About RGBQUAD type to OpenCV type

Question

m_DrawDepth.DrawFrame( (BYTE*) m_rgbWk );

Can I use m_rgbWk to get a lpimage data?

I want to use OpenCV to process Kinect photo.

Sorry,my english is poor! And, really thanks for all!!

Answer 1

I used cvSetData to get that data into an OpenCV image, and that works. You need to initialize the image header correctly, though.

Answer 2

Is like this?

IplImage * kinectDepthImage ;
DepthImage = cvCreateImage( cvSize(640,480), 16, 1);
cvSetData(DepthImage, (BYTE*) m_rgbWk, kinectDepthImage->widthStep);
cvShowImage("DepthImage",DepthImage);

But my code has some errors.

My OpenCV's Windows continued to wait state.

Can you help me to see which side wrong?

Really thank you very much!!

Answer 3

The part to set the data is correct.

For the depth image, since OpenCV cannot display an image which is of depth of 16, you'll need to convert it to a gray scale image. You can create do like:

DepthImageGray = cvCreateImage(cvSize(640, 480), 8, 1);

Then you can loop through the DepthImage and scale the data to 8 bits,

or you can try cvConvertScale(DepthImage, DepthImageGray, 1/16.0, 0);

 

For the color image, the format is BGRA from Kinect, so you have to create like ColorImage = cvCreateImage(cvSize(640, 480), 8, 4), then set the data.

 

And remember to add cvWaitKey(10) at the end of the loop to let the display window refresh the data.

Answer 4

I use the code that write "cvCreateImage(cvSize(320, 240), 8, 4)".

And it's success!!

Thank you so mush!!

Answer 5

Why dont use Mat instead?, since youre aready working in c++ (I suppose). Imho, its much simplier to manipulate.