Boundary detection using multisensor imagery: Application to ice sheet margin detection

Boundary mapping and feature extraction becomes the first step in the interpretation of any remotely sensed imagery. This type of segmentation is one of the most important process for automated image analysis because objects of interest are extracted from an image for subsequent processing. A boundary detection scheme was devised and implemented to achieve this end. The ice sheet margin mapping algorithm was successfully performed on multisensor remotely sensed data, comprising SAR amplitude data, visible imagery, and INSAR (Interferometric SAR) coherence data.; Due to the complex nature of the ice sheet boundary, a multisensor approach reduces potential errors in ice sheet margin mapping and aids in feature identification and differentiation. This study opens a new look to the hyper-spectral approach for detecting boundaries and demonstrates the utility of multisensors for detecting ice sheet margins. Also with successful application of detecting the ice sheet margin using multisensor data, it was possible to measure ice sheet fluctuations. This suggests that with repeat acquisition of remotely sensed data it is feasible to determine the extend of boundary changes in the environment over time.