11-30, 17:00–17:30 (Asia/Seoul), Gallery 3
The registration of multi-modal satellite images (e.g., SAR, EO) is essential in fields using multiple images such as time series analysis, change detection, and image fusion. The usability of satellite images acquired from various sensors can be increased only when the images are accurately registered to each other. However, multi-modal satellite image registration is still challenging due to the different characteristics of the images. Therefore, in order to explore an effective way for multi-modal satellite image registration, this study intends to apply a matching method using feature extractors of pre-trained deep neural networks and compare the matching performance of the models.
This work was supported by the Agency For Defense Development by the Korean Government.
-Present, Senior Researcher, Agency for Defense Development (ADD)
-2018, Ph. D. in Civil and Environmental Engineering, Yonsei University
-2012, Bachelor of Civil and Environmental Engineering, Yonsei University