Hi everyone, I have a PNG file which is 16bt per channel (total 48bit), but when I check the bit-depth with openCV using depth() it keeps returning CV_8U (zero).
int imgDepth = cvImage.depth();
Can someone please provide me with a solution to this problem, or at least explain what I'm doing wrong?
Thanks in advance, Mark.
Check this. It says that depth( ) function returns the depth of an individual channel and channels( ) returns the number of channels in image. Thus, in your case, image seems to have depth of 16bits with number of channels equal to 4.
I still don't understand...
You say depth() returns the depth of an individual channel. I already understand that part, and then you say that my image 'seems to have depth of 16bits'. How can you derive an answer of 16bits per channel if the value being passed is CV_8U which according to the manual is 8bits per channel?
If CV_8U means '16bit with 4 channels', then how do I check for 8bit with 4 channels?
This is very non-intuitive to me, because to me CV_16U should be 16bit, and CV_8U should be 8bit. and even the manual says so, yet that's not how it's calculated???
I would really appreciate it if you could provide me the code to calculate the depth for one channel: eg:
int imgDepth = <code to get depth per channel> // equals 8 or 16
Asked: 2015-05-24 08:46:30 -0600
Seen: 40,317 times
Last updated: May 25 '15
you can find here your answer
Depth number of bits by channel A jpg image (like lena.jpg in opencv/data ) is 8 bits by channel and 3 channels but chicky_512.png is 8 bits by channel and 4channels
Sorry but I still don't understand. The documentation says depth() "Returns the depth of a matrix element" and "CV_8U - 8-bit unsigned integers ( 0..255 )" and "CV_16U - 16-bit unsigned integers ( 0..65535 )"
I already know all of this, and it's not working. I'm new to openCV and new to programming so I don't really know what to do...
may be you can share image
Here is the image
16rows X16 cols X 3channels pixels size 16 bits
mat->rows,mat->cols,mat->channels() mat->depth() (CV_16S)
Why am I getting CV_8U?
Here is my code:
I will try installing newer version and see if it still doesn't work.
may be use cv::IMREAD_UNCHANGED instead of CV_LOAD_IMAGE_COLOR
I managed to fix the problem using: CV_LOAD_IMAGE_ANYDEPTH | CV_LOAD_IMAGE_ANYCOLOR But now the problem is it won't recognize a 4 channel image (says '3'). If I use UNCHANGED flag then it crashes the program when loading a 4 channel image.
It seems that you are not using opencv 3.0? cv::IMREAD_UNCHANGED is for opencv 3.0. I don't know if imread had changed between 2.4.11 and 3.0. May be you should test your program with images in opencv/data giving explicit path for jpg and png and check it