imquery() and overloaded imread() functions

[sturkmen72]

proposed solution to #6574 #17862

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

• I agree to contribute to the project under OpenCV (BSD) License.
• To the best of my knowledge, the proposed patch is not based on a code under GPL or other license that is incompatible with OpenCV
• The PR is proposed to proper branch
• There is reference to original bug report and related work
• There is accuracy test, performance test and test data in opencv_extra repository, if applicable
  Patch to opencv_extra has the same branch name.
• The feature is well documented and sample code can be built with the project CMake

[vpisarev]

@sturkmen72, thank you for the contribution! I think, we should discuss the API. Can you provide a little code snippet how people would use this new functionality?

In C++ one of the best approaches would be to create an image iterator. For Python it could be possible to use a similar approach, even though it would not look as idiomatic. Something like that:

class ImageCollection
{
public:
     ...
     CV_WRAP size_t nimages() const;
     CV_WRAP ImageIterator begin() const;
     ImageIterator end() const;
     ...

     CV_WRAP static ImageCollection fromMultiPageImage(const std::string& img);
     CV_WRAP static ImageCollection fromDirectory(const std::string& dir);
}; 

class ImageIterator
{
public:
   Mat operator *() const;
   CV_WRAP Mat read() const; // operator * for Python/Java
   CV_WRAP size_t currentIndex() const;
   CV_WRAP bool next(); // operator ++() for Python/Java
   ImageIterator& operator ++();
   ImageIterator operator ++(int);
protected:
   Ptr<ImageIteratorImpl> pimpl;
}

// C+++
for(cv::Mat img: ImageCollection::fromMultiPageImage("big_tiff_file.tif")) {
    ...
}

# Python
image_col = cv2.ImageCollection.fromMultiPageImage("big_tiff_file.tif")
nimg = image_col.nimages()
it = image_col.begin()
i  = 0
while i < nimg:
   img = it.read()
   # process img ...
   it.next()
   i += 1

of course, it would also be possible to create more Pythonic wrapper on top of it, like suggested here, for example:
https://treyhunner.com/2018/06/how-to-make-an-iterator-in-python/

[sturkmen72]

@vpisarev thank you for your feedback. let me work on the PR a bit more

[asmorkalov]

@sturkmen72 Do you have a chance to finish the PR in meantime?

[sturkmen72]

@asmorkalov , @vpisarev i am still thinking of an effective usage way considering #8511

[sturkmen72]

A simple review is required before continuing. please forgive my sloppiness i did not look codes of #8511 before. now i am aware both PR has basely similar changes.

@brian-armstrong-discord do you have time to look my changes

[sturkmen72]

@vpisarev, @alalek please take a look at this PR. I know it still need more work but i need to know if i am on the right path.

sample usage

C++

    ImageLoader imageloader;

    if( imageloader.open(filename, flags) )
        for (int i = 0; i < imageloader.getNumPages(); i++)
        {
            Mat page;
            if( imageloader.load(page) )
            {
                 // do something 
            }
            imageloader.nextPage();
        }

Python

import cv2

il = cv2.ImageLoader()
il.open('D:/Github/opencv_extra/testdata/highgui/readwrite/multipage.tif', ,cv2.IMREAD_REDUCED_GRAYSCALE_2)

for i in range(il.getNumPages()):
    ret,img = il.load()
    cv2.imshow("pages", img)
    cv2.waitKey()
    il.nextPage()

cv2.destroyAllWindows()

[vpisarev]

@sturkmen72, thank you for keep working on it. Could you, please, provide some sample code and the test for the functionality?

[sturkmen72]

@vpisarev, @alalek i created a repo for testing this PR. still i have some issues. is it possible to make a short online meeting about this PR.

note : indentations ignored for now to facilitate review. also there is many lines can be deleted if this PR will be mature.

[vpisarev]

@sturkmen72, yes, let's discuss it. I sent you e-mail

[vpisarev]

you can try the following:

ret,roi = cv.imread("images.tif", None, cv.IMREAD_REDUCED_COLOR_2, i)
img1[row:row+page_info3.height, column:column+page_info3.width] = roi

[sturkmen72]

mats shoud not be empty before calling this function

[sturkmen72]

If we asked for more images than multipage page count and we successfully read existing images should this be false ?

[sturkmen72]

i think GDAL driver loads only first page of multipage images. i will try to implement GdalDecoder::nextPage() after actual changes seems good.

[sturkmen72]

@vpisarev I think I've made significant progress after our meeting. could you review the actual code.