Abstract: In view of the problem of XPGR (crosspolarized gradient ratio) algorithm existed in the icesheet freezethaw detection， an XPGR combined with wavelet transform algorithm was proposed， which did not depend on the field observations. The proposed algorithm was used in the Antarctic icesheet freezethaw detection by microwave radiometer. That was the processing for the long time series data of XPGR algorithm by the use of wavelet transform model to automatically get the distribution of Antarctic icesheet melting onset date， melting duration and melting end date. The computational efficiency， usability and operability of detecting were improved in the icesheet freezethaw detection because many samples could be automatically selected by the algorithm with certain denoising effect. To some extent， the accuracy of detecting icesheet freezing and thawing was also improved and the goal of Antarctic icesheet monitoring system construction was better achieved.

Key words: icesheet freezethaw detection, microwave radiometer, XPGR, wavelet transform, accuracy