编译 opencv 代码时遇到错误

Facing error with compiling opencv code

本文关键字:遇到 错误 代码 opencv 编译      更新时间:2023-10-16

我正在使用opencv 2.4,下面是我正在尝试编译的代码。 我使用此命令编译我的代码

g++ -o "match" -ggdb `pkg-config --cflags opencv` match.cpp `pkg-config --libs opencv` 

为什么我收到此错误:

match.cpp: In function ‘int main(int, const char**)’:
match.cpp:18:37: error: expected type-specifier before ‘SurfFeatureDetector’
match.cpp:18:37: error: conversion from ‘int*’ to non-scalar type ‘cv::Ptr<cv::FeatureDetector>’ requested
match.cpp:18:37: error: expected ‘,’ or ‘;’ before ‘SurfFeatureDetector’
match.cpp:22:2: error: ‘SurfDescriptorExtractor’ was not declared in this scope
match.cpp:22:26: error: expected ‘;’ before ‘extractor’
match.cpp:26:2: error: ‘extractor’ was not declared in this scope
match.cpp:29:2: error: ‘BruteForceMatcher’ was not declared in this scope
match.cpp:29:30: error: expected primary-expression before ‘>’ token
match.cpp:29:32: error: ‘matcher’ was not declared in this scope

我认为我使用的 opencv 版本存在一些问题,因为相同的代码在 2.2 版本上运行良好,但我不确定它是什么。帮助!!

#include <opencv/cv.h>
#include <opencv/highgui.h>
#include <string.h>
#include <iostream>
using namespace std;
using namespace cv;

int main(int argc, const char* argv[])
{
    cout << argv[1] << endl << argv[2] << endl;
    Mat img1 = imread(argv[1] , CV_LOAD_IMAGE_GRAYSCALE );
    Mat img2 = imread(argv[2] , CV_LOAD_IMAGE_GRAYSCALE );
    vector<KeyPoint> keypoints1;
    vector<KeyPoint> keypoints2;
    Ptr<FeatureDetector> feature = new SurfFeatureDetector(2500);
    feature->detect(img1,keypoints1);
    feature->detect(img2,keypoints2);
    SurfDescriptorExtractor extractor;
    Mat desc1 , desc2;
    extractor.compute(img1,keypoints1,desc1);
    extractor.compute(img2,keypoints2,desc2);
    BruteForceMatcher<L2<float> > matcher;
    vector<vector<DMatch> > matches1;
    vector<vector<DMatch> > matches2;
    vector<DMatch> symMatches;
    vector<DMatch> outMatches;
    matcher.knnMatch(desc1,desc2,matches1,2);
    matcher.knnMatch(desc2,desc1,matches2,2);
    int count_inliers = 0 , count_matches = 0;
    for(vector<vector<DMatch> >::const_iterator matIt1 = matches1.begin(); matIt1 != matches1.end(); ++matIt1){
        count_matches++;
        if(matIt1->size() < 2)
            continue;
        for(vector<vector<DMatch> >::const_iterator matIt2 = matches2.begin(); matIt2 != matches2.end(); ++matIt2){
            if(matIt2->size() < 2)
                continue;
            if((*matIt1)[0].queryIdx == (*matIt2)[0].trainIdx && (*matIt2)[0].queryIdx == (*matIt1)[0].trainIdx){
                count_inliers++;
                symMatches.push_back(DMatch((*matIt1)[0].queryIdx,(*matIt1)[0].trainIdx,(*matIt1)[0].distance));
                break;
            }
        }
    }
    vector<Point2f> points1, points2;
    for(vector<DMatch>::const_iterator it = symMatches.begin(); it!=symMatches.end(); ++it){
        float x = keypoints1[it->queryIdx].pt.x;
        float y = keypoints1[it->queryIdx].pt.y;
        points1.push_back(Point2f(x,y));
        x = keypoints2[it->trainIdx].pt.x;
        y = keypoints2[it->trainIdx].pt.y;
        points2.push_back(Point2f(x,y));
    }
    vector<uchar> inliers(points1.size(),0);
    Mat fundamental;
    fundamental = findFundamentalMat(Mat(points2),Mat(points1),inliers,CV_FM_RANSAC,2,0.8);
    vector<uchar>::const_iterator itIn = inliers.begin();
    vector<DMatch>::const_iterator itM = symMatches.begin();
    for(;itIn!=inliers.end();++itIn,++itM){
        if(*itIn){
            outMatches.push_back(*itM);
        }
    }
    cout << count_inliers << endl;
    cout << count_matches << endl;
    cout << (float) count_inliers/(float) count_matches << endl;
    float diff = (float) count_inliers/(float) count_matches;
//  if(diff > 0.30){
//      cout << "Similar Images " << endl << "-----------------" << endl;
//      exit(1);
//  }
//  vector<uchar> inliers(points1.size(),0);
    Mat homography = findHomography(Mat(points2),Mat(points1),inliers,CV_RANSAC,1);
    vector<Point2f>::const_iterator itPts = points1.begin();
//  vector<uchar>::const_iterator itIn = inliers.begin();
/*  while(itPts != points1.end()){
        if(*itIn)
            circle(img1,*itPts,3,Scalar(255,255,255),2);
        ++itPts;
        ++itIn;
    }
    itPts = points2.begin();
    itIn = inliers.begin();
    while(itPts != points2.end()){
        if(*itIn)
            circle(img2,*itPts,3,Scalar(255,255,255),2);
        ++itPts;
        ++itIn;
    }
*/
    Mat result;
    warpPerspective(img2,result,homography,Size(2*img2.cols,img2.rows));
    Mat half(result,Rect(0,0,img1.cols,img1.rows));
    img1.copyTo(half);


    // Add results to image and save.
    char name[1000];
//    strcpy(name,"./surf/surf");
//    strcat(name,argv[1]);
    cv::Mat output1;
    cv::Mat output2;
    cv::drawKeypoints(img1, keypoints1, output1);
    cv::drawKeypoints(img2, keypoints2, output2);
    cv::imwrite("./surf/img11.png", img1);
    cv::imwrite("./surf/img21.png", img2);
    cv::imwrite("./surf/img31.png", result);
    cv::imwrite("./surf/tt.png", result);
    cv::imwrite("./surf/img41.png", half);
    cv::imwrite("./surf/img51.png", output1);
    cv::imwrite("./surf/img61.png", output2);
    return 0;
}

BruteForceMatcher现在称为

cv::BFMatcher

请参阅文档。

您可以像这样定义匹配器:

DescriptorMatcher* hammingMatcher = new BFMatcher(NORM_HAMMING,false);
//or
DescriptorMatcher* hammingMatcher = new BFMatcher(NORM_L2,false);

编辑

此外,在此示例代码中,您可以通过包含标头来了解如何使用旧版本的匹配器

#include "hammingseg.h"

阅读有关 SURF 和 SIFT 探测器的讨论 - 它们已被移动为非自由。

还添加为动态库以链接libopencv_nonfree.solibopencv_features2d.so

对于BruteForceMatcher来说,它看起来仍然是一个悬而未决的问题,但我很确定在其中一个.so内,我希望他们也确实改变了标题。如果您发现有关BruteForceMatcher的内容,我将不胜感激。

包括以下内容

#"opencv2/features2d/features2d.hpp"
#"opencv2/highgui/highgui.hpp"
#"opencv2/core/core.hpp"
#"opencv2/legacy/legacy.hpp" (BruteForceMatcher is defined here!)

并链接到以下内容

#opencv_core.so
#opencv_flann.soo (if youre using the FLANN matcher)
#opencv_highgui.so
#opencv_features2d.so
#opencv_nonfree.so

似乎为我做了这个把戏。希望这有帮助。

看起来你需要包含SurfFeatureDetector的头文件

这是 API,他们在其中提到了以下内容:

#include <features2d.hpp>

不要将文件包含整个路径,而应仅包含名称,并让编译器参数 (-I) 指定其路径。这更便携。

如果它没有在那里定义,那么寻找它。 在 Linux 中,您可以执行以下操作:

# find . -name "*.h*" | xargs grep SurfFeatureDetector | grep class
# find . -name "*.h*" | xargs grep BruteForceMatcher | grep class

这应该得到所有的 *.h 和 *.hpp 文件以及 SurfFeatureDetector 的 grep,对于这些结果,grep 用于类。

它非常简单:

//matching descriptors
cv::BFMatcher matcher(cv::NORM_L2, true);
std::vector<cv::DMatch> matches;
matcher.match(descriptor1, descriptor2, matches);

如果标志设置为 true,您已经进行了交叉检查。

#include <opencv2/legacy/legacy.hpp>

添加此行,它将像以前一样工作。