Interpretation of sea ice using NOAA-AVHRR data

The different channels of the NOAA-AVHRR (National Oceanic and Atmospheric Administration-Advanced Very High Resolution Radiometer) satellites were studied to determine the best single channel and channel combinations to use in distinguishing between clouds, sea-ice types and open water. The study a...

Full description

Bibliographic Details
Main Authors: Ahlnäs, Kristina, Kuittinen, Risto, Parmes, Eija
Format: Book
Language:English
Published: VTT Technical Research Centre of Finland 1990
Subjects:
Online Access:https://cris.vtt.fi/en/publications/b2213b28-be70-4f64-8817-fbed6b0c00cb
Description
Summary:The different channels of the NOAA-AVHRR (National Oceanic and Atmospheric Administration-Advanced Very High Resolution Radiometer) satellites were studied to determine the best single channel and channel combinations to use in distinguishing between clouds, sea-ice types and open water. The study area was situated in the Bay of Bothnia, the northernmost part of the Baltic Sea, in the winter of 1987. The best channel combination for cloud separation was 124/RGB (Red Green Blue) making clouds bright yellow and open water blue. This combination worked well from mid-February through May. However, when available, as in NOAA-9, Ch 125/RGB was a better choice than Ch 124/RGB. Leads and open water could be seen in channel 5 but not in channel 4. The near IR channel 2, in combination with one thermal IR channel, 4 or 5, was also sufficient to separate clouds from ice. Image enhancement by exponential and logarithmic functions improved the interpretation of the image. Numerical interpretation with spectral and textural information of NOAA images was used to interpret clouds, water and sea-ice types. In this task contrast and entropy were used as textural parameters. Supervised and unsupervised methods were used in classification. The results showed that water, open ice, close ice and partly ridged as well as fast ice can be discriminated although discriminating between fast ice and clouds as well as fast ice and ridges is difficult. Entropy was a better textural parameter than contrast but an algorithm has to be developed to make the interpretation faster.