Memory-Augmented Multimodal Large Language Models for Small Object Understanding in Streaming Aerial Videos

Penglei Sun Yehua Huang Zhuoli Tao Xiang Li Runwei Guan Yaoxian Song Kai Zhao Henghui Ding Bo Han Ieee Yang Yang Senior Member Ieee Xiaowen Chu Fellow

2026 · arXiv.org · Vol abs/2607.19857 · 0 citations · 47 references

Computer Science

TL;DR

This work proposes SkyAnchor, an MLLM with two designs to the above challenges: a Semantics-Aware Token Router that preserves small-target under a reduced visual-token budget, and a Hierarchical Memory Bank that keeps the target consistently understood on streams.

DOI Semantic Scholar

View source

Similar papers

Conference Jul 2026

Spatial Information and Metadata Generation in Aerial Robotics with Vision-Language Models

Unmanned aerial vehicles (UAVs) are used for a wide range of tasks, many of which are autonomous. Since UAVs operate in three-dimensional space, visual data presents a rich source of input for them. Meanwhile, the rapid development of artificial intelligence has given rise to vision-language models (VLMs). These models...

Elena Wachtler, Antonella Barisic Kulas, Tamara Petrović et al. · International Conference on... · 0 citations

Preprint Aug 2026

RSVideo: Are Your Vision-Language Models Ready for Remote Sensing Videos?

RSVideo, a reinforcement learning framework for small-target spatiotemporal focusing that selects question-relevant regions across frames and suppresses redundant background tokens is proposed, a reinforcement learning framework for small-target spatiotemporal focusing that selects question-relevant regions across fram...

Hong-Jie Zhou, Shiqin Wang, Haoyang Chen et al. · 1 citation

Preprint Aug 2026

ARIES-Mission2: A Zero-Shot Vision-Language-Action Framework for Fast Large-Scale Aerial Mission Generation

This work proposes ARIES-Mission2, a zero-shot Vision-Language-Action (VLA) framework that decouples visual-semantic perception from physical route optimization in low-altitude Unmanned Aerial Vehicle (UAV) mission generation and indicates that the TSP module maintains lower growth in computation time as the number of...

Jun-Hao Wei, Yan-Xiao Li, Hao-Chen Li et al. · 1 citation

Jul 2026

SkyVLaM: Multimodal Large Language Model for UAV Video Understanding in Remote Sensing

Experiments show that SkyVLaM provides a more effective allocation of the visual token budget and improves language-conditioned video segmentation in UAV scenarios.

Kaiwen Jing, Rui-Jian Jia, Bing-Yao Li et al. · arXiv.org · 0 citations

Jul 2026

M4World: A Multi-view Multimodal Driving World Model for Interactive Object Manipulation and Minute-long Streaming

M$^\text{4}$World, a Multi-view and Multimodal generative driving world model that synthesizes future surround-view video streams and synchronized LiDAR scans while supporting interactive object Manipulation and stable Minute-long streaming is presented.

Ke Cheng, Hanqiao Ye, Lei Shi et al. · arXiv.org · 1 citation

Jul 2026

Knowing the Self, Understanding the World: A Dual-Cognition Benchmark for UAV Spatio-temporal Reasoning with MLLMs

UAV-DualCog is presented, a benchmark for aerial multiview spatio-temporal reasoning built on this dual-cognition perspective that includes both image and video tasks to jointly evaluate self-state and environment-state reasoning, while requiring spatial or temporal grounding beyond discrete answer prediction.

Like Liu, Zheng Xu, Haitao He et al. · arXiv.org · 1 citation