Wind measurements were obtained by tracking water vapor features on Meteosat and GOES-VAS 6.7 micron water vapor images. While pure water vapor features are fuzzy, there are discernible features which can be tracked. An investigation of preprocessing algorithms designed to bring out the features to be tracked showed that high pass filters tended to bring out the noise in the image, while low pass filters washed out the features. Sliced linear gray scale stretching under operator control gave the best enhancement to the water vapor features.