2ros0
Posts: 1 Joined: Tue Jul 16, 2019 3:00 pm
Post
by 2ros0 » Tue Jul 16, 2019 3:08 pm

Hi,

I was looking to see what kind of depth error I could expect using the stereo cameras here:https://nerian.com/support/resources/calculator/

On plugging in different values for "expected disparity error", I expected the depth error to vary linearly. But that doesn't seem to be the case.

I was using the formula

where, delta_Z is depth error

Z is target depth

B is baseline

f is focal length

delta_D is expected disparity error

Is there something else in the nerian stereo cameras that I have to take into account?

k.schauwecker
Posts: 59 Joined: Mon Mar 25, 2019 1:12 pm
Post
by k.schauwecker » Tue Jul 16, 2019 3:18 pm

Hi,

your equation is wrong. Z is computed by:

so a delta in D will result in:

Code: Select all

```
delta_Z(D) = B*f / (D + delta_D) - B*f/D
```

this delta is dependent on the disparity (and hence the depth) and is thus not linear but it increases with the depth. The solution is approximately quadratic for small delta_D, which is what you see in the error chart.

Regards,

Konstantin