FLAVARS: A Multimodal Foundational Language and Vision Alignment Model for Remote Sensing