I was just learning about poisson processes and all the things you can model with them in eecs126. Cool connection to this class! (Photons arrive in a poisson process manner)
adam2451
How can we tell that this number is modeled by a Poission distribution?
S-Muddana
One interesting implication of photon shot noise is its impact on low-light imaging scenarios. In situations where the number of photons detected is low, such as astrophotography or imaging in dimly lit environments, photon shot noise becomes more pronounced. As a result, images captured under such conditions may exhibit graininess or speckle-like patterns, limiting the ability to discern fine details.
Liaminamerica2
The Poisson distribution is very cool as it is the binomial distribution with n approaching infinity and the probability p of success approaching zero. In this case, because there are lots of photons hitting the sensor and the probability of a photon getting picked up by the sensor is relatively small, then this can be modeled well with the distribution.
charshou
Do the number of photons that come in from each exposure stay the same for each unit of time (ei. the first second of exposure vs the last second of exposure)? If so, how do we scale the lambda value that defines the Poisson distribution to account for this?
Mehvix
@adam2451 the distribution of the difference between events from many IID geometric processes (e.g. photons hitting an image sensor) is Poisson. Poissons clump together by nature, so there are surely bursts of arrivals that we can't curb. This arises since each pixel has a fixed exposure time (or you could increase overall exposure time)
We just covered this (queueing theory) in CS162.
sparky-ed
I think it's very interesting to see how the number of photons hitting the sensor during exposure can make the image clear or noisy. This variance, which follows a Poisson distribution, is common in nature for modeling random events. It's fascinating to realize that shot noise is an inherent part of the imaging process, highlighting the limits of precision imposed by the laws of physics.
I was just learning about poisson processes and all the things you can model with them in eecs126. Cool connection to this class! (Photons arrive in a poisson process manner)
How can we tell that this number is modeled by a Poission distribution?
One interesting implication of photon shot noise is its impact on low-light imaging scenarios. In situations where the number of photons detected is low, such as astrophotography or imaging in dimly lit environments, photon shot noise becomes more pronounced. As a result, images captured under such conditions may exhibit graininess or speckle-like patterns, limiting the ability to discern fine details.
The Poisson distribution is very cool as it is the binomial distribution with n approaching infinity and the probability p of success approaching zero. In this case, because there are lots of photons hitting the sensor and the probability of a photon getting picked up by the sensor is relatively small, then this can be modeled well with the distribution.
Do the number of photons that come in from each exposure stay the same for each unit of time (ei. the first second of exposure vs the last second of exposure)? If so, how do we scale the lambda value that defines the Poisson distribution to account for this?
@adam2451 the distribution of the difference between events from many IID geometric processes (e.g. photons hitting an image sensor) is Poisson. Poissons clump together by nature, so there are surely bursts of arrivals that we can't curb. This arises since each pixel has a fixed exposure time (or you could increase overall exposure time)
We just covered this (queueing theory) in CS162.
I think it's very interesting to see how the number of photons hitting the sensor during exposure can make the image clear or noisy. This variance, which follows a Poisson distribution, is common in nature for modeling random events. It's fascinating to realize that shot noise is an inherent part of the imaging process, highlighting the limits of precision imposed by the laws of physics.