opencv编译以及形态学快速的原因

最新推荐文章于 2022-12-02 22:06:20 发布

原创最新推荐文章于 2022-12-02 22:06:20 发布 · 325 阅读

0 ·

CC 4.0 BY-SA版权

工作记录专栏收录该内容

5 篇文章

订阅专栏

本文详细介绍了如何从GitHub下载并编译OpenCV，以及在项目中正确配置OpenCV库的过程。通过对比自写算法与OpenCV的性能，发现OpenCV的ipp加速和OpenCL支持显著提升效率。文章还分享了配置过程中遇到的问题及解决方案，如动态库的拷贝、头文件和静态库的路径设置，以及在不同编译模式下如何引用库。

自己用线程写出的膨胀算法竟然比opencv慢，为此查看opencv源码

安装

直接搜opencv，官网

到github下载版本多但慢===很慢对应网站下载编译

https://github.com/opencv/opencv/releases

https://blog.youkuaiyun.com/moxiaonuo/article/details/77823990

build /install 中的东西有用

配置

头文件包含目录install 下的include
添加附加静态库库目录lib
把动态库拷贝到输出目录下面
加上下面到话不需要配置输入库名

#include "stdafx.h"
#include "opencv2\opencv.hpp"
#include <vector>
#ifdef _DEBUG
#pragma comment(lib, "opencv_calib3d330d.lib")
#pragma comment(lib, "opencv_core330d.lib")
#pragma comment(lib, "opencv_dnn330d.lib")
#pragma comment(lib, "opencv_features2d330d.lib")
#pragma comment(lib, "opencv_flann330d.lib")
#pragma comment(lib, "opencv_highgui330d.lib")
#pragma comment(lib, "opencv_imgcodecs330d.lib")
#pragma comment(lib, "opencv_imgproc330d.lib")
#pragma comment(lib, "opencv_ml330d.lib")
#pragma comment(lib, "opencv_objdetect330d.lib")
#pragma comment(lib, "opencv_photo330d.lib")
#pragma comment(lib, "opencv_shape330d.lib")
#pragma comment(lib, "opencv_stitching330d.lib")
#pragma comment(lib, "opencv_superres330d.lib")
#pragma comment(lib, "opencv_ts330d.lib")
#pragma comment(lib, "opencv_video330d.lib")
#pragma comment(lib, "opencv_videoio330d.lib")
#else
#pragma comment(lib, "opencv_calib3d330.lib")
#pragma comment(lib, "opencv_core330.lib")
#pragma comment(lib, "opencv_dnn330.lib")
#pragma comment(lib, "opencv_features2d330.lib")
#pragma comment(lib, "opencv_flann330.lib")
#pragma comment(lib, "opencv_highgui330.lib")
#pragma comment(lib, "opencv_imgcodecs330.lib")
#pragma comment(lib, "opencv_imgproc330.lib")
#pragma comment(lib, "opencv_ml330.lib")
#pragma comment(lib, "opencv_objdetect330.lib")
#pragma comment(lib, "opencv_photo330.lib")
#pragma comment(lib, "opencv_shape330.lib")
#pragma comment(lib, "opencv_stitching330.lib")
#pragma comment(lib, "opencv_superres330.lib")
#pragma comment(lib, "opencv_ts330.lib")
#pragma comment(lib, "opencv_video330.lib")
#pragma comment(lib, "opencv_videoio330.lib")
#endif

发现

opencv有 ipp加速

有专门的useOpenCL的函数库

核心函数为：

IppStatus ippStatus = ::iwiFilterMorphology(pSrcImage, pDstImage, morphType, pMaskImage, pAnchor, border, border, pRoi);

此代码无法溯源

后记

发现gpuz显示不支持opencl

重装显卡驱动后支持显卡驱动