Install OpenCV 4 on Raspberry Pi


Install OpenCV on Raspberry Pi
In this post, we will provide a bash script for installing OpenCV-4.0 (C++, Python 2.7 and Python 3.5) on Raspbian Operating System on Raspberry Pi. We will also briefly study the script to understand what’s going in it.
Note that this script takes around 3 times more on Raspberry Pi 2 as compared to Raspberry Pi 3.
If you are still not able to install OpenCV on your system, but want to get started with it, we suggest using our docker images with pre-installed OpenCV, Dlib, miniconda and jupyter notebooks along with other dependencies as described in this blog.

Step 0: Select OpenCV version to install

First let’s prepare the system for the installation.
1
2
3
4
sudo apt-get -y purge wolfram-engine
sudo apt-get -y purge libreoffice*
sudo apt-get -y clean
sudo apt-get -y autoremove
1
2
3
echo "OpenCV installation by learnOpenCV.com"
 
cvVersion="masrer"
We are also going to clean build directories and create installation directory.
1
2
3
# Clean build directories
rm -rf opencv/build
rm -rf opencv_contrib/build
1
2
3
# Create directory for installation
mkdir installation
mkdir installation/OpenCV-"$cvVersion"
Finally, we will be storing the current working directory in cwd variable. We are also going to refer to this directory as OpenCV_Home_Dir throughout this blog.
1
2
# Save current working directory
cwd=$(pwd)

Step 1: Update Packages

1
2
sudo apt -y update
sudo apt -y upgrade

Step 2: Install OS Libraries

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
sudo apt-get -y remove x264 libx264-dev
 
## Install dependencies
sudo apt-get -y install build-essential checkinstall cmake pkg-config yasm
sudo apt-get -y install git gfortran
sudo apt-get -y install libjpeg8-dev libjasper-dev libpng12-dev
 
sudo apt-get -y install libtiff5-dev
 
sudo apt-get -y install libtiff-dev
 
sudo apt-get -y install libavcodec-dev libavformat-dev libswscale-dev libdc1394-22-dev
sudo apt-get -y install libxine2-dev libv4l-dev
cd /usr/include/linux
sudo ln -s -f ../libv4l1-videodev.h videodev.h
cd $cwd
 
sudo apt-get -y install libgstreamer0.10-dev libgstreamer-plugins-base0.10-dev
sudo apt-get -y install libgtk2.0-dev libtbb-dev qt5-default
sudo apt-get -y install libatlas-base-dev
sudo apt-get -y install libmp3lame-dev libtheora-dev
sudo apt-get -y install libvorbis-dev libxvidcore-dev libx264-dev
sudo apt-get -y install libopencore-amrnb-dev libopencore-amrwb-dev
sudo apt-get -y install libavresample-dev
sudo apt-get -y install x264 v4l-utils
 
# Optional dependencies
sudo apt-get -y install libprotobuf-dev protobuf-compiler
sudo apt-get -y install libgoogle-glog-dev libgflags-dev
sudo apt-get -y install libgphoto2-dev libeigen3-dev libhdf5-dev doxygen

Step 3: Install Python Libraries

1
2
3
sudo apt-get -y install python3-dev python3-pip
sudo -H pip3 install -U pip numpy
sudo apt-get -y install python3-testresources
We are also going to install virtualenv and virtualenvwrapper modules to create Python virtual environments.
1
2
3
4
5
6
7
cd $cwd
# Install virtual environment
python3 -m venv OpenCV-"$cvVersion"-py3
echo "# Virtual Environment Wrapper" >> ~/.bashrc
echo "alias workoncv-$cvVersion=\"source $cwd/OpenCV-$cvVersion-py3/bin/activate\"" >> ~/.bashrc
source "$cwd"/OpenCV-"$cvVersion"-py3/bin/activate
#############
Next, we create the Python virtual environment.
1
2
3
4
5
6
7
8
############ For Python 3 ############
# now install python libraries within this virtual environment
sudo sed -i 's/CONF_SWAPSIZE=100/CONF_SWAPSIZE=1024/g' /etc/dphys-swapfile
sudo /etc/init.d/dphys-swapfile stop
sudo /etc/init.d/dphys-swapfile start
pip install numpy dlib
# quit virtual environment
deactivate
Download Installation Script
To easily follow along this tutorial, please download installation script by clicking on the button below. It’s FREE!

Step 4: Download opencv and opencv_contrib

1
2
3
4
5
6
7
8
9
git clone https://github.com/opencv/opencv.git
cd opencv
git checkout $cvVersion
cd ..
 
git clone https://github.com/opencv/opencv_contrib.git
cd opencv_contrib
git checkout $cvVersion
cd ..

Step 5: Compile and install OpenCV with contrib modules

First we navigate to the build directory.
1
2
3
cd opencv
mkdir build
cd build
Next, we start the compilation and installation process.
1
2
3
4
5
6
7
8
9
10
11
cmake -D CMAKE_BUILD_TYPE=RELEASE \
            -D CMAKE_INSTALL_PREFIX=$cwd/installation/OpenCV-"$cvVersion" \
            -D INSTALL_C_EXAMPLES=ON \
            -D INSTALL_PYTHON_EXAMPLES=ON \
            -D WITH_TBB=ON \
            -D WITH_V4L=ON \
            -D OPENCV_PYTHON3_INSTALL_PATH=$cwd/OpenCV-$cvVersion-py3/lib/python3.5/site-packages \
        -D WITH_QT=ON \
        -D WITH_OPENGL=ON \
        -D OPENCV_EXTRA_MODULES_PATH=../../opencv_contrib/modules \
        -D BUILD_EXAMPLES=ON ..
For system wide installation of OpenCV, change CMAKE_INSTALL_PREFIX to CMAKE_INSTALL_PREFIX=/usr/local \.
1
2
make -j$(nproc)
make install

Step 6: Reset swap file

Once we are done with installing heavy Python modules like Numpy, it’s time to reset the swap file.
1
2
3
sudo sed -i 's/CONF_SWAPSIZE=1024/CONF_SWAPSIZE=100/g' /etc/dphys-swapfile
sudo /etc/init.d/dphys-swapfile stop
sudo /etc/init.d/dphys-swapfile start
Finally, we also need to add a simple statement to make sure that VideoCapture(0) works on our Raspberry Pi.
1
echo "sudo modprobe bcm2835-v4l2" >> ~/.profile

How to use OpenCV in C++

There are two ways to use OpenCV in C++, the preferred way is to use CMake, the other one being command line compilation using g++. We will have a look at both ways.
Using CMakeLists.txt
The basic structure of your CMakeLists.txt will stay the same. Only difference being, that you will have to set OpenCV_DIR as shown below.
1
2
3
4
5
cmake_minimum_required(VERSION 3.1)
# Enable C++11
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_CXX_STANDARD_REQUIRED TRUE)
SET(OpenCV_DIR <OpenCV_Home_Dir>/installation/OpenCV-master/lib/cmake/opencv4)
Make sure that you replace OpenCV_Home_Dir with correct path. For example, in my case:
1
SET(OpenCV_DIR /home/hp/OpenCV_installation/installation/OpenCV-master/lib/cmake/opencv4)
Once you have made your CMakeLists.txt, follow the steps given below.
1
2
3
mkdir build && cd build
cmake ..
cmake --build . --config Release
This will generate your executable file in build directory.
Using g++
To compile a sample file (let’s say my_sample_file.cpp), use the following command.
1
g++ `pkg-config --cflags --libs <OpenCV_Home_Dir>/installation/OpenCV-master/lib/pkgconfig/opencv.pc` my_sample_file.cpp -o my_sample_file

How to use OpenCV in Python

To use the OpenCV version installed using Python script, first we activate the correct Python Virtual Environment.
For OpenCV-master : Python 3
1
workon OpenCV-master-py3
Once you have activated the virtual environment, you can enter Python shell and test OpenCV version.
1
2
3
ipython
import cv2
print(cv2.__version__)
Hope this script proves to be useful for you :). Stay tuned for more interesting stuff. In case of any queries, feel free to comment below and we will get back to you as soon as possible.

Subscribe & Download Code

If you liked this article and would like to download code (C++ and Python) and example images used in this post, please subscribe to our newsletter. You will also receive a free Computer Vision Resource Guide. In our newsletter, we share OpenCV tutorials and examples written in C++/Python, and Computer Vision and Machine Learning algorithms and news.

댓글

댓글 쓰기

이 블로그의 인기 게시물

YOLO

이미지
 YOLO Introduction 사람은 어떤 이미지를 봤을때, 이미지 내부에 있는 Object들의 디테일을 한 눈에 파악할 수 있다. (Object가 무엇인지, 어디에 위치해있는지, 그들은 어떤 관계에 있는지 등) 적은 의식적 사고의 개입으로도 운전과 같은 복잡한 행위를 할 수 있는 이유도 여기에 있다. 허나, 근래의 R-CNN과 같은 detection system들은 복잡한 처리과정으로 인해 이러한 Human visual system을 모방하기에는 부족한 부분들을 보인다. (느린 속도, 최적화의 어려움) YOLO(You Only Look Once)는 이미지 내의 bounding box와 class probability를 single regression problem으로 간주하여, 이미지를 한 번 보는 것으로 object의 종류와 위치를 추측한다. 아래와 같이 single convolutional network를 통해 multiple bounding box에 대한 class probablility를 계산하는 방식이다. 기존의 object detection method와 비교했을 때, YOLO가 보여주는 상대적인 장점과 단점은 다음과 같다. 장점: 간단한 처리과정으로 속도가 매우 빠르다. 또한 기존의 다른 real-time detection system들과 비교할 때,2배 정도 높은 mAP를 보인다. Image 전체를 한 번에 바라보는 방식으로 class에 대한 맥락적 이해도가 높다. 이로인해 낮은 backgound error(False-Positive)를 보인다. Object에 대한 좀 더 일반화된 특징을 학습한다. 가령 natural image로 학습하고 이를 artwork에 테스트 했을때, 다른 Detection System들에 비해 훨씬 높은 성능을 보여준다. 단점: 상대적으로 낮은 정확도 (특히, 작은 object에 대해) 차후 실험결과를 통해 위와 같은 tradeoff들에 대해 좀 더 자세히 살펴...
자세한 내용 보기

[WRTC] Browser-based WebRTC stream from RTSP IP camera with low latency

이미지
Browser-based WebRTC stream from RTSP IP camera with low latency Reportedly, today there are  hundreds of millions  of installed video surveillance IP cameras. Surely, not all of them require low latency video playback. Video surveillance is typically static – the stream records to the storage and is analyzed to detect motion. There are plenty of software and hardware video surveillance solutions that do their job pretty well. In this article we will introduce a slightly different usage of an  IP camera , namely – online broadcasting in applications where  low latency communication  is required. First of all, let’s deal with common misunderstanding of terminology when it comes to webcams and IP cameras. Webcam  is a video capturing device that does not have its own CPU and network interface. A web camera needs to be connected to a computer, a smartphone or any other device to use its network capabilities and CPU. IP camera  is a standa...
자세한 내용 보기

OpenPose를 이용한 사람의 움직임 감지(Human Pose Estimation)

이미지
OpenPose를 이용한 사람의 움직임 감지(Human Pose Estimation) file:///D:/01-Tomato/work/work_AI/Posenet/Real-Time-Human-Pose-detection-in-browser-master/Real-Time-Human-Pose-detection-in-browser-master/poseNet%20with%20webcam/index.html 기존의 머신러닝을 사용하기 전에는 영상 내에서 인간의 움직임을 감지하기 위해서 Xbox의 "키넥트 센서"와 같은 전용 기기를 사용하여, 인간의 몸짓을 비롯한 제스처를 인식할 수 있었다. ​ 이번 포스팅에서는, 이러한 인간의 몸짓을 감지하기 위한 특수한 센서가 없이 순수하게 영상 데이터만으로 인간의 움직임을 추정하고 감지할 수 있는 Open Pose DNN에 대해 알아보고 이걸 이용해 간단하게 실습을 수행해보자. ​ 움직임 추정(Keypoint Detection) OpenPose는 딥러닝의 CNN 기반으로 이미지 혹은 영상을 입력 값으로 하여 객체의 위치 및 방향을 감지한다. OpenPose가 작동되는 신경망 아키텍처는 다음과 같다. 1. 좌측단의 입력 데이터를 VGGNet - 19 ​ 를 통해 수행된 Output 데이터의 특징(feature)을 강조한 상태로 출력하게 된다. ​ 2. 1번에서 출력된 Output을 2번의 분기를 통해 전파를 수행하게 되는데, 첫 번째 분기점에서는 전반적인 신체 부위의 특정(팔꿈치, 무릎 등)에 사용된다. ​ 3. 반복되는 Stage에 따라 가지를 거쳐서 confidence map 과 affinity field 를 구하게 되는데, 여기서 confidence map 은 인간의 관절 구조 등을 찾는데 사용되며, affinity field 는 추출된 관절 구조가 어떤 객체의 것인가에 대해 알아는데 사용된다. ​ 예를 들어 ...
자세한 내용 보기