Loading [a11y]/accessibility-menu.js
Use of Intensity and Coherence of X-Band SAR Data to Map Thermokarst Lakes on the Northern Tibetan Plateau | IEEE Journals & Magazine | IEEE Xplore

Use of Intensity and Coherence of X-Band SAR Data to Map Thermokarst Lakes on the Northern Tibetan Plateau


Abstract:

In order to monitor the response of thermokarst lakes on the Tibetan Plateau to rapid climatic changes and human activities, an automated shoreline extraction approach fr...Show More

Abstract:

In order to monitor the response of thermokarst lakes on the Tibetan Plateau to rapid climatic changes and human activities, an automated shoreline extraction approach from the high-resolution TerraSAR/TanDEM-X imagery is proposed. First, a preprocessing scheme is applied to calibrate, coregister, and estimate the intensity and coherence of the imagery. Second, the statistical distribution of intensity and coherence are used to refine the data term of a region-based level set model. A distance regularized term and a variable weighting parameter are also added to improve the algorithm efficiency. Finally, a postprocessing is applied to remove false edges related to spurious segments and generate a vector map by geocoding the detected edges. Experiments on the intensity and coherence images demonstrate that: 1) the Fisher distribution can improve the usability of the level set method for extraction of shorelines using the intensity feature, because it can flexibly model the speckle fluctuations of many land classes; 2) the coherence information can significantly improve the efficiency and accuracy of segmentation by suppressing the misclassification caused by the sands around the lakes and the local winds over the lakes; 3) using a variable parameter to weight the data term and a regular term in a level set model can produce more robust and fast segmentation; and 4) using of intensity and coherence in unison can locate the exact position of most of the shorelines and reduce the relative uncertainty within 3 pixels with 90% confidence level.
Page(s): 3164 - 3176
Date of Publication: 25 April 2016

ISSN Information:

Funding Agency:


Contact IEEE to Subscribe

References

References is not available for this document.