🚀 go-pugleaf

In article <1992Jul23.201102.7331@emr1.emr.ca>, naunheim@ccrs.emr.ca (J Naunheimer) writes:
|> 
|> We have recently received some ERS-1 data from the 35-day Multi-disciplinary
|> Phase.  The data were processed to 4 looks in slant range, at a pixel dimension
|> of 35 by 30 metres.
|> 
|> We are interested in filtering the data for display purposes, and we were
|> wondering if any other research groups have investigated filter types and 
|> sizes which maintain radiometric and geometric quality, but effectively 
|> reduce speckle.
|> 
|> If anyone has any information or success stories with using filtering
|> algorithms on ERS-1 data, my collegues and I would appreciate you sharing
|> them.
|> 

Speckle reduction in ERS-1 images (and SAR images) are performed by using a low pass
filter of any kind, which means :
  a) using multi-looks images (summing independant pixels)
  b) performing low pass filtering (loosing resolution)
  c) using a MAP filter (Maximum A Priori Probablility) developed by the CESR team
     Touzi, Lopes, Nezri, et al. (cf IGARSS proceedings, 1990-91-92)

     they can be reached by snail mail at
     CESR, avenue du colonel Roche
     31055 Toulouse Cedex, France.

     There is also a paper in the ESA free bulletin Earth Remote Sensing of July 92
     (or is it August ?)
--
vutien@cenatls.cena.dgac.fr
vutien@capsogeti.fr