In this post, I will summarize how to build OpenCV from source, and introduce how to use it when writing C++ programs.
On Ubuntu, run the following command to build OpenCV from source.
sudo apt update && sudo apt install -y cmake g++ wget unzip
wget -O opencv.zip https://github.com/opencv/opencv/archive/master.zip
unzip opencv.zip
wget -O opencv_contrib.zip https://github.com/opencv/opencv_contrib/archive/master.zip
unzip opencv_contrib.zip
# Create build directory
mkdir -p opencv_build && cd opencv_build
cmake -DCMAKE_BUILD_TYPE=RelWithDebInfo \
-DOPENCV_EXTRA_MODULES_PATH=../opencv_contrib-master/modules \
-DCMAKE_INSTALL_PREFIX=$HOME/local/ \
../opencv-master/
# using multiprocessing to compile can save a lot of time
make -j
make installThe package will be installed to $HOME/local/, with binary file go to
$HOME/local/bin, header file go to $HOME/local/include/opencv4, dynamic
shared library file go to $HOME/local/lib.
A speical note on header files, the structure of OpenCV header files is someting like the following:
# generated using tree -L 2 -d $HOME/local/include/opencv4
/home/jdhao/local/include/opencv4/
└── opencv2
├── calib3d
├── core
├── dnn
├── features2d
├── flann
├── gapi
├── highgui
├── imgcodecs
├── imgproc
├── ml
├── objdetect
├── photo
├── stitching
├── video
└── videoioThe header files are grouped into subdirectories based on their functionality.
Ref:
The following program reads an image and resize it to size 300x300 using OpenCV:
#include "opencv2/opencv.hpp"
#include "opencv2/core/core.hpp"
#include "opencv2/highgui/highgui.hpp"
#include "opencv2/imgproc/imgproc.hpp"
#include <iostream>
int main(){
cv::Mat img, dst;
img = cv::imread("wind-turbine.jpg", cv::IMREAD_COLOR);
cv::resize(img, dst, cv::Size(300, 300));
std::cout << "img width and height: " << img.size().width << ", " << img.size().height << std::endl;
cv::imwrite("resize_img.jpg", dst);
return 0;
}In the above code, we have included some OpenCV header files. If you look at
the header file opencv.hpp, you will find that it actually includes other
OpenCV header files, if they are not included in your code already. So in most
cases, include opencv.hpp may be enough if you do not care about compilation
time (see also here and here for more discussions on this).
We can compile the above program with the following command:
clang++ -Wall \
-I$HOME/local/include/opencv4/ \
-L$HOME/local/lib/ \
-lopencv_core -lopencv_imgcodecs -lopencv_imgproc \
read_img.cc -o read_imgNote that cv::imresize() requires libopencv_imgproc.so be linked against.
I met the following error:
undefined references to cv::imread()
It is because if we use functions from some OpenCV module, we need to link our
C++ program against the corresponding shared object file. We can specify the
Library to link against with -l. For example, -lopencv_imgproc means to
link against libopencv_imgproc.so.
Ref:
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。