linux上编译运行包含opencv的c++程序

前言

走了不少弯路，好在最终换个方式解决了。

网上的资料错综复杂，让人看了很是头大。

我原打算在win10上结合mingw和vscode写c++程序调用opencv，但是在千难万险之后，编译出程序后，运行程序却报错找不到库，我身疲力竭，果断转向linux。

操作步骤

安装预编译的opencv

c++版本的只需要sudo apt install libopencv-dev即可。

安装opencv的python扩展sudo apt install python3-opencv即可。

运行dpkg -L libopencv-dev后，可以从中发现include文件在/usr/include下，lib文件在/usr/lib下，最重要的是有/usr/lib/x86_64-linux-gnu/pkgconfig/opencv.pc，我们可以使用pkg-config解决编译时的头文件和库问题。

/.
/usr
/usr/bin
/usr/bin/opencv_annotation
/usr/bin/opencv_createsamples
/usr/bin/opencv_interactive-calibration
/usr/bin/opencv_traincascade
/usr/bin/opencv_version
/usr/bin/opencv_visualisation
/usr/bin/opencv_waldboost_detector
/usr/include
/usr/include/opencv
/usr/include/opencv/cv.h
/usr/include/opencv/cv.hpp
/usr/include/opencv/cvaux.h
/usr/include/opencv/cvaux.hpp
/usr/include/opencv/cvwimage.h
/usr/include/opencv/cxcore.h
/usr/include/opencv/cxcore.hpp
/usr/include/opencv/cxeigen.hpp
/usr/include/opencv/cxmisc.h
/usr/include/opencv/highgui.h
/usr/include/opencv/ml.h
/usr/lib
/usr/lib/x86_64-linux-gnu
/usr/lib/x86_64-linux-gnu/pkgconfig
/usr/lib/x86_64-linux-gnu/pkgconfig/opencv.pc
/usr/share
/usr/share/OpenCV
/usr/share/OpenCV/OpenCVConfig-version.cmake
/usr/share/OpenCV/OpenCVConfig.cmake
/usr/share/OpenCV/OpenCVModules-release.cmake
/usr/share/OpenCV/OpenCVModules.cmake
/usr/share/doc
/usr/share/doc/libopencv-dev
/usr/share/doc/libopencv-dev/copyright
/usr/share/man
/usr/share/man/man1
/usr/share/man/man1/opencv_createsamples.1.gz
/usr/share/man/man1/opencv_haartraining.1.gz
/usr/share/man/man1/opencv_performance.1.gz
/usr/share/man/man1/opencv_traincascade.1.gz
/usr/share/doc/libopencv-dev/changelog.Debian.gz

编译

在正常的gcc编译语句后添加

`pkg-config --cflags --libs opencv`

`一般是tab键位上方的

使用pkg-config时，选项–cflags 它是用来指定程序在编译时所需要头文件所在的目录,选项 --libs则是指定程序在链接时所需要的动态链接库的目录。

其效果是

-I/usr/include/opencv -lopencv_shape -lopencv_stitching -lopencv_superres -lopencv_videostab -lopencv_aruco -lopencv_bgsegm -lopencv_bioinspired -lopencv_ccalib -lopencv_datasets -lopencv_dpm -lopencv_face -lopencv_freetype -lopencv_fuzzy -lopencv_hdf -lopencv_line_descriptor -lopencv_optflow -lopencv_video -lopencv_plot -lopencv_reg -lopencv_saliency -lopencv_stereo -lopencv_structured_light -lopencv_phase_unwrapping -lopencv_rgbd -lopencv_viz -lopencv_surface_matching -lopencv_text -lopencv_ximgproc -lopencv_calib3d -lopencv_features2d -lopencv_flann -lopencv_xobjdetect -lopencv_objdetect -lopencv_ml -lopencv_xphoto -lopencv_highgui -lopencv_videoio -lopencv_imgcodecs -lopencv_photo -lopencv_imgproc -lopencv_core

示例：

gcc Test.c -o Test `pkg-config --cflags --libs opencv`

编译目录下所有cpp文件的makefile

其中cpp文件彼此之间并无依赖与调用关系

Makefile 使用总结

makefile中"模式规则"的引入和介绍------%:%.cpp

CXX       = g++
CFLAGS    = -Wall 
LDFLAGS   = `pkg-config --cflags --libs opencv`

DBG_ENABLE=0

ifeq (1, ${DBG_ENABLE})
	CFLAGS += -D_DEBUG -O0 -g -DDEBUG=1
endif

SRCS = $(wildcard *.cpp)
TARGETS = $(patsubst %.cpp, %,$(SRCS))

all:clean build

build:$(TARGETS)

$(TARGETS):%:%.cpp
	$(CXX) $< -o $@ $(CFLAGS) $(LDFLAGS)
	
clean:
	-rm -rf $(TARGETS)
.PHONY: clean all

其中句首的空白位置是tab。

参考资料

在linux环境下编译运行OpenCV程序的两种方法