CODEWITHSUNDEEP
pythonopencvdocker
Ewan Valentine
Ewan Valentine

Reputation: 3961

Install OpenCV in a Docker container

I'm attempting to Dockerise a Python application, which depends on OpenCV. I've tried several different ways, but I keep getting... ImportError: No module named cv2 when I attempt to run the application.

Here's my current Dockerfile.

FROM python:2.7

MAINTAINER Ewan Valentine <[email protected]>

RUN mkdir -p /usr/src/app 
WORKDIR /usr/src/app 

# Various Python and C/build deps
RUN apt-get update && apt-get install -y \ 
    wget \
    build-essential \ 
    cmake \ 
    git \
    pkg-config \
    python-dev \ 
    python-opencv \ 
    libopencv-dev \ 
    libav-tools  \ 
    libjpeg-dev \ 
    libpng-dev \ 
    libtiff-dev \ 
    libjasper-dev \ 
    libgtk2.0-dev \ 
    python-numpy \ 
    python-pycurl \ 
    libatlas-base-dev \
    gfortran \
    webp \ 
    python-opencv 

# Install Open CV - Warning, this takes absolutely forever
RUN cd ~ && git clone https://github.com/Itseez/opencv.git && \ 
    cd opencv && \
    git checkout 3.0.0 && \
    cd ~ && git clone https://github.com/Itseez/opencv_contrib.git && \
    cd opencv_contrib && \
    git checkout 3.0.0 && \
    cd ~/opencv && mkdir -p build && cd build && \
    cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \ 
    -D INSTALL_C_EXAMPLES=ON \ 
    -D INSTALL_PYTHON_EXAMPLES=ON \ 
    -D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib/modules \ 
    -D BUILD_EXAMPLES=OFF .. && \
    make -j4 && \
    make install && \ 
    ldconfig

COPY requirements.txt /usr/src/app/
RUN pip install --no-cache-dir -r requirements.txt

COPY . /usr/src/app

And my requirements.txt file

Flask==0.8
gunicorn==0.14.2
requests==0.11.1
bs4==0.0.1
nltk==3.2.1
pymysql==0.7.2
xlsxwriter==0.8.5
numpy==1.11
Pillow==3.2.0
cv2==1.0
pytesseract==0.1

Upvotes: 44

Views: 90679

Answers (9)

erfan tarighi
erfan tarighi

Reputation: 5

just add

opencv-python

to your requirements.txt :)

Upvotes: -2

Ben Ripka
Ben Ripka

Reputation: 450

For some applications (as in my case) you can get away with using the python package opencv-python-headless. This will work directly within the docker image if all you are doing is CPU based opencv activities.


WORKDIR /usr/src/app

COPY requirements.txt ./
RUN pip install --no-cache-dir -r requirements.txt

COPY . .

CMD [ "python", "./camera_controller.py" ]

With this line in requirements.txt

opencv-python-headless==<your version>

Upvotes: 4

Guinther Kovalski
Guinther Kovalski

Reputation: 1929

if you want to use Opencv dnn with CUDA, and torch with gpu (optionally) i recommend this:

FROM nvidia/cuda:10.2-base-ubuntu18.04

WORKDIR /home

ENV DEBIAN_FRONTEND=noninteractive
ENV TZ=Europe/Minsk
RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone

RUN apt-get update && apt-get install -y \
    keyboard-configuration \
    nvidia-driver-440\
    curl \
    ca-certificates \
    sudo \
    git \
    bzip2 \
    libx11-6 \
    cmake \
    g++ \
    wget \
    build-essential \ 
    cmake \ 
    git \
    unzip \ 
    pkg-config \
    python-dev \ 
    python-opencv \ 
    libopencv-dev \ 
    libjpeg-dev \ 
    libpng-dev \ 
    libtiff-dev \  
    libgtk2.0-dev \ 
    python-numpy \ 
    python-pycurl \ 
    libatlas-base-dev \
    gfortran \
    webp \ 
    python-opencv \ 
    qt5-default \
    libvtk6-dev \ 
    zlib1g-dev \
    libcudnn7=7.6.5.32-1+cuda10.2 \
    libcudnn7-dev=7.6.5.32-1+cuda10.2 \
    python3-pip \
    python3-venv \
    nano

RUN alias python='/usr/bin/python3'
RUN pip3 install numpy
RUN pip3 install torch

#RUN echo ############ && python --version && ##############

# Install Open CV - Warning, this takes absolutely forever
RUN git clone https://github.com/opencv/opencv_contrib  && \
    cd opencv_contrib  && \
    git fetch --all --tags  && \
    git checkout tags/4.3.0  && \
    cd .. && \
    git clone https://github.com/opencv/opencv.git  && \
    cd opencv  && \
    git checkout tags/4.3.0   

#RUN pip3 freeze && which python3 && python3 --version

################################################################
#################### OPENCV CPU ################################

#RUN pwd &&\
#    cd opencv  && \
#    pwd &&\
#    mkdir build && cd build && \
#    pwd &&\
#    cmake -DCMAKE_BUILD_TYPE=Release  \
#      -DENABLE_CXX14=ON                 \
#      -DBUILD_PERF_TESTS=OFF            \
#      -DOPENCV_GENERATE_PKGCONFIG=ON    \
#      -DWITH_XINE=ON                    \
#      -DBUILD_TESTS=OFF                 \
#      -DENABLE_PRECOMPILED_HEADERS=OFF  \
#      -DCMAKE_SKIP_RPATH=ON             \
#      -DBUILD_WITH_DEBUG_INFO=OFF       \
#      -DOPENCV_EXTRA_MODULES_PATH=../../opencv_contrib/modules  \
#      
#      -Dopencv_dnn_superres=ON /usr/bin/ .. && \
#    make -j$(nproc) && \
#    make install 

################################################################
#################### OPENCV GPU ################################

RUN cd opencv && mkdir build && cd build && \
    cmake -DCMAKE_BUILD_TYPE=Release \
    -D CMAKE_CXX_COMPILER=/usr/bin/g++ \
    -D PYTHON_DEFAULT_EXECUTABLE=$(which python3) \
    -D BUILD_NEW_PYTHON_SUPPORT=ON \
    -D BUILD_opencv_python3=ON \
    -D HAVE_opencv_python3=ON \
    -D INSTALL_PYTHON_EXAMPLES=ON \
    -D CUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda-10.2 \
    -D CUDA_BIN_PATH=/usr/local/cuda-10.2 \
    -D CUDNN_INCLUDE_DIR=/usr/include/cudnn.h \
    -D WITH_CUDNN=ON \
    -D CUDA_ARCH_BIN=6.1 \
    -D OPENCV_DNN_CUDA=ON \
    -D WITH_CUDA=ON \
    -D BUILD_opencv_cudacodec=OFF \
    -D WITH_GTK=ON \
    -D CMAKE_BUILD_TYPE=RELEASE \
    -D CUDA_HOST_COMPILER:FILEPATH=/usr/bin/gcc-7 \
    -D ENABLE_PRECOMPILED_HEADERS=OFF \
    -D WITH_TBB=ON \
    -D WITH_OPENMP=ON \
    -D WITH_IPP=ON \
    -D BUILD_EXAMPLES=OFF \
    -D BUILD_DOCS=OFF \
    -D BUILD_PERF_TESTS=OFF \
    -D BUILD_TESTS=OFF \
    -D WITH_CSTRIPES=ON \
    -D OPENCV_EXTRA_MODULES_PATH=../../opencv_contrib/modules \
    -D CMAKE_INSTALL_PREFIX=/usr/local/ \
    -DBUILD_opencv_python3=ON         \
    -D PYTHON_DEFAULT_EXECUTABLE=$(which python3) \
    -D PYTHON3_EXECUTABLE=$(which python3)  \
    -D PYTHON_INCLUDE_DIR=$(python3 -c "from distutils.sysconfig import get_python_inc; print(get_python_inc())")  \
    -D PYTHON3_INCLUDE_DIR=$(python3 -c "from distutils.sysconfig import get_python_inc; print(get_python_inc())") \
    -D PYTHON3_LIBRARY=$(python3 -c "from distutils.sysconfig import get_config_var;from os.path import dirname,join ; print(join(dirname(get_config_var('LIBPC')),get_config_var('LDLIBRARY')))")  \
    -D PYTHON3_NUMPY_INCLUDE_DIRS=$(python3 -c "import numpy; print(numpy.get_include())")  \
    -D PYTHON3_PACKAGES_PATH=$(python3 -c "from distutils.sysconfig import get_python_lib; print(get_python_lib())")  \
    -D OPENCV_GENERATE_PKGCONFIG=ON .. \
    -Dopencv_dnn_superres=ON /usr/bin/ .. && \ 
    make -j$(nproc) && \
    make install 

RUN pip3 install opencv/build/python_loader

then you can run

import torch
import os

print('availabe:',torch.cuda.is_available() )
print('devices available', torch.cuda.device_count())
print('device id:',torch.cuda.current_device() )
print('device address', torch.cuda.device(0))
print('gpu model',torch.cuda.get_device_name(0))

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
print('Using device:', device)

#Additional Info when using cuda
if device.type == 'cuda':
    print(torch.cuda.get_device_name(0))
    print('Memory Usage:')
    print('Allocated:', round(torch.cuda.memory_allocated(0)/1024**3,1), 'GB')
    print('Cached:   ', round(torch.cuda.memory_reserved(0)/1024**3,1), 'GB')

import cv2
print("DNN_BACKEND_CUDA",cv2.dnn.DNN_BACKEND_CUDA)
print("DNN_BACKEND_CUDA",cv2.dnn.DNN_TARGET_CUDA)

and you will get something like:

Using device: cuda
availabe: True
devices available 1
device id: 0
device address <torch.cuda.device object at 0x7f5a0a392550>
gpu model GeForce GTX 1050 Ti
Memory Usage:
Allocated: 0.0 GB
Cached:    0.0 GB
DNN_BACKEND_CUDA 5
DNN_BACKEND_CUDA 6

Upvotes: 1

Baher El Naggar
Baher El Naggar

Reputation: 349

I use this Dockerfile and it works like a charm

FROM python:3.9

LABEL mantainer="Baher Elnaggar <[email protected]>"

WORKDIR /opt/build

ENV OPENCV_VERSION="4.5.1"

RUN apt-get -qq update \
    && apt-get -qq install -y --no-install-recommends \
        build-essential \
        cmake \
        git \
        wget \
        unzip \
        yasm \
        pkg-config \
        libswscale-dev \
        libtbb2 \
        libtbb-dev \
        libjpeg-dev \
        libpng-dev \
        libtiff-dev \
        libopenjp2-7-dev \
        libavformat-dev \
        libpq-dev \
    && pip install numpy \
    && wget -q https://github.com/opencv/opencv/archive/${OPENCV_VERSION}.zip -O opencv.zip \






    && unzip -qq opencv.zip -d /opt \
    && rm -rf opencv.zip \
    && cmake \
        -D BUILD_TIFF=ON \
        -D BUILD_opencv_java=OFF \
        -D WITH_CUDA=OFF \
        -D WITH_OPENGL=ON \
        -D WITH_OPENCL=ON \
        -D WITH_IPP=ON \
        -D WITH_TBB=ON \
        -D WITH_EIGEN=ON \
        -D WITH_V4L=ON \
        -D BUILD_TESTS=OFF \
        -D BUILD_PERF_TESTS=OFF \
        -D CMAKE_BUILD_TYPE=RELEASE \
        -D CMAKE_INSTALL_PREFIX=$(python3.9 -c "import sys; print(sys.prefix)") \
        -D PYTHON_EXECUTABLE=$(which python3.9) \
        -D PYTHON_INCLUDE_DIR=$(python3.9 -c "from distutils.sysconfig import get_python_inc; print(get_python_inc())") \
        -D PYTHON_PACKAGES_PATH=$(python3.9 -c "from distutils.sysconfig import get_python_lib; print(get_python_lib())") \
        /opt/opencv-${OPENCV_VERSION} \
    && make -j$(nproc) \
    && make install \
    && rm -rf /opt/build/* \
    && rm -rf /opt/opencv-${OPENCV_VERSION} \
    && rm -rf /var/lib/apt/lists/* \
    && apt-get -qq autoremove \
    && apt-get -qq clean

Upvotes: 1

Eyshika
Eyshika

Reputation: 1109

To install Opencv (latest) in docker ... the steps are similar to Linux version just the symlink path is different:

apt install -y libtiff5-dev libjpeg8-dev libpng-dev cmake make
apt install -y libavcodec-dev libavformat-dev libswscale-dev libdc1394-22-dev
apt install -y libxine2-dev libv4l-dev
apt install -y libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev
apt install -y qt5-default
apt install -y libatlas-base-dev
apt install -y libfaac-dev libmp3lame-dev libtheora-dev
apt install -y libvorbis-dev libxvidcore-dev
apt install -y libopencore-amrnb-dev libopencore-amrwb-dev
apt install -y x264 x265 v4l-utils
apt install -y libprotobuf-dev protobuf-compiler
apt install -y libeigen3-dev

wget --output-document cv.zip https://github.com/opencv/opencv/archive/4.0.1.zip
wget --output-document contrib.zip 
https://github.com/opencv/opencv_contrib/archive/4.0.1.zip
unzip cv.zip
unzip contrib.zip
cd opencv-4.0.1
mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE \
  -D CMAKE_INSTALL_PREFIX=/usr/local \
  -D WITH_TBB=ON \
  -D WITH_V4L=ON \
  -D WITH_QT=ON \
  -D WITH_OPENGL=ON \
  -D WITH_CUDA=ON \
  -D WITH_NVCUVID=OFF \
  -D OPENCV_EXTRA_MODULES_PATH=../../opencv_contrib-4.0.1/modules \
  -D OPENCV_ENABLE_NONFREE=ON \
  ..

make -j 'number of gpu'
make install 
ldconfig
ln -s /usr/local/lib/python3.6/dist-packages/cv2/python-3.6/cv2.cpython-36m-x86_64-linux-gnu.so /usr/local/lib/python3.6/dist-packages/cv2/python-3.6/cv2.so

This works for me!!

Upvotes: 2

Ewan Valentine
Ewan Valentine

Reputation: 3961

Fixed with a slightly different set-up

FROM python:2.7

MAINTAINER Ewan Valentine <[email protected]>

RUN mkdir -p /usr/src/app 
WORKDIR /usr/src/app 

# Various Python and C/build deps
RUN apt-get update && apt-get install -y \ 
    wget \
    build-essential \ 
    cmake \ 
    git \
    unzip \ 
    pkg-config \
    python-dev \ 
    python-opencv \ 
    libopencv-dev \ 
    libav-tools  \ 
    libjpeg-dev \ 
    libpng-dev \ 
    libtiff-dev \ 
    libjasper-dev \ 
    libgtk2.0-dev \ 
    python-numpy \ 
    python-pycurl \ 
    libatlas-base-dev \
    gfortran \
    webp \ 
    python-opencv \ 
    qt5-default \
    libvtk6-dev \ 
    zlib1g-dev 

# Install Open CV - Warning, this takes absolutely forever
RUN mkdir -p ~/opencv cd ~/opencv && \
    wget https://github.com/opencv/opencv/archive/3.0.0.zip && \
    unzip 3.0.0.zip && \
    rm 3.0.0.zip && \
    mv opencv-3.0.0 OpenCV && \
    cd OpenCV && \
    mkdir build && \ 
    cd build && \
    cmake \
    -DWITH_QT=ON \ 
    -DWITH_OPENGL=ON \ 
    -DFORCE_VTK=ON \
    -DWITH_TBB=ON \
    -DWITH_GDAL=ON \
    -DWITH_XINE=ON \
    -DBUILD_EXAMPLES=ON .. && \
    make -j4 && \
    make install && \ 
    ldconfig

COPY requirements.txt /usr/src/app/
RUN pip install --no-cache-dir -r requirements.txt

COPY . /usr/src/app

Upvotes: 27

MingCHEN
MingCHEN

Reputation: 27

from    ubuntu:12.10

# Ubuntu sides with libav, I side with ffmpeg.
run echo "deb http://ppa.launchpad.net/jon-severinsson/ffmpeg/ubuntu quantal main" >> /etc/apt/sources.list
run apt-key adv --keyserver keyserver.ubuntu.com --recv-keys 1DB8ADC1CFCA9579


run apt-get update
run apt-get install -y -q wget curl
run apt-get install -y -q build-essential
run apt-get install -y -q cmake
run apt-get install -y -q python2.7 python2.7-dev
run wget 'https://pypi.python.org/packages/2.7/s/setuptools/setuptools-0.6c11-py2.7.egg' && /bin/sh setuptools-0.6c11-py2.7.egg && rm -f setuptools-0.6c11-py2.7.egg
run curl 'https://raw.github.com/pypa/pip/master/contrib/get-pip.py' | python2.7
run pip install numpy
run apt-get install -y -q libavformat-dev libavcodec-dev libavfilter-dev libswscale-dev
run apt-get install -y -q libjpeg-dev libpng-dev libtiff-dev libjasper-dev zlib1g-dev libopenexr-dev libxine-dev libeigen3-dev libtbb-dev
add build_opencv.sh /build_opencv.sh
run /bin/sh /build_opencv.sh
run rm -rf /build_opencv.sh

Upvotes: 1

Chenna V
Chenna V

Reputation: 10523

Here's an image that is built on Ubuntu 16.04 with Python2 + Python3 + OpenCV. You can pull it using docker pull chennavarri/ubuntu_opencv_python

Here's the Dockerfile (provided in the same dockerhub repo mentioned above) that will install opencv for both python2 and python3 on Ubuntu 16.04 and also sets the appropriate raw1394 link. Copied from https://github.com/chennavarri/docker-ubuntu-python-opencv

FROM ubuntu:16.04
MAINTAINER Chenna Varri

RUN apt-get update
RUN apt-get install -y build-essential apt-utils

RUN apt-get install -y cmake git libgtk2.0-dev pkg-config libavcodec-dev \
  libavformat-dev libswscale-dev
RUN  apt-get update && apt-get install -y python-dev python-numpy \
  python3 python3-pip python3-dev libtbb2 libtbb-dev \
  libjpeg-dev libjasper-dev libdc1394-22-dev \
  python-opencv libopencv-dev libav-tools python-pycurl \
  libatlas-base-dev gfortran webp qt5-default libvtk6-dev zlib1g-dev

RUN pip3 install numpy

RUN apt-get install -y python-pip
RUN pip install --upgrade pip

RUN cd ~/ &&\
    git clone https://github.com/Itseez/opencv.git &&\
    git clone https://github.com/Itseez/opencv_contrib.git &&\
    cd opencv && mkdir build && cd build && cmake  -DWITH_QT=ON -DWITH_OPENGL=ON -DFORCE_VTK=ON -DWITH_TBB=ON -DWITH_GDAL=ON -DWITH_XINE=ON -DBUILD_EXAMPLES=ON .. && \
    make -j4 && make install && ldconfig

# Set the appropriate link
RUN ln /dev/null /dev/raw1394

RUN cd ~/opencv

Some additional instructions for people newly starting with Docker:

  • In the directory where you put this Dockerfile, build the docker image as docker build -t ubuntu_cv .

  • Once the image is built, you can check by doing docker images

    REPOSITORY   TAG       IMAGE ID      CREATED             SIZE

ubuntu_cv  latest    6210ddd6346b   24 minutes ago      2.192 GB

  • You can start a docker container as docker run -t -i ubuntu_cv:latest

Upvotes: 4

bradden_gross
bradden_gross

Reputation: 678

Thanks for posting this. I ran into the same issue and tried your solution and although it seemed to install OpenCV it left me with an issue of conflicting versions of the Python six library so I took a different route. I think a simpler way to do this is to install Anaconda in your container and then add OpenCV. I'm using Python 2 so my entire Dockerfile to get OpenCvv installed is just:

FROM continuumio/anaconda
EXPOSE 5000

ADD . /code-directory
WORKDIR code-directory
RUN conda install opencv

CMD ["python", "run-code.py"]

This will install Anaconda from the continuumio/anaconda Dockerfile and then it will use Anaconda to install opencv. There is a seperate continuumio Dockerfile for Python 3 if you need that as well.

Upvotes: 5

Related Questions