RS-OVC: Open-Vocabulary Counting for Remote-Sensing Data

TL;DR

RS-OVC introduces an open-vocabulary counting model for remote-sensing imagery, enabling accurate counting of unseen object classes using textual and visual cues, thus overcoming the limitations of closed-set methods.

Contribution

This work presents the first open-vocabulary counting model for remote sensing, allowing for flexible counting of novel objects without re-annotation or re-training.

Findings

Model accurately counts unseen object classes.

Enables flexible, real-world remote sensing monitoring.

Uses textual and visual conditioning for open vocabulary understanding.

Abstract

Object-Counting for remote-sensing (RS) imagery is attracting increasing research interest due to its crucial role in a wide and diverse set of applications. While several promising methods for RS object-counting have been proposed, existing methods focus on a closed, pre-defined set of object classes. This limitation necessitates costly re-annotation and model re-training to adapt current approaches for counting of novel objects that have not been seen during training, and severely inhibits their application in dynamic, real-world monitoring scenarios. To address this gap, in this work we propose RS-OVC - the first Open Vocabulary Counting (OVC) model for Remote-Sensing and aerial imagery. We show that our model is capable of accurate counting of novel object classes, that were unseen during training, based solely on textual and/or visual conditioning.