zoukankan      html  css  js  c++  java
  • OpenCV学习参考 即时贴

    注意:本博文在github上日常更新(保持GitHub最新)

    https://github.com/SylvesterLi/MyOpenCVCode

    基本安装:https://blog.csdn.net/nicewe/article/details/79173346

    Contribute编译-安装:https://blog.csdn.net/zmdsjtu/article/details/78069739

    注意:我在make的时候被卡住了好半天。因为后台在补全(下载)package,跟网络有关 
    
    实践证明,没有build文件夹直接使用newbuild是可以的。(另外,第二个教程里教你配置环境变量F:OCVopencv
    ewbuildinstallx64vc15in)
    
    

    我的Github链接:https://github.com/SylvesterLi/MyOpenCVCode

    Here is my README.md

    Here is my OpenCV Learning Code

    Well, I am not so professional in this, but I'd like to share my learning experinence.

    maybe someday it goes to work out my questions~

    Update 1

    2018/08/25 Now we come a new stage and I should learn more professional knowledge.

    What I have learnt from

    This tells you how to normally setup

    https://blog.csdn.net/jia20003/article/details/54583431

    and when you got some trouble:Pls check your network(My home network trapped me for a long while)

    Problems such as

    Download face_landmark_model.dat Failed (or time out)
    missing ffmpeg

    but the download is real slow.
    you could use your mobile and provide hotspot for your PC.(As I done)

    if auto-download can't do to help, you could check these issues :

    https://github.com/opencv/opencv_contrib/issues

    At last ,google can be your best driver.

    Update 2

    Finally, got it! After check and check again, it works!

    Below is the code of printing all files' name in current catalog
    Just paste it in the PowerShell
    Attention : the F:OCVopencv ewbuildinstallx64vc15lib is my file directory path

    
    Get-ChildItem F:OCVopencv
    ewbuildinstallx64vc15lib | ForEach-Object -Process{
    if($_ -is [System.IO.FileInfo])
    {
    Write-Host($_.name);
    }
    }
    
    

    Update 3

    In the OCV3 Project I don't use src as default input image but use img_1.Which leads me misleading and make some mistakes. In the last days, I should take care of this!!

    Update 4

    These two days, i just watch corner detection, one of detection methods is Harris , and another is Shi-Tomasi Corner detetion. In my point of view , Good Features To Track (aka Shi-Tomasi) performs better than Harris detection .

    But both of them did not mark the top of roof which human could easily recognize.

    Pic blow is good feature to track

    and till now, I should have finished custom corner detection, but I think it is unnecessary to learn. When I meet such kind of projects or problems, I would come back have a careful seek.

    Update 5

    These two days I have tried SURF and SIFT,both of them are using to detect KeyPoints in the image which is hard for human beings to recognize what it is.

    And the result of Experiments is that there are not so much differences between SURF and SIFT, but you still say, the KeyPoints of image shows their own features, which we can conclude that SIFT seems better?

    Almost forget to say, their sample code looks same.

    
    //SIFT跟SURF代码是一模一样的
    //numOfFeatures指的是特征点的个数
    int numOfFeatures = 400;
    //现在创建检测器
    Ptr<SIFT> detector = SIFT::create(numOfFeatures);
    vector<KeyPoint> keypoints;//存到这来
    //检测 
    detector->detect(src, keypoints);
    Mat kpImage;
    //绘制关键点
    drawKeypoints(src, keypoints, kpImage);
    
    namedWindow("result", WINDOW_AUTOSIZE);
    imshow("result", kpImage);
    
    

    Update 6

    Comming soon....

    These days I am too lazy to update, but from now on. I will keep code updating!!!

  • 相关阅读:
    深入理解泛型之JAVA泛型的继承和实现、泛型擦除
    hadoop过程中遇到的错误与解决方法
    微服务拆分到什么粒度合适——康威定律
    墨菲定律(设计系统)和康威定律(系统划分)
    Hadoop-Impala学习笔记之SQL参考
    Hadoop-Impala学习笔记之管理
    Hadoop2-HDFS学习笔记之入门(不含YARN及MR的调度功能)
    Hadoop-Impala学习笔记之入门
    解决 Invalid character found in method name. HTTP method names must be tokens 异常信息
    从康威定律和技术债看研发之痛
  • 原文地址:https://www.cnblogs.com/-SANG/p/9655749.html
Copyright © 2011-2022 走看看