Efficient On-Board Processing of Oblique AAV Video for Rapid Flood Extent Mapping

Abstract

Effective disaster response relies on rapid situational awareness, where oblique aerial video is the primary modality for initial scouting due to its ability to maximize spatial coverage and situational awareness in limited flight time. However, the on-board processing of high-resolution oblique streams is severely bottlenecked by the strict size, weight, and power constraints of autonomous aerial vehicles (AAVs). The computational density required to process these wide-field-of-view streams precludes low-latency inference on standard edge hardware. To address this, we propose temporal token reuse (TTR), an adaptive inference framework capable of accelerating video segmentation on embedded devices. TTR exploits the intrinsic spatiotemporal redundancy of aerial video by formulating image patches as tokens; it utilizes a lightweight similarity metric to dynamically identify static regions and propagate their precomputed deep features, thereby bypassing redundant backbone computations. We validate the framework on standard benchmarks and a newly curated Oblique Floodwater Dataset designed for hydrological monitoring. Experimental results on edge-grade hardware demonstrate that TTR achieves speedups ranging from 1.58× to 1.69× on video datasets (Floodwater, AAVid, Audi Autonomous Driving Dataset) with negligible degradation in segmentation accuracy (<1% mean intersection over union). On the image-based FloodNet benchmark, where temporal redundancy is inherently absent, TTR preserves full accuracy without degradation. These findings confirm that TTR effectively shifts the operational Pareto frontier for oblique video understanding, with the magnitude of efficiency gains scaling directly with the degree of temporal redundancy present in the input stream.