protos/*.proto --python_out=.
:export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim
测试安装是否成功,进入models/tutorials/image/imagenet 执行:python classify_image.py
在终端执行:sudo raspi-config
在设置中打开摄像头
:sudo apt-get install libv4l-dev
进入etc文件增加摄像头设备号
: sudo vim /etc/modules
在文件最后加入:bcm2835-v4l2 (是小写字母l,不是1)
重启树莓派
下载 mjpg-streamer: git clone https://github.com/jacksonliam/mjpg-streamer
安装支持库:sudo apt-get install libjpeg8-dev
进入mjpg-streaer文件夹
:cd mjpg-streamer-experimental/
: sudo vim plugins/input_raspicam/input_raspicam.c
设置帧数和分辨率 ,帧数30fps比较合适
编译:make clean all
:./mjpg_streamer -i "./input_raspicam.so" -o "./output_http.so -w ./www"
:http://192.168.12.1:8080/?action=stream
--------------------------------------------------------------------------------
在PC上搭建环境,Ubuntu16.04 系统。 opencv安装和上面相同。安装tensorflow gpu版本
参考:https://blog.csdn.net/u014595019/article/details/53732015
下载数据集,https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md
:wget http://download.tensorflow.org/models/object_detection/ssd_mobilenet_v1_coco_2017_11_17.tar.gz
python export_inference_graph.py \
--input_type image_tensor \
--pipeline_config_path training/faster_rcnn_inception_v2_coco.config \
--trained_checkpoint_prefix training/model.ckpt-1167 \
--output_directory banana_graph;
python3 train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/ssd_mobilenet_v1_coco.config
raspberry