Near Earth Object Detection Using a Poisson Statistical Model for Detection on Images Modeled from the Panoramic Survey Telescope and Rapid Response System
Date of Award
Master of Science
Department of Electrical and Computer Engineering
Stephen C. Cain, PhD.
The purpose of this research effort is to develop, simulate, and test a new algorithm to detect Near Earth Objects (NEOs) using a Likelihood Ratio Test (LRT) based on a Poisson statistical model for the arrival of photons. One detection algorithm currently in use is based on a Gaussian approximation of the arrival of photons, and is compared to the proposed Poisson model. The research includes three key components. The first is a quantitative analysis of the performance of both algorithms. The second is a system model for simulating detection statistics. The last component is a collection of measured data to apply comparatively to both algorithms. A Congressional mandate directs NASA and the DoD to catalogue 90% of all NEOs by the year 2020.  Results from this research effort could feasibly be applied directly to operations in the Pan-Starrs program to facilitate the accomplishment of the Congressional mandate. Improvements in the size of detectable NEOs and in the probability of detecting larger NEOs would increase the state of readiness of the world for possible catastrophic impact events. Improvements in detection probability of measured data were as high as a factor of seven, and the expected average improvement is around 10%.
DTIC Accession Number
Peterson, Curtis J. R., "Near Earth Object Detection Using a Poisson Statistical Model for Detection on Images Modeled from the Panoramic Survey Telescope and Rapid Response System" (2012). Theses and Dissertations. 1146.