简体   繁体   中英

Using cv::Mat1f as cv::Mat

I have this method:

static void WriteMatVect(const std::string& filename, const std::vector<cv::Mat>& mats);

...

void FileSystem::WriteMatVect(const std::string& filename, const std::vector<cv::Mat>& mats){
    size_t size = mats.size();
    FileSystem::saveArray(&size,1,filename);
    if(mats.empty()){
        std::cerr<<"error WriteMatVect: no cv::Mat in mats"<<std::endl;
        return;
    }
    for(size_t i=0 ; i<mats.size() ; i++)
        FileSystem::WriteMat(filename, mats[i], true);
}

Which is called passing a std::vector<cv::Mat1f> as mats . But this returns the following error:

../FileSystem.hpp:28:14: note:   no known conversion for argument 2 from ‘std::vector<cv::Mat_<float> >’ to ‘const std::vector<cv::Mat>&’

A simple workaround could be changing WriteMatVect signature using std::vector<cv::Mat1f>& mats , but this would make WriteMatVect too strict (it would work only with float matrices), while I would like to do it as general as generic as possible. The only solution that comes to my mind is using templates, so const std::vector<T> &mats . Any other solution?

The problem: I think you are mixing derived with base class:

(see the compiler error)

no known conversion for argument 2 from std::vector<cv::Mat_<float> >
to const std::vector<cv::Mat>&

template<typename _Tp> class Mat_ : public Mat
{
public:
    // ... some specific methods
    //         and
    // no new extra fields
};

Solution 1: The template class Mat_ is derived from Mat , so maybe you can change your code from vector<cv::Mat> to vector<cv::Mat*> and the argument from std::vector<cv::Mat1f> to std::vector<cv::Mat*> , the downcasting will be implicit when you push the matrix to the vectors.

Then your method will be:

static void WriteMatVect(const std::string& filename, const std::vector<cv::Mat*>& mats);

[EDITED] Solution 2: Another possible workaround (if you prefer not to change the calling method) is to slice the objects in the std::vector<cv::Mat1f> mats , so use a std::vector<cv::Mat> mats instead, for example when you push the objects:

cv::Mat1f matFelement;
std::vector<cv::Mat> mats
mats.push_back(matFelement); //object sliced form Mat1f to Mat

A Mat1f is convertible to a Mat , but a vector<Mat1f> is not convertible to a vector<Mat> .

A simple workaround is to copy your vector<Mat1f> to the correct vector<Mat> . Remember that data are not copied, so it shouldn't be that slow.

vector<Mat1f> v;
...
vector<Mat> u;
u.reserve(v.size());
for(const auto& m : v) { u.push_back(m); }

WriteMatVect(u);

The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM