Detection of oil fire smoke over water in the Persian Gulf region

Smoke from the recent Kuwait oil fires and the regional dust storms, collectively referred to as aerosol features, attracted enormous attention among a wide variety of scientists and environmentalists. Yet many real-time interpreters of NOAA Advanced Very High Resolution Radiometer (AVHRR) images have difficulties in consistently detecting and distinguishing these features. To identify aerosoI features, automated statistical and non-statistical texture measures have been investigated. The appropriate measures were extended and successfully applied to this problem. The difficulty in detecting aerosol features arises from two different phenomena. First, aerosol features such as dust, smoke, and fog have very similar signatures in the spectral dimension of AVHRR data