article · 2014

Binary Detection using Multi-Hypothesis Log-Likelihood, Image Processing

Cited by 1Abstract and details only

Abstract

Abstract : One of the United States Air Force missions is to track space objects. Finding planets, stars, and other natural and synthetic objects are all impacted by how well the tools of measurement can distinguish between these objects when they are in close proximity. In astronomy, the term binary commonly refers to two closely spaced objects. Splitting a binary occurs when two objects are successfully detected. The physics of light, atmospheric distortion, and measurement imperfections can make binary detection a challenge. Binary detection using various post processing techniques can significantly increase the probability of detection. This paper explores the potential of using a multi-hypothesis approach. Each hypothesis assumes one two or no points exists in a given image. The log-likelihood of each hypothesis are compared to obtain detection results. Both simulated and measured data are used to demonstrate performance with various amounts of atmosphere, and signal to noise ratios. Initial results show a significant improvement when compared to a detection via imaging by correlation. More work exists to compare this technique to other binary detection algorithms and to explore cluster detection.

Cite this paper

APA 7

Gessel, B. H. (2014). Binary Detection using Multi-Hypothesis Log-Likelihood, Image Processing.

MLA 9

Gessel, Brent H. “Binary Detection using Multi-Hypothesis Log-Likelihood, Image Processing.” 2014.

Chicago

Gessel, Brent H. 2014. “Binary Detection using Multi-Hypothesis Log-Likelihood, Image Processing.”

Harvard

Gessel, B.H. (2014) ‘Binary Detection using Multi-Hypothesis Log-Likelihood, Image Processing’.

BibTeX

@article{gessel2014binary,
  title = {{Binary Detection using Multi-Hypothesis Log-Likelihood, Image Processing}},
  author = {Gessel, Brent H.},
  year = {2014}
}

Made from the index record. Check names, capitalisation and page numbers against the paper. Need another style? AI for Research has over 10,000.