76 
THEMATIC COMPRESSION OF 
METEOROLOGICAL IMAGE DATA 
Hans-Jörg Grund man n, Adolf Günther, Olaf Hellmuth 
Academy of Sciences of the GDR 
Institute for Space Research 
Satellite Ground Station Neustrelitz 
Kalkhorstweg 53, Neustrelitz 5. 2080 
Introduction 
Currently used digital imagery from meteorological satellites represents large 
data volumes. One scene in the visible channel of METROS AT IV imagery, for 
example, contains 25 000 000 pixels. Considerably higher data volumes are ex 
pected in future, since additional channels and sensors (i.e. radar imagery) will 
complement current satellite payloads. These data volumes require large storage 
capacities. A compression of data is, therefore, of gener 
In order to distribute satellite imagery between a centi: 
al interest. 
•al acquisition and process- 
ing facility and decentralized users data links using available lines would be re 
quired. Transmission times would, however, be prohibitively long for the large 
data volumes involved. A data reduction by compression would be needed, too. A 
design of relevant algorithms would be based on the following assumptions: 
♦ elimination of meteorologically non relevant information (i.g. separating 
variable image parts from non-variable). Transformation of imagery data 
into forms advantageous for visual presentation, by data reduction 
* reduction of redundance by parallel and s 
techniques. 
compression 
A combination of both assumptions is feasible, leading to further reduction by 
relevant algorithms. The described procedures of data reduction have been de- 
veloped at. the Satellite Ground Station Neustrelitz in close co-operation with me 
teorological users. Chapter 2 describes an adaptive compression procedure based 
on cloud classification with respect, to main pressure levels. 
An adaptive compression procedure is outlined in chapter 3 combining quad-tree 
decomposition and controlled adaptive coding. Data processing is implemented 
by a video processor presenting the results in visual form as well as preparing the 
data for transmission within a few seconds. 
2. Thematic compression of digital METEOSÄT infrared 
Th.e thematic compression is implemented by cloud classification with respect to 
main pressure levels (850 h Pa. 700 ii Pa, 500 h Pa. 300 h Pa). Reduction of the 
data volume is achieved by limitation to four cloud classes plus one rest class, as 
well as. through differentiation of land, water and clouds. Transmission is reduced 
to cloud classes. The transmission is aimed at providing the user only such data