转换std :: vector <uint8_t> 到QImage

Question

I am trying to create QImage from std::vector. I tried;

void FaceCutThread::convertImage(std::vector<uint8_t> &buf)
{
    QImage img(&buf[0], 300, 300, QImage::Format_ARGB32);

    emit isFinisedFaceCut(img);
}

and

void FaceCutThread::convertImage(std::vector<uint8_t> &buf)
{
    QImage img(buf.data(), 300, 300, QImage::Format_ARGB32);

    emit isFinisedFaceCut(img);
}

both two type of convertion didn't give any error;however, when I tried to show it in QLabel like;

void MainWindow::handleFaceResults(QImage& buf)
{
    QImage myImage(buf);
    ui->lblRawImage->setPixmap(QPixmap::fromImage(myImage));
}

my app always unexpectedly finishing. Could you please help how I can convert to QImage correctly?

EDIT:

I can write image into file without any problem like;

bool save_file (const string &path, const vector<uint8_t> &data)
{
    std::ofstream os;
    os.open(path.c_str(), std::ios::out | std::ios::binary);
    if (!os.is_open())
        return false;

    os.write((const char*)data.data(), data.size());
    return true;
}

Answer 1

I solved a similar problem with QImageReader and QBuffer

QByteArray data = QByteArray::fromRawData(reinterpret_cast<const char*>(buf.data()), buf.size());
QBuffer buffer(data);
QImageReader reader(&buffer);
QImage img = reader.read();

Answer 2

You need to make sure that the std::vector<uint8_t> &buf holds enough data for the picture. For QImage::Format_ARGB32 I would expect that 4 bytes are required per sample, so buf.size() must be at least 300*300*4.

What to do it it's less? Well, in that case probably the size of the image is not 300x300 and you need to figure out how to discover that. Or you can fill in buf with '0' to get the required size just to check that in that case everything works (or, better, nothing bad happens)

Answer 3

You can use this.

QImage img;
img.loadFromData(buf.data(), buf.size());
img.scaled(QSize(300,300), Qt::KeepAspectRatio);
...