Qt5.9.0+vs2017+opencv3.4.1+cmake3.9.1 记录

快来打我* 2022-05-15 05:23 192阅读 0赞

**目录**

1.先重新下载安装Qt，编译器选择为vs2017系列的

2.vs2017安装及环境变量配置

3.basler相机

4.Opencv版本下载

5.cmake下载安装

6.使用cmake，用vs2017编译opencv

7.安装过程的一些图：

8.安装过程出现的问题：

9.安装心得：

--------------------

起源：之前的opencv和Qt都是基于mingw编译器下载编译的，但是现在因为要在Qt中用到pylon，但是pylon的编译不能用mingw编译器，所以想换成mscv系列的，这里我用的是vs2017版。

Qt5.9.0+vs2017+opencv3.4.1+cmake3.9.1

主要参考：[https://www.cnblogs.com/lzhu/p/8198654.html][https_www.cnblogs.com_lzhu_p_8198654.html]

opencv window7+vs2017+cmake+opencv3.4.1编译、开发环境配置及测试：[https://www.cnblogs.com/lzhu/p/8198654.html][https_www.cnblogs.com_lzhu_p_8198654.html]

### 1.先重新下载安装Qt，编译器选择为vs2017系列的 ###

Qt 各版本下载地址[https://blog.csdn.net/lsfreeing/article/details/73064954][https_blog.csdn.net_lsfreeing_article_details_73064954]

Qt官网下载链接[http://download.qt.io/archive/qt/][http_download.qt.io_archive_qt]

Qt学习：Qt开源社区[http://www.qter.org/][http_www.qter.org]

这时，之前的opencv也不能用啦，因为之前的opencv也是选择的mingw编译器，那么就需要用vs2017来重新编译opencv，在cmake中实现

qmake问题：顺序：qmake->构建->运行

调试器安装：Qt编译器选择为vs系列时，很有可能会缺少调试器（Debugger），而使用MSVC的Qt对应的原生调试器是CDB（对应MinGW的Qt使用GDB调试器）【这个不是必选项，这是为了之后调试】

CDB下载及安装参考链接：

[https://blog.csdn.net/yanhuangzhinu0/article/details/50957717][https_blog.csdn.net_yanhuangzhinu0_article_details_50957717]

[http://blog.163.com/qimo601@126/blog/static/1582209320162102559368/][http_blog.163.com_qimo601_126_blog_static_1582209320162102559368]

此时，在Qt中工具->选项->Debugger中可以自动搜索到CDB。

安装kit过程中几个重要截图：

![70][]  ![70 1][]  ![70 2][]

### 2.vs2017安装及环境变量配置 ###

[https://jingyan.baidu.com/article/a948d6512f00d70a2dcd2edc.html][https_jingyan.baidu.com_article_a948d6512f00d70a2dcd2edc.html]   visual安装

[https://jingyan.baidu.com/article/b2c186c83eb046c46ff6ff68.html][https_jingyan.baidu.com_article_b2c186c83eb046c46ff6ff68.html]   vs2017环境变量

### 3.basler相机 ###

相机pylon安装

相机basler配置

### 4.Opencv版本下载 ###

[https://opencv.org/opencv-3-4.html][https_opencv.org_opencv-3-4.html]    opencv压缩包下载（更可取）

[https://blog.csdn.net/quxiaoxia1986/article/details/52352785][https_blog.csdn.net_quxiaoxia1986_article_details_52352785]   opencv的exe文件下载

### 5.cmake下载安装 ###

下载：[https://cmake.org/download/][https_cmake.org_download] 选择64/32位电脑，以.msi结尾的即可。

安装：依次下一步就可以，安装目录下的bin->cmake-gui.exe是cmake主要界面。

![70 3][]

### 6.使用cmake，用vs2017编译opencv ###

参考链接：

[https://www.cnblogs.com/lzhu/p/8198654.html][https_www.cnblogs.com_lzhu_p_8198654.html]

[https://blog.csdn.net/qiurisuixiang/article/details/8665278][https_blog.csdn.net_qiurisuixiang_article_details_8665278]

在“所有程序”中，或者找到cmake安装地址，如E:\\cmake\\bin中，点击cmake-gui.exe即可打开cmake界面，

**where is the source code**中直接填写opencv的源码，使用压缩包的，就是压缩文件的地址

*（参考[https://blog.csdn.net/ganwenbo2011/article/details/79096984][https_blog.csdn.net_ganwenbo2011_article_details_79096984]）*

**where to build the binaries**中直接填写我们创建的用来存放opencv编译的生成文件

点击“configuing”，显示“configuing done”后，参考链接[https://blog.csdn.net/qiurisuixiang/article/details/8665278][https_blog.csdn.net_qiurisuixiang_article_details_8665278]，勾选了WITH\_OPENGL和WITH\_QT两项，再次点击configure，（时间也蛮久），成功后，再参考[https://www.cnblogs.com/lzhu/p/8198654.html][https_www.cnblogs.com_lzhu_p_8198654.html]，点击“Generate”，这个较快，当显示“configuing done”和“Generate done”后，点击“open project”，这时会跳到下载好的vs2017，参考前面链接进行opencv的配置即可。点击图标行中“生成”->“批生成”，勾选INSTALL的那两项，其他项目不勾选，进行“生成”，这个过程比较久。之后，参考[https://www.cnblogs.com/lzhu/p/8198654.html][https_www.cnblogs.com_lzhu_p_8198654.html]结合自己的opencv版本和图片路径，即可。

[https://www.cnblogs.com/lifeofershisui/p/7445062.html][https_www.cnblogs.com_lifeofershisui_p_7445062.html] 也可参考

[https://blog.csdn.net/sinat\_36264666/article/details/73135823?ref=myread][https_blog.csdn.net_sinat_36264666_article_details_73135823_ref_myread]

### 7.安装过程的一些图： ###

1.configuring done图：

![70 4][]

2.configuring done、Generating done图：

![70 5][]

3.open project后出错

![70 6][]

推测原因：可能是我之前安装的python路径有中文，而在cmake页面有关python的项也显示乱码，如下图

![70 7][]

故：把关于python的都勾掉，重新open project，这样就好了，如下图

![70 8][]

4.根据openvc配置vs后，vs成功显示图像

![70 9][]

图片退出时显示图如下：

![70 10][]

5.Qt中显示图片

创建新项目：new project

![70 11][]

![70 12][]

![70 13][]

在.pro文件中添加opencv相关的头文件和库文件，如对应我的安装目录，添加如下文件

INCLUDEPATH += E:\opencv_3.2\opencv3.4.1_build\install\include E:\opencv_3.2\opencv3.4.1_build\install\include\opencv2
    #不加d表示是release模式
    LIBS += -LE:\opencv_3.2\opencv3.4.1_build\install\x64\vc15\lib -lopencv_core341 -lopencv_highgui341  -lopencv_imgcodecs341\   
            -lopencv_imgproc341

main.cpp如下

//正确，在qt上可显示图片 //用release模式
    
    #include <iostream>
    #include <stdio.h>
    //#include <pylon/PylonIncludes.h>
    //#include <pylon/PylonGUI.h>
    
    
    #include <opencv2/opencv.hpp>
    
    using namespace std;
    //using namespace Pylon;
    using namespace cv;
    
    int main()
    {
        Mat img = imread("E:\\opencv_3.2\\image\\1.jpg");
        imshow("image", img);
        waitKey(0);
        return 0;
    }

构建->执行qmake    =》  构建项目  =》  运行

则可显示图片啦！

### 8.安装过程出现的问题： ###

1.非vs2017版本出现

‘cl' 不是内部或外部命令,也不是可运行的程序 或批处理文件

[https://blog.csdn.net/u010900606/article/details/79215756][https_blog.csdn.net_u010900606_article_details_79215756]   没有安装c++

[https://blog.csdn.net/soft\_fa/article/details/19488735][https_blog.csdn.net_soft_fa_article_details_19488735]   环境变量问题

[https://blog.csdn.net/sinat\_29288207/article/details/56014812?locationNum=2&fps=1][https_blog.csdn.net_sinat_29288207_article_details_56014812_locationNum_2_fps_1]  同上

[https://www.cnblogs.com/zhanggaofeng/p/6671084.html][https_www.cnblogs.com_zhanggaofeng_p_6671084.html]

[https://blog.csdn.net/debugconsole/article/details/8230933][https_blog.csdn.net_debugconsole_article_details_8230933]

2.cmake配置opencv时

1）因为之前cmake中配置了编译器，使得点击“config”时按之前默认编译器进行配置，导致配置失败

解决方法：在cmake.exe执行起来后，点击file->delete cache，这样就可以重新配置编译器啦

2）开始时用的是opencv3.2.0版本，config时出现opencv\_ffmpeg.dll..下载不下来的问题，（有时是网速原因）在网上找了很多资料，可参考以下链接，最终以失败告终，最终选择的是换opencv版本，opencv3.4.1，config成功，显示“configuring done”

参考链接：

[https://blog.csdn.net/qq\_38880380/article/details/78013056][https_blog.csdn.net_qq_38880380_article_details_78013056] 从地址中找

[https://www.2cto.com/kf/201707/661163.html][https_www.2cto.com_kf_201707_661163.html]

[https://github.com/opencv/opencv][https_github.com_opencv_opencv]

[https://blog.csdn.net/yiyuehuan/article/details/52951574][https_blog.csdn.net_yiyuehuan_article_details_52951574]

[https://blog.csdn.net/love1055259415/article/details/81095408][https_blog.csdn.net_love1055259415_article_details_81095408]

[https://blog.csdn.net/nullnull2011/article/details/75000071][https_blog.csdn.net_nullnull2011_article_details_75000071]

其中

[https://raw.githubusercontent.com/opencv/opencv\_3rdparty/2a19d0006415955c79431116e4634f04d5eb5a74/ffmpeg/opencv\_ffmpeg.dll][https_raw.githubusercontent.com_opencv_opencv_3rdparty_2a19d0006415955c79431116e4634f04d5eb5a74_ffmpeg_opencv_ffmpeg.dll] 成功

[https://raw.githubusercontent.com/opencv/opencv\_3rdparty/2a19d0006415955c79431116e4634f04d5eb5a74/ffmpeg/opencv\_ffmpeg\_64.dll][https_raw.githubusercontent.com_opencv_opencv_3rdparty_2a19d0006415955c79431116e4634f04d5eb5a74_ffmpeg_opencv_ffmpeg_64.dll] 成功

[https://raw.githubusercontent.com/opencv/opencv\_3rdparty/2a19d0006415955c79431116e4634f04d5eb5a74/ffmpeg/ffmpeg\_version.cmake][https_raw.githubusercontent.com_opencv_opencv_3rdparty_2a19d0006415955c79431116e4634f04d5eb5a74_ffmpeg_ffmpeg_version.cmake] 未成功，不存在

### 9.安装心得： ###

1.版本明确

一：要搞清楚你的是什么版本的opencv，然后下载对应的opencv\_ffmpeg\_64.dll，ffmpeg\_version.cmake，opencv\_ffmpeg.dll

二：要明白上述三个文件和ffmpeg.cmake的关系，在ffmpeg.cmake中的，，是下载地址，可参考  进行下载，但是这里没有找到ffmpeg\_version.cmake的相应链接（试了几个都不是）

2.点击“open project”，在vs2017中显示python问题，（失败：vs编译opencv时error MSB3073  targets(133,5)）可能是因为我的python安装在了中文目录下，后来去掉python相关的，就可以成功了。

3.从opencv官网上下载opencv压缩包形式的文件时，（开始时显示下载速度特别慢，最后突然下载完成）压缩时提示“不可预料的压缩文件末端”，修复未果，最终选择从师父那儿拷贝压缩包过来，下载不成功应该是网络不好导致

4.vs2017上亲测opencv时，出现已加载“C:\\Windows\\SysWOW64\\ntdll.dll”。无法查找或打开 PDB 文件。错误 ，参考链接 [https://www.cnblogs.com/ranson7zop/p/8039676.html][https_www.cnblogs.com_ranson7zop_p_8039676.html] （有图文）（调试-》选项  -》常规+符号），下载过程较慢，2018.7.27日可以正常显示图片了，但是因为加载过程时间太慢，加载一次之后，尝试把以上面链接勾选的项取消掉，亲测仍然可以显示图片，虽然会显示一些项无法加载。

5.Qt上测试opencv是否好用时，出现main.obj:-1: error: LNK2019: 无法解析的外部符号

"void \_\_cdecl cv::imshow(class cv::String const &,class cv::debug\_build\_guard::\_InputArray const &)" (?imshow@cv@@YAXAEBVString@1@AEBV\_InputArray@debug\_build\_guard@1@@Z)，该符号在函数 main 中被引用 的错误

可能原因：

（1）[https://blog.csdn.net/lzm18064126848/article/details/50532204][https_blog.csdn.net_lzm18064126848_article_details_50532204]中说可能是函数有定义，没有实现的缘故，这个对于我的不管用。还可以参考[https://blog.csdn.net/shenziheng1/article/details/54588457][https_blog.csdn.net_shenziheng1_article_details_54588457]

（2）我的：（debug和release模式切换一下就可以了）

将debug模式改为release模式，这是由引进的库有关的，如

-lopencv_core341   默认的是release模式

-lopencv_core341d  是debug模式

（qt中读取图片的路径最好不要有中文，路径中斜杠方向为“/”和“\\\\”均可）

debug和release的区别：[https://zhidao.baidu.com/question/1113218618582190379.html][https_zhidao.baidu.com_question_1113218618582190379.html]

6.CMake编译时出现“error in configuration process project files may be invalid”

参考链接：

[https://blog.csdn.net/ganwenbo2011/article/details/79096984][https_blog.csdn.net_ganwenbo2011_article_details_79096984]   源文件路径决定

[https://blog.csdn.net/c2a2o2/article/details/77921976][https_blog.csdn.net_c2a2o2_article_details_77921976]

[https://blog.csdn.net/sduvec/article/details/50084963][https_blog.csdn.net_sduvec_article_details_50084963]

[https://www.cnblogs.com/spoorer/p/6970401.html][https_www.cnblogs.com_spoorer_p_6970401.html] 是由于默认了之前的编译器

7.运行qt出现d:\\Program Files (x86)\\SogouInput\\...错误。

未解决。。。

之前保存的关于opencv的相关有用链接：

win7下配置opencv的Qt开发环境hy：[https://blog.csdn.net/qiurisuixiang/article/details/8665278][https_blog.csdn.net_qiurisuixiang_article_details_8665278]

opencv+vs开发环境配置hy：[https://blog.csdn.net/poem\_qianmo/article/details/19809337/][https_blog.csdn.net_poem_qianmo_article_details_19809337]

由于surf检测和sift检测等要用到xfeature2d等扩展模块，故在最开始的时候也要下载opencv\_contrib-3.4.1.tar.gz，编译过程一样

下载：[https://github.com/opencv/opencv\_contrib/releases][https_github.com_opencv_opencv_contrib_releases]

【opencv\_contrib-3.4.1压缩包：链接：https://pan.baidu.com/s/1cczu-TMxdl0OzqAfBDOp2Q   提取码：h20k 】

说明参考[https://blog.csdn.net/lixuminglxm/article/details/81486528][https_blog.csdn.net_lixuminglxm_article_details_81486528]

[https_www.cnblogs.com_lzhu_p_8198654.html]: https://www.cnblogs.com/lzhu/p/8198654.html
[https_blog.csdn.net_lsfreeing_article_details_73064954]: https://blog.csdn.net/lsfreeing/article/details/73064954
[http_download.qt.io_archive_qt]: http://download.qt.io/archive/qt/
[http_www.qter.org]: http://www.qter.org/
[https_blog.csdn.net_yanhuangzhinu0_article_details_50957717]: https://blog.csdn.net/yanhuangzhinu0/article/details/50957717
[http_blog.163.com_qimo601_126_blog_static_1582209320162102559368]: http://blog.163.com/qimo601@126/blog/static/1582209320162102559368/
[70]: https://img-blog.csdn.net/20180825105242949?watermark/2/text/aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2p1bGlhcmp1bGlhcg==/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70
[70 1]: https://img-blog.csdn.net/20180825105305119?watermark/2/text/aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2p1bGlhcmp1bGlhcg==/font/5a6L5L2T/fontsize/400/fill/I0JBQkFCMA==/dissolve/70
[70 2]: /images/20220515/d015063d553d4719ab8e246eeaa51b72.png
[https_jingyan.baidu.com_article_a948d6512f00d70a2dcd2edc.html]: https://jingyan.baidu.com/article/a948d6512f00d70a2dcd2edc.html
[https_jingyan.baidu.com_article_b2c186c83eb046c46ff6ff68.html]: https://jingyan.baidu.com/article/b2c186c83eb046c46ff6ff68.html
[https_opencv.org_opencv-3-4.html]: https://opencv.org/opencv-3-4.html
[https_blog.csdn.net_quxiaoxia1986_article_details_52352785]: https://blog.csdn.net/quxiaoxia1986/article/details/52352785
[https_cmake.org_download]: https://cmake.org/download/
[70 3]: /images/20220515/9a87950c98a845efad5204f5fc530c4e.png
[https_blog.csdn.net_qiurisuixiang_article_details_8665278]: https://blog.csdn.net/qiurisuixiang/article/details/8665278
[https_blog.csdn.net_ganwenbo2011_article_details_79096984]: https://blog.csdn.net/ganwenbo2011/article/details/79096984
[https_www.cnblogs.com_lifeofershisui_p_7445062.html]: https://www.cnblogs.com/lifeofershisui/p/7445062.html
[https_blog.csdn.net_sinat_36264666_article_details_73135823_ref_myread]: https://blog.csdn.net/sinat_36264666/article/details/73135823?ref=myread
[70 4]: /images/20220515/007498af5adf4565aa4aa7f5e85e48a4.png
[70 5]: /images/20220515/d55ec844e7364cdc8ad60f5b0b6c1ad2.png
[70 6]: /images/20220515/112f4555d13e4ea0a698a3da1aed021f.png
[70 7]: /images/20220515/daeac168d70742e8b684edbbc244135c.png
[70 8]: /images/20220515/451b73e95e1c4e7eb2ef4eaa9f604b29.png
[70 9]: /images/20220515/62127c540bed49bbbb10d1a2ca04ff11.png
[70 10]: /images/20220515/ebfa9ee0bd384e20bb796cbba503ce15.png
[70 11]: /images/20220515/86a257ac90564bb192d4f1d73a8a7571.png
[70 12]: /images/20220515/8bd67037d75447d4aab78cce6e56d0c7.png
[70 13]: /images/20220515/91050892c420498fa4f1c9c239d2a966.png
[https_blog.csdn.net_u010900606_article_details_79215756]: https://blog.csdn.net/u010900606/article/details/79215756
[https_blog.csdn.net_soft_fa_article_details_19488735]: https://blog.csdn.net/soft_fa/article/details/19488735
[https_blog.csdn.net_sinat_29288207_article_details_56014812_locationNum_2_fps_1]: https://blog.csdn.net/sinat_29288207/article/details/56014812?locationNum=2&fps=1
[https_www.cnblogs.com_zhanggaofeng_p_6671084.html]: https://www.cnblogs.com/zhanggaofeng/p/6671084.html
[https_blog.csdn.net_debugconsole_article_details_8230933]: https://blog.csdn.net/debugconsole/article/details/8230933
[https_blog.csdn.net_qq_38880380_article_details_78013056]: https://blog.csdn.net/qq_38880380/article/details/78013056
[https_www.2cto.com_kf_201707_661163.html]: https://www.2cto.com/kf/201707/661163.html
[https_github.com_opencv_opencv]: https://github.com/opencv/opencv
[https_blog.csdn.net_yiyuehuan_article_details_52951574]: https://blog.csdn.net/yiyuehuan/article/details/52951574
[https_blog.csdn.net_love1055259415_article_details_81095408]: https://blog.csdn.net/love1055259415/article/details/81095408
[https_blog.csdn.net_nullnull2011_article_details_75000071]: https://blog.csdn.net/nullnull2011/article/details/75000071
[https_raw.githubusercontent.com_opencv_opencv_3rdparty_2a19d0006415955c79431116e4634f04d5eb5a74_ffmpeg_opencv_ffmpeg.dll]: https://raw.githubusercontent.com/opencv/opencv_3rdparty/2a19d0006415955c79431116e4634f04d5eb5a74/ffmpeg/opencv_ffmpeg.dll
[https_raw.githubusercontent.com_opencv_opencv_3rdparty_2a19d0006415955c79431116e4634f04d5eb5a74_ffmpeg_opencv_ffmpeg_64.dll]: https://raw.githubusercontent.com/opencv/opencv_3rdparty/2a19d0006415955c79431116e4634f04d5eb5a74/ffmpeg/opencv_ffmpeg_64.dll
[https_raw.githubusercontent.com_opencv_opencv_3rdparty_2a19d0006415955c79431116e4634f04d5eb5a74_ffmpeg_ffmpeg_version.cmake]: https://raw.githubusercontent.com/opencv/opencv_3rdparty/2a19d0006415955c79431116e4634f04d5eb5a74/ffmpeg/ffmpeg_version.cmake
[https_www.cnblogs.com_ranson7zop_p_8039676.html]: https://www.cnblogs.com/ranson7zop/p/8039676.html
[https_blog.csdn.net_lzm18064126848_article_details_50532204]: https://blog.csdn.net/lzm18064126848/article/details/50532204
[https_blog.csdn.net_shenziheng1_article_details_54588457]: https://blog.csdn.net/shenziheng1/article/details/54588457
[https_zhidao.baidu.com_question_1113218618582190379.html]: https://zhidao.baidu.com/question/1113218618582190379.html
[https_blog.csdn.net_c2a2o2_article_details_77921976]: https://blog.csdn.net/c2a2o2/article/details/77921976
[https_blog.csdn.net_sduvec_article_details_50084963]: https://blog.csdn.net/sduvec/article/details/50084963
[https_www.cnblogs.com_spoorer_p_6970401.html]: https://www.cnblogs.com/spoorer/p/6970401.html
[https_blog.csdn.net_poem_qianmo_article_details_19809337]: https://blog.csdn.net/poem_qianmo/article/details/19809337/
[https_github.com_opencv_opencv_contrib_releases]: https://github.com/opencv/opencv_contrib/releases
[https_blog.csdn.net_lixuminglxm_article_details_81486528]: https://blog.csdn.net/lixuminglxm/article/details/81486528