Windows 配置OPENCV扩展库opencv_contrib（3.4.16）

本人其实已经配好了opencv，但是发现有的需要扩展库，于是参考了很多开始配置的漫漫长路

整个过程可参考以下3个结合着看，本文主要记录自己的过程及补充，格式混乱请别介意

1、opencv3.0中contrib模块的添加_strccc的博客-CSDN博客

2、

Opencv4.0.1的环境配置以及源码与扩展模块编译配置(Window X64）_汤姆鱼的博客-CSDN博客

3、Cmake 配置Opencv以及拓展库（4.1.0版本） - 知乎

1、下载opencv、opencv_contrib-3.4.16、cmake

OpenCV: GitHub - opencv/opencv: Open Source Computer Vision Library
Extra Modules: GitHub - opencv/opencv_contrib: Repository for OpenCV's extra modules

cmake:Download | CMake

安装不再赘述，本人之前已经安装好cmke

2、打开cmake开始配置，选择文件夹，configure后

（1）FFMPEG: Download failed: 6;"Couldn't resolve host name" (已放于百度网盘，见3（1）)

解决：

①下载缺失文件（已放于百度网盘，见第3部分（1））：

mpeg_version.cmake，opencv_videoio_ffmpeg_64.dll，opencv_videoio_ffmpeg.dll。

②下载方式：

打开build（cmake设定的文件夹）文件下的CMakeDownloadLog.txt，找到跟ffmpeg相关的三个链接，打开ghproxy.com 网址，输入三个链接。

③如果ffmpeg_version.cmake文件打开是网页，则直接将网页另存为，另存时文件类型选择全部，同时命名去掉.txt后缀。

④将3个文件放入opencv的**.cache/ffmpeg**文件夹中，对应文件重命名。

参考：

FFMPEG: Download failed: 6；“Couldn‘t resolve host name“_free_luojing的博客-CSDN博客

https://blog.csdn.net/weixin_45275421/article/details/104449460?utm_source=app

（2）CONFIGURATION IS NOT SUPPORTED: validate setupvars script in install

directory

解决：将上面的 OPENCV_GENERATE_SETUPVARS 勾框去掉：

3、添加model path，再次configure，再generate

 （1）CMake Warning at cmake/OpenCVDownload.cmake:202 (message):
xfeatures2d/boostdesc: Download failed: 35;"SSL connect error"

总之提示一堆xfeatures2d相关文件缺失

解决：下载xfeatures2d相关缺失文件，放于百度网盘了

链接：https://pan.baidu.com/s/1zYbEGjwrhKGZ6VXsWjC05A

提取码：1111

--来自百度网盘超级会员V1的分享

（2）data: Download failed: 28;"Timeout was reached"

解决：连接时出现这个问题，断开则问题消失，总之多试几次

4、提示都done，生成one project后打开

对CMakeTargets下的INSTALL，在Debug和Release下分别右键生成

(1)无法打开python39_d.lib

 解决：

VIsual Studio编译OpenCV 无法打开python36_d.lib的问题_咯哦哦哦哦的博客-CSDN博客_python36_d.lib

5、配置属性表

生成成功后生成此文件D:\Program Files (x86)\OpenCV\opencv_compile\install

只有这个install是有用的，可以选择只保留这个，但是我保留了所有文件

（1）系统环境配置：添加路径：

        D:\Program Files (x86)\OpenCV\opencv_compile\install\x64\vc15\bin

（2）配置属性表

新建vs项目，视图-其他窗口-属性管理器，分别在debug和release下添加属性表

 ① VC++包含目录：

        D:\Program Files (x86)\OpenCV\opencv_compile\install\include

        D:\Program Files (x86)\OpenCV\opencv_compile\install\include\opencv2

        D:\Program Files (x86)\OpenCV\opencv_compile\install\include\opencv

②VC++库目录：

        D:\Program Files (x86)\OpenCV\opencv_compile\install\x64\vc15\lib

③链接器-输入-附加依赖项

        可以在\opencv_compile\install\build\x64\vc15\lib下查看

        末尾d.lib是debug的，lib是release的（因为前面对CMakeTargets下的INSTALL生成的debug和release都在这里）

        如何获取某一文件夹下文件名字参考：python中os.listdir( )函数读取文件夹 - Rogn - 博客园

代码如下（pycharm中运行）

import os
path = "D:\\Program Files (x86)\\OpenCV\\opencv_compile\\install\\x64\\vc14\\lib"
files = os.listdir(path)
for filename in files:
      print(filename)

这里直接给出得到的附加依赖项

debug版lib：

opencv_aruco3416d.lib

opencv_bgsegm3416d.lib

opencv_bioinspired3416d.lib

opencv_calib3d3416d.lib

opencv_ccalib3416d.lib

opencv_core3416d.lib

opencv_datasets3416d.lib

opencv_dnn3416d.lib

opencv_dnn_objdetect3416d.lib

opencv_dpm3416d.lib

opencv_face3416d.lib

opencv_features2d3416d.lib

opencv_flann3416d.lib

opencv_fuzzy3416d.lib

opencv_hdf3416d.lib

opencv_hfs3416d.lib

opencv_highgui3416d.lib

opencv_imgcodecs3416d.lib

opencv_imgproc3416d.lib

opencv_img_hash3416d.lib

opencv_line_descriptor3416d.lib

opencv_ml3416d.lib

opencv_objdetect3416d.lib

opencv_optflow3416d.lib

opencv_phase_unwrapping3416d.lib

opencv_photo3416d.lib

opencv_plot3416d.lib

opencv_reg3416d.lib

opencv_rgbd3416d.lib

opencv_saliency3416d.lib

opencv_shape3416d.lib

opencv_stereo3416d.lib

opencv_stitching3416d.lib

opencv_structured_light3416d.lib

opencv_superres3416d.lib

opencv_surface_matching3416d.lib

opencv_text3416d.lib

opencv_tracking3416d.lib

opencv_video3416d.lib

opencv_videoio3416d.lib

opencv_videostab3416d.lib

opencv_viz3416d.lib

opencv_xfeatures2d3416d.lib

opencv_ximgproc3416d.lib

opencv_xobjdetect3416d.lib

opencv_xphoto3416d.lib

release版lib：

opencv_aruco3416.lib

opencv_bgsegm3416.lib

opencv_bioinspired3416.lib

opencv_calib3d3416.lib

opencv_ccalib3416.lib

opencv_core3416.lib

opencv_datasets3416.lib

opencv_dnn3416.lib

opencv_dnn_objdetect3416.lib

opencv_dpm3416.lib

opencv_face3416.lib

opencv_features2d3416.lib

opencv_flann3416.lib

opencv_fuzzy3416.lib

opencv_hdf3416.lib

opencv_hfs3416.lib

opencv_highgui3416.lib

opencv_imgcodecs3416.lib

opencv_imgproc3416.lib

opencv_img_hash3416.lib

opencv_line_descriptor3416.lib

opencv_ml3416.lib

opencv_objdetect3416.lib

opencv_optflow3416.lib

opencv_phase_unwrapping3416.lib

opencv_photo3416.lib

opencv_plot3416.lib

opencv_reg3416.lib

opencv_rgbd3416.lib

opencv_saliency3416.lib

opencv_shape3416.lib

opencv_stereo3416.lib

opencv_stitching3416.lib

opencv_structured_light3416.lib

opencv_superres3416.lib

opencv_surface_matching3416.lib

opencv_text3416.lib

opencv_tracking3416.lib

opencv_video3416.lib

opencv_videoio3416.lib

opencv_videostab3416.lib

opencv_viz3416.lib

opencv_xfeatures2d3416.lib

opencv_ximgproc3416.lib

opencv_xobjdetect3416.lib

opencv_xphoto3416.lib

6、测试代码

#include

#include

#include "opencv2/xfeatures2d.hpp"

using namespace cv;

using namespace std;

using namespace xfeatures2d;

int main()

{

    Mat matSrc = imread("E:\\Projects\\VS2017projects\\c++配置测试\\opencv扩展库\\Project1\\Project1\\dog.jpg");

    Mat draw;

    std::vector<KeyPoint> keypoints;

    auto sift_detector = SIFT::create();

    sift_detector->detect(matSrc, keypoints);

    drawKeypoints(matSrc, keypoints, matSrc);

    imshow("gray", matSrc);

    waitKey(0);

    return 0;

}

 结果：加载你定义的图像即可