Help with normalising
Hi all, trying to get useful data out of my Flir Lepton thermal camera.
It's giving me an 80x60 array of numbers, which can range from 0 to 65536, but they're usually around the 8000-mark.
The default pylepton_overlay code just uses cv2.normalize to stretch everything based on the min and max samples in the image. But the normalisation obviously changes based o n what's in the image.
What I want to do is clip everything below 7000 and above 9000, and normalise 7000 -> 0 and 9000 -> 65536.
Clipping is easy:
cv2.max(image,7000,image)
cv2.min(image,9000,image)
Now how do I nromalise to only the range of 7000:9000?
So far the auto-min/max code is:
cv2.normalize(image,image,0,65536,cv2.NORM_MINMAX,-1) # it needs another ',mask' at the end, that's what I want to know the format of.
The docs page is rather unhelpfully scant, "If you want to only use the mask to calculate the norm or min-max but modify the whole array, you can use norm() and Mat::convertTo()."
Yes, that's exactly what I want to do. Use a mask of ranges 7000 to 9000 as the min-max and normalise the whole array. Can anyone help me out as to how/syntax please?
I don't understand your problem. If you use the
max()andmin()functions like above, you end up with an array with values between 7000 and 9000. If then you callcv2.normalize(image, image, 0, 65536, cv2.NORM_MINMAX), the result is what you're expected, I think. The mask is optional, as the previous -1 (it indicated depth of destination matrix). If that is not working for you, please provide an example to better understand the problemSorry, I wasn't exactly clear. Yes, it works fine if I've got a range of 6500 to 10000, clip to 7000-9000 then normalise. The problem is for a dull flat background I've got a range of 7800-8000. Clipping does nothing obviously, then normalising sends 7800->0 and 8000->65536.