Computer Estimation of the Presence of Sea Ice in Satellite Pictures.

• Main Author: Gerson,Donald
• Other Authors: MARYLAND UNIV COLLEGE PARK COMPUTER SCIENCE CENTER
• Format: Text
• Language: English
• Published: 1975
• Subjects: Snow; Ice and Permafrost; *PATTERN RECOGNITION; *SEA ICE; *ARCTIC REGIONS; FOURIER TRANSFORMATION; CLOUDS; POWER SPECTRA; STANDARD DEVIATION; ENTROPY; Spaceborne photography; Feature extraction; Arctic; Ice; permafrost; Sea ice
• Online Access: http://www.dtic.mil/docs/citations/ADA012960; http://oai.dtic.mil/oai/oai?&verb=getRecord&metadataPrefix=html&identifier=ADA012960

The need for ice data in widespread Arctic regions is discussed and the requirement that data be objectively obtained by computer analysis of satellite imagery is established. Since former attempts at such analysis which used only information contained in the satellite image produced poor results, a method is established by which data external to the image are used. A decision tree is employed to arrive at a conclusion as to the contents of an incremental area. The decisions are (1) land vs. non-land, based on a map; (2) water vs. non-water, based on mean grey-level, and (3) ice vs. cloud, based on texture analysis of the image. If cloud is identified an a priori estimation of ice vs. water is used. The ice vs. cloud decision is the main work of the paper.