Blind Search of Faint Moving Objects in 3D Data Sets
Abstract
Targets that move in the image of 2-D framing sensors are detected with less sensitivity than stationary ones. The reduction in sensitivity occurs when the object streaks during a frame s exposure time. Detection schemes that utilize photon uncorrupted coordinates (on imager) and arrival times can potentially outperform those that pixelate photon coordinates and bin arrival times by frame. Photon counting imager (PCI) technology, e.g. Los Alamos National Laboratory s Remote Ultra Low Light Imager, offers large format imaging sensors with sub-nanosecond time resolution, and hence provides a unique opportunity to enhance sensitivity with respect to the detection of moving objects. The ability to time tag and locate each photon results in a clear advantage in detecting moving objects; that can be attributed to the much reduced number density on a 3-D streak in comparison to its 2-D projection. In spite of a promise of higher detection sensitivity since the concept was introduced and proposed in 1993, to our best knowledge, no automated algorithm has been implemented to demonstrate the predicted gain in sensitivity. We demonstrate the anticipated sensitivity enhancement using a simulated object signature superimposed on measured background, and show that the limiting magnitude can be improved by up to 6 visual magnitudes. A quasi blind search algorithm that identifies the streak of photons, assuming no prior knowledge of orbital information, will be discussed and results shown. Performance of PCI sensors for this application will be presented for several telescope/imager configurations.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2013
- Accession Number
- ADA591502
Entities
People
- Patrick Mcnicholl
- Peter Crabtree
- Phan Dao
- Tamar Payne
Organizations
- Air Force Research Laboratory