Zero-Parameter Geometric Gating for Temporally Stable Low-Altitude UAV Video Semantic Segmentation

2026-06-08

AI Summary

This work addresses temporal inconsistency in semantic segmentation of low-altitude UAV videos, where optical flow introduces structured noise over planar regions. To mitigate this issue, the authors propose a zero-parameter geometric gating mechanism that dynamically selects between homography and optical flow for feature alignment on a 16×16 spatial grid, based on the RANSAC homography inlier ratio. The selected alignment is then fused with results from semantic similarity propagation. Relying solely on a median-threshold decision derived from RANSAC statistics, the method introduces no additional learnable parameters and imposes no extra computational burden on the backbone network. Evaluated on the UAVid dataset, it improves mIoU by 4.24–4.91% and boosts temporal consistency over planar regions from 62% to 92%.

Abstract

Video semantic segmentation for low-altitude UAVs requires temporal consistency, yet dense optical flow introduces spatially structured noise in the planar regions that dominate aerial imagery. We propose a zero-parameter geometric gate that uses RANSAC homography inlier ratios on a $16\times16$ spatial grid to route each region to either homography or optical flow warp before fusion via Semantic Similarity Propagation. The gate requires no learned parameters -- only a median-threshold binary decision on RANSAC statistics -- adding only 211K trainable parameters (the SSP fusion layer) to a frozen backbone. On synthetic UAVid, the method achieves +4.24--4.91\% mIoU improvement over base models across two architectures (SegFormer-b2 and Hiera-S+UPerNet). Mechanism diagnostics reveal that flow residuals in planar regions are spatially autocorrelated (Moran's I = 0.32, $p < 0.001$), predict boundary instability (Spearman $ρ= 0.66$), and that rigidification recovers temporal consistency from 62\% to 92\% (+29.5pp) in homography-valid regions.

Problem

• temporal consistency
• semantic segmentation
• low-altitude UAV
• optical flow
• spatially structured noise

Innovation

• zero-parameter gating
• temporal consistency
• RANSAC homography
• semantic similarity propagation
• UAV video segmentation