Tracking cloud systems using ERBE scanner data: image construction and segmentation

Satellite and other types of images are familiar fools to most weather forecasters. Visual analysis of the moving patterns which appear in the images is used together with the forecaster's experience to guide in the forecasting of weather trends. Indeed, it is relatively easy for those interested in the weather to get a sense of cloud cover, cold fronts, rising humidities, etc., from images commonly seen on television. Although some of these atmospheric patterns are fairly recognizable from everyday experience, there has been relatively little research in the area of using automated methods to quantify these patterns on a global scale. The authors present their application of image segmentation to images built from global satellite observations over an extended time period. These techniques are the first steps in tracking the life histories of cloud systems.<>