s093294's profile - activity

Is there anyway to extract dense sift/surf features with opencv.

With dense I mean, that in a regular grid of some size, compute features at each vertex.

What is the point of

    Point2f p = features1.keypoints[m.queryIdx].pt;
    p.x -= features1.img_size.width * 0.5f;
    p.y -= features1.img_size.height * 0.5f;
    src_points.at<Point2f>(0, static_cast<int>(i)) = p;

    p = features2.keypoints[m.trainIdx].pt;
    p.x -= features2.img_size.width * 0.5f;
    p.y -= features2.img_size.height * 0.5f;
    dst_points.at<Point2f>(0, static_cast<int>(i)) = p;

in BestOf2NearestMatcher ?

the surf homography example dont seem to do this, but simply uses the points as is.

The reason why i ask is that I see a small ofset (permanent) in warped images if I use the BestOf2NearestMatcher to estimate H, but not if just using the translated points.

I made the following test and was able to verify that the point translations are causing my issues.

    for (size_t i = 0; i < good_matches.size(); i++)
    {
        //-- Get the keypoints from the good matches
        obj.push_back(features1.keypoints[good_matches[i].queryIdx].pt);
        scene.push_back(features2.keypoints[good_matches[i].trainIdx].pt);
        const cv::DMatch& m = good_matches[i];

        cv::Point2f p = features1.keypoints[m.queryIdx].pt;
        p.x -= features1.img_size.width * 0.5f;
        p.y -= features1.img_size.height * 0.5f;
        src_points.at<cv::Point2f>(0, static_cast<int>(i)) = p;

        p = features2.keypoints[m.trainIdx].pt;
        p.x -= features2.img_size.width * 0.5f;
        p.y -= features2.img_size.height * 0.5f;
        dst_points.at<cv::Point2f>(0, static_cast<int>(i)) = p;
    }


    info.H = findHomography(src_points, dst_points, cv::RANSAC); //this is wronge
    info.H = findHomography(obj, scene, cv::RANSAC); //this is correct

I would like to take benefit of the BestOf2NearestMatcher instead of manually finding best matches.

Any chance that you have some code that can read the IIQ file format to opencv cv::Mat for model IQ180 ?

Thanks anyhow. Just abit anoyed with it. We are using view NX from nikon right now which is an awesome tool but I need to automate this.

Thats not the problem. I already tried that and also confirmed that a Red car is actual red. I found out how to tell libraw to use the camera WB, and the image is more truth to its colors, but still to light

the difference from teh two images can be seen here. https://04o62a.dm1.livefilestore.com/y2p7TPrRkirmyRkqZWfcwKZCK0Htr9OwJMeXop9NfXsw7xgzbI9qaxk-r-pVoOlcc6VBrY1daOhldAXLJMFg43mMuYQ7x2ilAQm8niKFJPNHSw/libraw_d800e.png?psid=1

I just tried libraw outside opencv environement and opened the ppm image it generated from the nef. It have done something to the colors. Will try your link and see if the same results will be present there. If so it must be something in libraw that is manipulating the image.

http://lclevy.free.fr/nef/index.html (the problem fits with the specs of the 160x120 extra image)

I tried to open a .nef file in my opencv application. Everything worked but the image was 160x120 pixels instead of the full size. No errors.

What do it take to open raw .nef files and then use them.

I updated my issues at : http://stackoverflow.com/questions/22355491/libraw-is-making-my-images-too-bright-compared-to-nikons-own-converter