HR: 1340h
AN: SH13A-1096 [Abstracts]
TI: Application of Statistical Image Segmentation to Recognition of Solar Magnetic Network
AU: * Jones, H P
EM: hjones@nso.edu
AF: National Solar Observatory, PO Box 26732, Tucson, AZ 85726, United States
AU: Malanushenko, O V
EM: elena@noao.edu
AF: Apache Point Observatory, PO Box 59, Sunspot, NM 88349, United States
AU: Pap, J M
EM: Judit.M.Pap.1@gsfc.nasa.gov
AF: NASA's Goddard Space Flight Center, GEST, UMBC
Code 612.1
NASA's Goddard Space Flight Center, Greenbelt, MD 20771, United States
AU: Turmon, M J
EM: turmon@jpl.nasa.gov
AF: Jet Propulsion Laboratory, Jet Propulsion Laboratory
California Institute of Technology, Pasadena, CA 91109, United States
AB:
We have developed a statistical method for feature identification in NSO multidimensional imagery which
requires a training set of independently determined image segmentations. The large spatial scale of our initial
training set determined by the algorithm of Harvey and White (1999, ApJ 515, p. 812) mixes the details of
magnetic network which are contained in the observations with quiet Sun and other features. We have found it
difficult to reproduce this large scale in models of conditional and prior probabilities and are in fact interested in
marking smaller scale structures for comparison with variation of total and spectral solar irradiance. We describe
in this paper the performance of our technique with finer scale training sets determined by observations from
other instruments and independently for the NSO data.
DE: 7538 Solar irradiance
DE: 7594 Instruments and techniques
SC: SPA-Solar and Heliospheric Physics [SH]
MN: 2007 Fall Meeting