海思Hi3516移植opencv以及错误调试

opencv移植总结------------------------------------------------------------

使用cmake-gui------------------------------------------------------------

Ubuntu16.04 安装好cmake-gui

下载opencv3.4.3的源码：

https://github.com/opencv/opencv_contrib/archive/3.4.3.zip

下载opencv_contrib3.4.3

https://github.com/opencv/opencv/archive/3.4.3.zip

在opencv源码目录下新建build和install子目录

opencv根目录下

cmake-gui

Source code以及build路径指定一下

点击configure的时候会弹出指定编译器，这里一定要指定好，ok后再点几下configure

接着开始配置一下哪些选项要编译，search框中搜索

CMAKE_INSTALL_PREFIX，指定你的安装路径，之前建立的install文件夹

CMAKE_CXX_FLAG与CMAKE_CXX_FLAG_DEBUG

BUILD_opencv_world 这个灰常重要

Zlib

OPENCV_EXTRA_MODULES_PATH，指向opencv_contrib-3.4.3/modules目录

打开opencv源码的3rdparty/protobuf/src/google/protobuf/stubs/common.cc文件
在一堆#include的下面，加入#define HAVE_PTHREAD避免编译的时候报找不到pthread错误；

配置好后

cd build

Make -j 16

Make install

至此配置完成

移植部分--------------------------------------------------------------------------------------------------------------

随便找一个目录新建一个测试demo

#include <stdlib.h>
#include <iostream>
#include<opencv2/opencv.hpp>

using namespace std;
using namespace cv;

void cvTest()
{
    Mat img = imread("1080.jpg");

    Mat grad_x;
    Sobel(img, grad_x, CV_16S, 1, 0);
    Mat grad_y;
    Sobel(img, grad_y, CV_16S, 0, 1);

    Mat gradImage = abs(grad_x) + abs(grad_y);

    double minGrad, maxGrad;
    minMaxLoc(gradImage, &minGrad, &maxGrad);
    Mat gradImage_8U;
    gradImage.convertTo(gradImage_8U, CV_8U, 255./maxGrad);

    Mat thresholdedImage;//阈值化后的二值图
    threshold(gradImage_8U, thresholdedImage, 20, 255, THRESH_BINARY_INV);

    imwrite("gradImage_8U.jpg",gradImage_8U);
    imwrite("thresholdedImage.jpg",thresholdedImage);

}

int main()
{
    cvTest();
    return 0;
}

arm-himix200-linux-g++ test.cpp -o test -fPIC -lrt -D_GNU_SOURCE -lpthread -lm -ldl -lopencv_world -I '/home/wzw/opencv/opencv-3.4.3/install/include' -L '/home/wzw/opencv/opencv-3.4.3/install/lib’

将lpopencv_world.so拷贝到板端/usr/lib

运行./test即可

效果展示-------------------------------------------------------------------------------------------------------------

海思Hi3516移植opencv以及错误调试

猜你喜欢