Byte array to QImage

[Kent-Dorfman]

what does QtGui.QImageWriter.supportedImageFormats() report? That should tell you something.

[Christian Ehrlicher]

Do you really think the camera is sending jpeg images?

[beecksche]

@hardikgarg19
Maybe the set format is wrong. Try Format_Grayscale8 or Format_RGB32.

Or try to leave the format blank, so QImage will detect the reading format by itself.

QImage::fromData((uchar*)pBuf, iBytes*sy);

All described in the docs

[SGaist]

Hi,

Besides the points of my fellows, are you sure about the size of the image ? 1200x1920 seems a bit non standard for a camera.

On that point, what camera is that ?
How do you get data from it ?
How did you configure it ?

[hardikgarg19]

@beecksche
I have tried the change, but, I am getting the bool value as false and no image is created.

    QImage img(1200,1920,QImage::Format_Grayscale8);
    img = QImage::fromData((uchar*)pBuf,iBytes*sy);
    bool b = img.save( "C:\\test_folder\\file_img.BMP");

[Christian Ehrlicher]

You really should try to find out what format the camera is sending - guessing is not the correct solution here.

[mranger90]

I'm curious about the "iBytes * sy" in your code.
If you are truly grabbing an 8 bit gray scale image from the camera then the size of the data will likely be just "iBytes" which, and I'm going out on a limb here, should be imageWidth * imageHeight.
If you are multiplying this by some factor, the mysterious "sy", then the image is probably RGB24 (sy == 3 ?) or RGB32 (sy == 4 ?).

[hardikgarg19]

@SGaist Hi,
I am using a FPGA based camera (http://isgcameras.com/allegro-usb3-family/).

Below is the code:

long sx,sy,iBpp;
//get the image size

//pMedleyUcam is the camera object
pMedleyUcam->get_SizeX(&sx);
pMedleyUcam->get_SizeY(&sy);

pMedleyUcam->GetBytesPerPixel(&iBpp);
int iBytes=sx*iBpp; // horizontal image size in bytes

BYTE* pBuf=(BYTE*)malloc(iBytes*sy);

short v;
pMedleyUcam->Grab(&v);
BYTE *ptr=pBuf;
for(int y=0;y<sy;y++)
    {
        pMedleyUcam->GetImagePointer(0, sy-1-y,&var);
        memcpy(ptr,var.pbVal,iBytes);
        ptr+=iBytes;
    }

//form the image using the pBuf
QImage img(1200,1920,QImage::Format_Grayscale8);
img = QImage::fromData((uchar*)pBuf,iBytes*sy);
bool b = img.save( "C:\test_folder\file_img.BMP");

b is returning false here

[hskoglund]

Hi, if you try with double slashes

bool b = img.save( "C:\\test_folder\\file_img.BMP");

is b still returning false?

[hardikgarg19]

@Christian-Ehrlicher
The camera allows me to check the available pixel format. Below is the list:
Mono8
Mono10p
Mono12p
BayerRGB
BayerRGB10p
BayerRGB12p
RGB8
RGB565P
BGR8
YCbCr422_8
YUV400_8
YUV8_UYV

[hardikgarg19]

@hskoglund Hi,
Yes, it is still returning false.

[hskoglund]

Looks like you're only getting the raw pixels from the camera (in the specified pixel format). To save to a file, like a BMP file., you need an header as well.
To get one, you could try create an empty QImage with the specified x and y size, and then fill it with those pixels you get from the camera, say something like:

QImage img(1200,1920,QImage::Format_Grayscale8);
for (int y = 0; (y < 1920); ++y)
    for (int x = 0; (x < 1200); ++x)
        img.setPixel(x,y,((uchar*)pBuf) + x + 1200 * y);

[fcarney]

According to this SO question you can do a line at a time:
https://stackoverflow.com/questions/14563180/raw-data-to-qimage

[hardikgarg19]

@hskoglund Hi,
Yes, that was the problem. Thank you so much.

[hardikgarg19]

@fcarney Hi,
Thank you so much for sharing the link. I have used the QImage.scanLine and it worked.