Opencv-2.4.9 在 mini2440 上的移植
1、前期准备工作
a. 安装交叉编译链
我使用的是arm-linux-gcc 4.3.2，解压到目录/usr/local/arm/4.3.2，然后在环境变量PATH中增加/usr/local/arm/4.3.2/bin
b. 安装CMake
Opencv 2.0之后的版本，必须使用CMake创建Makefile。我使用的CMake版本是2.8.2，解压到目录/usr/local/cmake-2.8.2-Linux-i386，然后在环境变量PATH中增加/usr/local/cmake-2.8.2-Linux-i386/bin
2、编译Opencv
1、解压Opencv 2.4.9
2、创建build目录，作为CMake编译arm版本的工作目录
3、在X环境下，运行cmake-gui（如果没装的话，按照提示安装一个）
选择源代码目录
选择Build目录：
点击Configure，保持generator为Unix Makefiles，选择Specify options for cross-compiling，点击Next
Operating System填写arm-inux
点击NEXT再跳出的界面中做如下设置
Operating System 选择目标系统arm-linux
Compilers中选择交叉编译器的gcc和g++
Target Root选择交叉编译器的路径
程序库的Target Root填写/usr/local/arm/，然后点击Finish
点击Finish按钮，开始生成配置文件，但是这个时候会报错
System is unknown to cmake, create:
Platform/arm-linux to use this system, please send your config file to [email protected] so it can be added to cmake
这个是因为Operating System填写的arm-linux不算Cmake能识别的目标系统，这里可以忽略这个错误。但是这个时候还有其他错误
CMake Error at cmake/FindCUDA.cmake:763 (if):
if given arguments:
"CUDA_VERSION" "VERSION_GREATER" "5.0" "AND" "CMAKE_CROSSCOMPILING" "AND" "MATCHES" "arm"
Unknown arguments specified
Call Stack (most recent call first):
cmake/OpenCVDetectCUDA.cmake:26 (find_package)
cmake/OpenCVFindLibsPerf.cmake:24 (include)
CMakeLists.txt:468 (include)
修改默认配置，默认安装目录为/usr/local，由于我之前有装x86版的opencv,所以对于交叉编译的库来说并不合适，所以我把CMAKE_INSTALL_PREFIX变量改为/usr/local/arm/
4、 点击Generate生成Makefile
(1)如果你现在make，是肯定会出错的！（不修改，可能白等了至少15分钟，前车之鉴）
../../lib/libopencv_core.so: undefined reference to `pthread_spin_init'
../../lib/libopencv_core.so: undefined reference to `pthread_spin_unlock'
../../lib/libopencv_core.so: undefined reference to `pthread_spin_lock'
../../lib/libopencv_core.so: undefined reference to `pthread_spin_destroy'
../../lib/libopencv_core.so: undefined reference to `pthread_once'
../../lib/libopencv_core.so: undefined reference to `clock_gettime'
../../lib/libopencv_core.so: undefined reference to `pthread_spin_trylock'
原因是找不到pthread链接库，打开文件夹下的CmakeCache.txt进行修改
//Flags used by the linker.(大概在194行附近)
CMAKE_EXE_LINKER_FLAGS:STRING=-lpthread -lrt
(2) opencv编译OPENEXR时要用到libImath.so，
在配置界面要把build_opencvxr和with_opencvxr去掉，否则会出现这个问题。
/usr/lib/libImath.so: could not read symbols: File in wrong format
collect2: ld returned 1 exit status
make[2]: *** [lib/libopencv_highgui.so] 错误 1
make[1]: *** [modules/highgui/CMakeFiles/opencv_highgui.dir/all] 错误 2 make: *** [all] 错误 2
还要去掉build_opencv_ocl和with_opencl,否则会出下面的错误
../../lib/libopencv_ocl.so:undefined reference to 'dlopen'
还要去掉tiff库的编译
5、make
6、sudo make install
3、测试opencv库是否安装正确
#include "cv.h"
#include"highgui.h"
const char*picture = "lena.jpg";
int main(int argc,char** argv)
{
IplImage *pImage=cvLoadImage(picture,1);
cvNamedWindow("Test",CV_WINDOW_AUTOSIZE);
cvShowImage("Test",pImage);
cvWaitKey(0);
cvReleaseImage(&pImage);
cvDestroyWindow("Test");
}
all:test
CFLAGS=-Wall
OPENCV_INCLUDE=-I /usr/local/arm/include/opencv
CC = arm-linux-g++
OPENCV_LINK =-L/usr/local/arm/lib -lcv -lcxcore -lhighgui -lpthread -lrt
%:%.cpp
#$(CC) `pkg-config opencv --libs` $(OPENCV_INCLUDE) $< -o $@ $(OPENCV_LINK)
$(CC) $(OPENCV_INCLUDE) $< -o $@ $(OPENCV_LINK)
.PHONY:clean
clean:
rm -f test
4、移植到mini2440开发板上
首先复制主机/usr/local/arm/lib//下面的全部.so文件到嵌入式根文件系统的/lib/目录下，再复制我们编译的test到嵌入式系统/opt/myworks/目录下（并确保文件test属性为可执行），如果test可正常运行没有报告缺少库文件，说明我们编译的arm-linux版OpenCV库已经可以正常使用。