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

arXiv cs.CV / 4/13/2026

📰 NewsSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

The paper highlights a key limitation in remote-sensing object-counting: most methods only work for a closed set of object classes seen during training, requiring re-annotation and retraining to handle new classes.
It proposes RS-OVC, the first open-vocabulary counting model tailored to remote-sensing and aerial imagery, enabling counting of novel object categories without having seen them during training.
RS-OVC is designed to perform this open-vocabulary counting using textual and/or visual conditioning as guidance signals for which objects to count.
The authors report that the model can accurately count classes that are unseen during training, aiming to make RS monitoring more adaptable to real-world dynamics.

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.