HDR imaging with 48 bit RGB images

asked 2015-05-12 08:08:31 -0600

Dolphin gravatar image

updated 2015-05-12 08:26:19 -0600

berak gravatar image


I'm wondering if anyone has been able to do HDR images using 48 bit RGB images (raw)? The tutorial http://docs.opencv.org/master/d3/db7/... only describes doing HDR with 32 bit images. I've implemented the example in python, but it fails on the type of image, which has to be 8 bits per colour channel, and not the 16 bit per channel I have from my raw files.

Kind regards


edit retag flag offensive close merge delete


Thank you, following the example (with my 48 bit RGB (16 bit per colour), the problem occurs when calculating the camera response using

imc=.. # list of images, each a numpy array MxNx3,of numpy float 32
exp=.. # list of exposures 
calibrate = cv2.createCalibrateDebevec()
respone = np.array()
calibrate.process(imc, response, exp)

This fails telling me that the image depth is wrong: OpenCV Error: Assertion failed (images[0].depth() == CV_8U) in process, file ../opencv/modules/photo/src/calibrate.cpp, line 72

Dolphin gravatar imageDolphin ( 2015-05-12 08:58:18 -0600 )edit

yea, you're right.

(had to delete the comment, because it only worked in some special cases. ;( )

berak gravatar imageberak ( 2015-05-12 09:01:33 -0600 )edit