Deep Learning Meets Satellite Images -- An Evaluation on Handcrafted and Learning-based Features for Multi-date Satellite Stereo Images
This work addresses feature matching challenges for digital surface model generation in remote sensing, but it is incremental as it compares existing methods without introducing new techniques.
The paper evaluated handcrafted and learning-based feature matching methods for multi-date satellite stereo images, finding that traditional methods like SIFT remain competitive while deep learning approaches show promise in specific scenarios, based on tests with around 500 stereo pairs across two sites.
A critical step in the digital surface models(DSM) generation is feature matching. Off-track (or multi-date) satellite stereo images, in particular, can challenge the performance of feature matching due to spectral distortions between images, long baseline, and wide intersection angles. Feature matching methods have evolved over the years from handcrafted methods (e.g., SIFT) to learning-based methods (e.g., SuperPoint and SuperGlue). In this paper, we compare the performance of different features, also known as feature extraction and matching methods, applied to satellite imagery. A wide range of stereo pairs(~500) covering two separate study sites are used. SIFT, as a widely used classic feature extraction and matching algorithm, is compared with seven deep-learning matching methods: SuperGlue, LightGlue, LoFTR, ASpanFormer, DKM, GIM-LightGlue, and GIM-DKM. Results demonstrate that traditional matching methods are still competitive in this age of deep learning, although for particular scenarios learning-based methods are very promising.