3
$\begingroup$

Disclaimer: I am not an astronomer or astrophysicist. I'm a computer scientist.

Astronomers collect vast amounts of image data through astronmical surveys. The data is then analysed for signs of asteroids and comets.

  1. What are the most common algorithms used in this process?
  2. What characteristics does software 'look' for?
  3. Are there any relevant open-source tools?
$\endgroup$
0
$\begingroup$

Near-Earth Objects (NEO’s) are detected from imagery frames. The methodology is to ‘look’ for a tiny bright spot from several images, that has its pattern of brightness, shape and movement.

  1. Currently for NEO analysis Matched Filter Processing for Asteroid Detection is used. Matched Filter provides significant gain for processing images, used for asteroid detection and tracking.
  2. The methodology involves a two-step algorithm: i. mean removal and background noise; ii. enhancing the signal component. Raw images are preprocessed, removing pixel artifacts, registering the frames spatially, and equalizing the background intensities. A quadratic warp-fitting function used for the interpolation necessary to create a set of registered frames. Then a second-order warp with high accuracy positioning used in order to eliminate misregistering. Higher order registration techniques might be used. MF algorithm itself equalizes the images.
  3. I don’t know anything relevant as MF is not an open algorithm.

More information might be useful for the ‘Near-Earth Object Observations Program’ from NASA. https://www.nasa.gov/planetarydefense/neoo

$\endgroup$
1

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.