opencv convertto need to manually scale down inputs ?

scale

asked 2017-01-20 11:28:02 -0600

Nbb
731 ●12 ●23 ●38

if i am converting a cv_32s to cv_8u do i need to scale down the values or will opencv do that for me. what is the equation to scale everything ?

new = (current - min) / (max - min) * 255 ?

edit retag flag offensive close merge delete

Comments

1

can it be, you want a normalization (resulting in a CV_8U image) in the 1st place ?

berak ( 2017-01-20 11:47:53 -0600 )edit

add a comment

1

answered 2017-01-24 12:06:03 -0600

Pedro Batista

2213 ●5 ●18 ●42 https://lnkd.in/bw9XGH8

updated 2017-01-24 12:08:56 -0600

Just to complement @pi-null-mezon's answer.

Before converting image type from 32S to 8U, you must normalize the image. For that you don't need to create your own algorithm, OpenCV has a normalize function

This call

cv::normalize(inMat, outMat, CV_MINMAX, 0, 255)

Will transform the smallest pixels to 0, the highest one to 255, and all the in-between accordingly to the right proportion. After that, you convert outMat to 8U and you are ready to display it.

edit flag offensive delete link

add a comment

1

answered 2017-01-24 04:44:02 -0600

pi-null-mezon

536 ●2 ●13

Hi! As docs says saturate_cast will be applied after conversion. If you downscale precision (from CV_32S to CV_8U) than all converted values that are more than 255 becomes 255 and less than 0 becomes 0. So, now you should decide is that behavior one that you want?

edit flag offensive delete link

add a comment

opencv convertto need to manually scale down inputs ?

Comments

2 answers

Links

Question Tools

Stats

Related questions

opencv convertto need to manually scale down inputs ? edit

Comments

2 answers

Links

Question Tools

Stats

Related questions

opencv convertto need to manually scale down inputs ?