Description for confidence image values

[Astronomia-maker]

Hello,
I have O3R 225 camera and installled ifm3dpy==1.1.1 for my python environment. I get array for confidence image, but values in arrays does not match this description.
This is the output for arrays that i am getting.
[[ 577 577 577 ... 513 513 577] [ 577 577 577 ... 513 513 577] [ 577 577 97 ... 41 513 513] ... [ 513 513 16417 ... 4129 513 513] [ 513 513 513 ... 513 513 513] [ 513 513 513 ... 513 513 513]]
So where can i find the description for the confidence image values?

[lola-masson]

Hi @Astronomia-maker,
The values in the confidence image can be combinations of multiple values. For instance:

577 = 512 + 64 + 1 -> Edge or isolated pixel (512) + invalid range (64) + invalid (1)
513 = 512 + 1 -> Edge or isolated pixel + invalid
97 = 64 + 32 + 1 -> Invalid range + Exposure index (32) + invalid

[lola-masson]

Unfortunately, this is the only way the confidence image is streamed from the device.

[BigBoot]

Hello @Astronomia-maker,

could you please clarify what value you're interested in exactly?

The values in the confidence image are 16bit bitflags, that means each bit in the value represents whether a flag is set or not.
To check if a pixel has a specific flag set you can use the bitwise and operator like this:

if img[42][42]&1024: 
    print("Pixel 42 42 is UNPLAUSIBLE")

You can also apply the bitwise and to the whole image like this:

unplausible = img&1024

which would result in an image where each value is either 1024 if the flag is set or 0 if it isn't.

[Astronomia-maker]

Hi @BigBoot, Thank you for the explanation. Actually i want to eliminate all pixel which have crosstalk error, some reflected pixel from the object if there is a box then there are some flying pixel around the box (not every time i have). So the confidence image value could be ( document reference:-32768,4096,16384,4,2) as far my knowledge.

[BigBoot]

In that case you could use something like the following:

flags = (32768|16384|4096|4|2)
mask = np.not_equal(confidence&flags, 0)

which results in a boolean mask like this:

[[ True  True  True ...  True  True  True]
 [ True  True  True ...  True  True  True]
 [False  True  True ...  True False  True]
 ...
 [ True  True  True ...  True  True  True]
 [ True  True  True ...  True  True  True]
 [ True  True  True ...  True  True  True]]

which can then be used to mask your initial image:

masked = np.ma.masked_array(img, mask=mask)

[Astronomia-maker]

Hi @BigBoot, Thank you for the suggestion and explanation. I will check out this suggested solution.