Stacking multiple astrophotography images—often called image stacking or integration—is one of the most important techniques in modern astronomy imaging. It allows photographers to reveal faint celestial details that would be lost in a single exposure.
Instead of taking one very long exposure (which can be ruined by a plane crossing the field or wind moving the equipment), astrophotographers take many shorter exposures of the same object (a galaxy, nebula, star field, or planet). Each image records the signal (real light from stars and deep-sky objects) as well as noise (random fluctuations from the sensor, heat, electronics, and sky background).
Because astronomical objects are effectively static over short time spans, these multiple images contain the same signal but different random noise patterns. Since the signal is consistent across images it adds together as the individual subs are built up, but the noise is random and averages out toward zero. As a result, the signal-to-noise ratio (SNR) improves by the square root of the number of images stacked (√N).
Indeed, the sky background, due to excessive light pollution, limits exposures of under 3-5 minutes in order to prevent the object from being swamped out. In very dark skies this is not generally a problem; but near cities where streetlights and buildings are dense, it is.
Noise is the biggest enemy in low-light imaging. Stacking reduces random noise far more effectively than any single long exposure, producing smoother backgrounds and cleaner detail. In addition, stacking allows bright stars and faint structures to coexist in the same image without clipping highlights or crushing shadows, revealing more tonal information across the scene.
In summary, image stacking works because math beats physics: Physics limits how much light a single exposure can collect without noise while statistics allow many imperfect images to combine into one high-quality result allowing modern amateur astrophotographers to produce images that rival professional observatory photographs from decades ago.
Of course, you need to determine an appropriate exposure time to capture the needed photons (if no photons are picked up by the sensor it doesn't matter how many subs you stack in the end). But you can limit the exposure time by quite a bit depending on the object of interest.
As an example of just how well this works, I imaged the famous Orion Nebula, M42, with my EdgeHD telescope in Hyperstar mode (an f/2 optic train). This nebula is the brightest nebula in the night sky. A single 60 second exposure can give you a nice result (albeit a bit noisy). A ten-minute exposure would really bring out the detail. But 10 minutes is way too long of an exposure in the light polluted sky of my backyard. The science of stacking, however, claims that a stack of ten, 1-minute exposures would provide the same image as a single ten-minute exposure with the added benefit of reduced noise, hence better SNR.
Here are three final photos of the Orion Nebula taken with differing sub-exposures and stacked to provide the total integration time of ten minutes.
Instead of taking one very long exposure (which can be ruined by a plane crossing the field or wind moving the equipment), astrophotographers take many shorter exposures of the same object (a galaxy, nebula, star field, or planet). Each image records the signal (real light from stars and deep-sky objects) as well as noise (random fluctuations from the sensor, heat, electronics, and sky background).
Because astronomical objects are effectively static over short time spans, these multiple images contain the same signal but different random noise patterns. Since the signal is consistent across images it adds together as the individual subs are built up, but the noise is random and averages out toward zero. As a result, the signal-to-noise ratio (SNR) improves by the square root of the number of images stacked (√N).
Indeed, the sky background, due to excessive light pollution, limits exposures of under 3-5 minutes in order to prevent the object from being swamped out. In very dark skies this is not generally a problem; but near cities where streetlights and buildings are dense, it is.
Noise is the biggest enemy in low-light imaging. Stacking reduces random noise far more effectively than any single long exposure, producing smoother backgrounds and cleaner detail. In addition, stacking allows bright stars and faint structures to coexist in the same image without clipping highlights or crushing shadows, revealing more tonal information across the scene.
In summary, image stacking works because math beats physics: Physics limits how much light a single exposure can collect without noise while statistics allow many imperfect images to combine into one high-quality result allowing modern amateur astrophotographers to produce images that rival professional observatory photographs from decades ago.
Of course, you need to determine an appropriate exposure time to capture the needed photons (if no photons are picked up by the sensor it doesn't matter how many subs you stack in the end). But you can limit the exposure time by quite a bit depending on the object of interest.
As an example of just how well this works, I imaged the famous Orion Nebula, M42, with my EdgeHD telescope in Hyperstar mode (an f/2 optic train). This nebula is the brightest nebula in the night sky. A single 60 second exposure can give you a nice result (albeit a bit noisy). A ten-minute exposure would really bring out the detail. But 10 minutes is way too long of an exposure in the light polluted sky of my backyard. The science of stacking, however, claims that a stack of ten, 1-minute exposures would provide the same image as a single ten-minute exposure with the added benefit of reduced noise, hence better SNR.
Here are three final photos of the Orion Nebula taken with differing sub-exposures and stacked to provide the total integration time of ten minutes.
 |
| 10x60sec |
 |
| 40x15sec |
 |
| 80x8sec |
Can you see any differences? In fact, looking at the original images (these are reduced size JPGs for posting here) the 80x8sec stacked version is noticeably better in terms of SNR, although the post processing software I use has some terrific noise reduction tools that were used in the creation of these images.
The drawbacks of this are few, but worth mentioning. It takes 8 times the storage to hold the 80 8-sec subs vs the 10 60-sec subs and the post processing time (and required storage on the computer) also goes up. But this is a small price to pay for such wonderful results.
Next time I'll target a fairly faint nebula to demonstrate how this really shines!
No comments:
Post a Comment