InstructSAM: A Training-Free Framework for Instruction-Oriented Remote Sensing Object Recognition

TL;DR

InstructSAM is a training-free, instruction-driven framework for remote sensing object recognition that leverages large vision-language models to interpret instructions and efficiently identify objects without extensive training.

Contribution

The paper introduces InstructSAM, a novel training-free approach that interprets instructions for remote sensing object recognition, along with a new benchmark dataset and tasks for open-vocabulary scenarios.

Findings

InstructSAM matches or surpasses specialized baselines in multiple tasks.

It maintains near-constant inference time regardless of object count.

Reduces output tokens by 89% and runtime by over 32% compared to direct generation.

Abstract

Language-Guided object recognition in remote sensing imagery is crucial for large-scale mapping and automated data annotation. However, existing open-vocabulary and visual grounding methods rely on explicit category cues, limiting their ability to handle complex or implicit queries that require advanced reasoning. To address this issue, we introduce a new suite of tasks, including Instruction-Oriented Object Counting, Detection, and Segmentation (InstructCDS), covering open-vocabulary, open-ended, and open-subclass scenarios. We further present EarthInstruct, the first InstructCDS benchmark for earth observation. It is constructed from two diverse remote sensing datasets with varying spatial resolutions and annotation rules across 20 categories, necessitating models to interpret dataset-specific instructions. Given the scarcity of semantically rich labeled data in remote sensing, we propose InstructSAM, a training-free framework for instruction-driven object recognition. InstructSAM leverages large vision-language models to interpret user instructions and estimate object counts, employs SAM2 for mask proposal, and formulates mask-label assignment as a binary integer programming problem. By integrating semantic similarity with counting constraints, InstructSAM efficiently assigns categories to predicted masks without relying on confidence thresholds. Experiments demonstrate that InstructSAM matches or surpasses specialized baselines across multiple tasks while maintaining near-constant inference time regardless of object count, reducing output tokens by 89% and overall runtime by over 32% compared to direct generation approaches. We believe the contributions of the proposed tasks, benchmark, and effective approach will advance future research in developing versatile object recognition systems.