媒体基金会捕获反转图像

Media Foundation capture reversed image

这有点奇怪,我正在使用 SourceReader 从我的相机中捕捉样本。我最终得到了一个颠倒的 HBITMAP。

捕获代码如下:

for (;;)
{
    DWORD streamIndex = 0, flags = 0;
    LONGLONG llTimeStamp = 0;
    CComPtr<IMFSample> pSample;

    hr = sr->ReadSample(MF_SOURCE_READER_FIRST_VIDEO_STREAM,0,&streamIndex,&flags,&llTimeStamp,&pSample);
...
}

我得到一个 IMFSample,我正在尝试绘制它(目前不是使用 Direct3D,而是使用 Direct2D)。捕获的格式是 MFVideoFormat_NV12,我想将它转换为 MFVideoFormat_RGB32,这样我就可以从中创建一个 HBITMAP。因此,我创建了一个 IMFTransform,其源类型设置为从捕获返回的媒体类型,并以 RGB 类型为目标:

if (prvtrs)
{
    DWORD is = 0, os = 0;
    hr = prvtrs->GetStreamCount(&is, &os);
    if (is > 0 && os > 0)
    {
        hr = prvtrs->SetInputType(iids[0], fmt, 0);
        if (SUCCEEDED(hr))
        {
            PROPVARIANT pv;
            InitPropVariantFromCLSID(MFVideoFormat_RGB32, &pv);
            hr = prmt->SetItem(MF_MT_SUBTYPE, pv);
            if (SUCCEEDED(hr))
            {
                hr = prvtrs->SetOutputType(oods[0], prmt, 0);
                if (SUCCEEDED(hr))
                {
// OK
                }
            }
        }
    }
}

捕获后,我转换 IMFSample:

CComPtr<IMFSample> pSample2;
MFCreateSample(&pSample2);

if (si.cbSize == 0)
    prvtrs->GetOutputStreamInfo(oods[0], &si);

CComPtr<IMFMediaBuffer> bb;
MFCreateMemoryBuffer(si.cbSize, &bb);
pSample2->AddBuffer(bb);


hr = prvtrs->ProcessInput(iids[0], pSample, 0);
MFT_OUTPUT_DATA_BUFFER db = { 0 };
db.dwStreamID = oods[0];
db.pSample = pSample2;
DWORD st;
hr = prvtrs->ProcessOutput(0, 1, &db, &st);
if (db.pEvents)
    db.pEvents->Release();
if (SUCCEEDED(hr))
{
    // Show it
    CComPtr<IMFMediaBuffer> b4;
    pSample2->ConvertToContiguousBuffer(&b4);
    if (b4)
    {
        vector<char> o;
        auto bi = SaveSample(b4, wi, he);
....

这是我的 SaveSample:

HBITMAP SaveSample(CComPtr<IMFMediaBuffer> mediaBuffer, int width32, int height32)
{
    HRESULT hr = 0;
    BYTE *pData = NULL;
    DWORD le = 0;
    mediaBuffer->GetCurrentLength(&le);
    hr = mediaBuffer->Lock(&pData, NULL, NULL);
    if (!pData)
        return 0;
    unsigned char*  pixels = pData;
    // at this point we have some input

    BITMAPINFOHEADER bmih;
    bmih.biSize = sizeof(BITMAPINFOHEADER);
    bmih.biWidth = width32;
    bmih.biHeight = height32;
    bmih.biPlanes = 1;
    bmih.biBitCount = 32;
    bmih.biCompression = BI_RGB;
    bmih.biSizeImage = 0;
    bmih.biXPelsPerMeter = 0;
    bmih.biYPelsPerMeter = 0;
    bmih.biClrUsed = 0;
    bmih.biClrImportant = 0;

    BITMAPINFO dbmi = { 0 };
    dbmi.bmiHeader = bmih;
    dbmi.bmiColors->rgbBlue = 0;
    dbmi.bmiColors->rgbGreen = 0;
    dbmi.bmiColors->rgbRed = 0;
    dbmi.bmiColors->rgbReserved = 0;

    HDC hdc = ::GetDC(NULL);
    HBITMAP hbmp = CreateDIBitmap(hdc, &bmih, CBM_INIT, pixels, &dbmi, DIB_RGB_COLORS);
    ::ReleaseDC(NULL, hdc);

    mediaBuffer->Unlock();
    return hbmp;
}

我得到的是倒置的位图。我做错了什么?

您错过了 specification of BMP files 的关键部分:

The pixel array is a block of 32-bit DWORDs, that describes the image pixel by pixel. Usually pixels are stored "bottom-up", starting in the lower left corner, going from left to right, and then row by row from the bottom to the top of the image.[4] Unless BITMAPCOREHEADER is used, uncompressed Windows bitmaps also can be stored from the top to bottom, when the Image Height value is negative.

考虑到这一点,您可以轻松修复错误。