I'm working with a stock Microscan Microhawk MV-40 (running the ID firmware), and need to lower the exposure time. The built in light is giving me 600 lux at the location of the bar code. This results in an exposure time of 4150 microseconds (us). If I increase the light, more lux, the exposure will decrease, less microseconds. Ignoring the sensor gain, there should be an inverse linear relation between lux and exposure time. So that got me thinking...
Is exposure time multiplied by lux a constant for a camera?
Can I use a new unit, us*lux to create a constant?
I think the answer is yex, er, yes sort of, almost, very close to yes with the tools I have.
The built in LED yields ~600 lux at 190 mm. The camera set it's exposure to 4150 us.
4150 us * 600 lux = 2,490,000 us*lux
Adding a white smart vision L300, I can get to 3000 lux. The camera decided to use an exposure of 850 us.
850 us * 3000 lux = 2,550,000 us*lux
So under either light sources the microsecond * lux values are about 2,500,000, or 2.5 million us*lux.
I should be able to get an exposure of 50 us if I have 50,000 lux...
us = 2,500,000 us*lux / 50,000 = 50 us
This is more useful the other way around...
If I know that I MUST have an exposure not more than 50 us, how much light is needed?
lux = 2,500,000 us*lux / 50 us = 50,000 lux.
Look, don't get too wrapped up in the details here.
Problem 1 is Lux is a measurement of light striking a surface, not light emitted by a light source.
Problem 2 is that I am ignoring the effect of the sensor's gain setting. Increasing the gain will act as a multiplier to the us*lux value
Problem 3 maybe the unit should be lux*us instead of us*lux. That's not really a problem, it's just a preference. You decide.
I just like the unit us*lux as a short hand for how much light I'm going to need when a vision system gets designed.
No comments:
Post a Comment