Automatic classification and compression of GMS cloud imagery in heavy rainfall monitoring application

The Bayes Decision method (BD) has been implemented on GMS Stretched VISSR (S.VISSR) images. Several objects are identified. These are earth’s surface, stratiform, convective and cirrus clouds. Overlaid 6-hour surface rainfall observations show that the rain area on the ground matches with precipitable cloud area identified by BD method. Finally the data of classified image are compressed to about one-eighth amount by use of Line-Run-Coding technique. It is helpful for data archive and remote transmission.