Please use this identifier to cite or link to this item:
|Title:||Antarctic snowmelt detection for QuikSCAT scatterometer data based on mathematical morphology combined with wavelet transform|
|Keywords:||Generalized Gaussian model;Mathematical morphology;Microwave scatterometer;Snowmelt detection;Wavelet transform|
|Abstract:||Microwave scatterometer is sensitive to the melting snow. When the freeze-thaw phenomenon occurs, the backscatter coefficients will have a sharp rising and falling mutation. Mathematical morphology has the characteristics with edge-preserving filter and wavelet transform has the characteristics with the automatic edge extraction, which does not depend on the priori snowmelt information. A new automatic Antarctic snowmelt detection method was proposed based on mathematical morphology combined with wavelet transform. This method improves the snowmelt detection accuracy, because this method can remove the interference of the edge extraction. Melt onset date, end date and duration can be obtained with high accuracy by identifying and tracking the sharp rising and falling edge. Compare the snowmelt results in this work with the temperature of ten automatic weather stations (AWS), which shows that the snowmelt detection method proposed in the paper improves the detection accuracy from about 50 % to 62.5 % in AWS Cape Denison.|
|ISSN:||0975-1033 (Online); 0379-5136 (Print)|
|Appears in Collections:||IJMS Vol.49(02) [February 2020]|
Items in NOPR are protected by copyright, with all rights reserved, unless otherwise indicated.