OpenCV 逐帧视频拼接

Question

我试图拼接来自两个不同角度设置的不同相机的输入。 我应该怎么做才能将摄像机的输出（每一帧一帧一帧）合并为一帧？

当我将输入作为以前录制的视频（“InputVideo1.mp4”和“InputVideo2.mp4”，如下所述）时，视频输入生成的帧会拼接。 但是，当我启用相机作为输入时，OpenCV Stitcher 返回枚举“ERR_NEED_MORE_IMGS = 1”。

// Create a VideoCapture object and open the input file
//VideoCapture vCap("InputVideo1.mp4");
//VideoCapture vCap1("InputVideo2.mp4");

// If the input is the web camera, pass 0 instead of the video file name
VideoCapture vCap(0);
VideoCapture vCap1(1);

// Check if the camera opened successfully
if (!vCap.isOpened()) {
    cout << "Error opening video stream or file" << endl;
    return -1;
}
if (!vCap1.isOpened()) {
    cout << "Error opening video stream or file" << endl;
    return -2;
}

while (1) {

    Mat frame, frame1, finalFrame;

    // Capture frame-by-frame
    vCap >> frame;
    vCap1 >> frame1;

    // If the frame is empty, break immediately
    if (frame.empty()) {
        cout << "Frame is empty" << endl;
        break;
    }
    if (frame1.empty()) {
        cout << "Frame1 is empty" << endl;
        break;
    }

    // Display the resulting frame
    imshow("First frame", frame);
    imshow("Second frame", frame1);
    
    // Push each frame one by one into the final vector 
    vector<Mat> finalFrameImages;

    finalFrameImages.push_back(frame);
    finalFrameImages.push_back(frame1);


    // Stitch all frames that are stored the final frame images vector set
    Stitcher::Mode mode = Stitcher::PANORAMA;
    Ptr<Stitcher> stitcher = Stitcher::create(mode);
    Stitcher::Status status = stitcher->stitch(finalFrameImages, finalFrame);

    if (status != Stitcher::OK) {
        cout << "Can't stitch images" << endl;
        return status;
    }
    

    imshow("Resulting frame", finalFrame);

}

Stitcher::Status 状态应该返回 0 但它返回 1。

Answer 1

可能是图片一样。 看看这张图： 现在我们看到在配准块中，该方法查看两个图像之间的特征并匹配它们以正确进行拼接。 建议您在拼接之前检查是否导入了好的图像，我看不出其他问题。 除此之外，请查看此链接