analiza aiqbench
This research addresses the critical gap of adapting state-of-the-art foundation models like Depth Anything 3 to the specific spatially structured noise inherent in analog FPV drone video transmission. By utilizing a real-noise bank and parameter-efficient fine-tuning techniques such as LoRA, the proposed AnalogDepth pipeline significantly outperforms standard digital noise baselines in depth estimation accuracy. The work underscores that effective AI adaptation for robotics applications requires training data that faithfully replicates the physical degradation mechanisms of the target deployment environment rather than relying on synthetic approximations.
AnalogDepth: Multi-view Geometry from FPV drones under Analog Video Transmission
arXiv:2609.24312v1 Announce Type: cross Abstract: Analog video transmission (VTX) remains widespread in FPV drones due to low latency, weight and low cost. However analog VTX suffers from complex spatially structured image degradation which differ fundamentally from digital image corruption (e.g. AWGN) used in standard training augmentation. This work shows that this type of noise severely degrades the accuracy of Depth Anything 3 (DA3), a state-of-the-art feed forward visual geometry foundation model. To address this gap, we present AnalogDepth, a parameter-efficient training pipeline that adapts DA3 to analog FPV imagery using student-teacher knowledge distillation with Low-Rank Adaptation (LoRA) injected into the DINOv2 backbone. Rather than synthesizing noise analytically, we build a noise bank from static FPV recordings under diverse conditions and compare real-noise injection against PSD-matched Gaussian synthesis and AWGN as baselines. Experiments on six real FPV flight sequences across three indoor scenes show that training with our noise bank consistently reduces per-frame depth RMSE and 3D reconstruction Chamfer distance compared to the pretrained DA3 baseline and both Gaussian noise variants. These results demonstrate that replicating the spatial structure of real analog transmission noise is critical for effective adaptation.
Abstract:Analog video transmission (VTX) remains widespread in FPV drones due to low latency, weight and low cost. However analog VTX suffers from complex spatially structured image degradation which differ fundamentally from digital image corruption (e.g. AWGN) used in standard training augmentation. This work shows that this type of noise severely degrades the accuracy of Depth Anything 3 (DA3), a state-of-the-art feed forward visual geometry foundation model. To address this gap, we present AnalogDepth, a parameter-efficient training pipeline that adapts DA3 to analog FPV imagery using student-teacher knowledge distillation with Low-Rank Adaptation (LoRA) injected into the DINOv2 backbone. Rather than synthesizing noise analytically, we build a noise bank from static FPV recordings under diverse conditions and compare real-noise injection against PSD-matched Gaussian synthesis and AWGN as baselines. Experiments on six real FPV flight sequences across three indoor scenes show that training with our noise bank consistently reduces per-frame depth RMSE and 3D reconstruction Chamfer distance compared to the pretrained DA3 baseline and both Gaussian noise variants. These results demonstrate that replicating the spatial structure of real analog transmission noise is critical for effective adaptation.
Computer Vision and Pattern Recognition (cs.CV); Robotics (cs.RO)
arXiv:2609.24312 [cs.CV]
(or arXiv:2609.24312v1 [cs.CV] for this version)
https://doi.org/10.48550/arXiv.2609.24312 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 282 cuvinte
Subiecte conexe
Skydio CEO Adam Bry on why Silicon Valley shouldn’t draw red lines for drone use
The Verge · 15 iun. 2026
Neglected Basics of AI Alignment
LessWrong · 7 iun. 2026
IEEE Celebrates Innovators Shaping the Future
IEEE Spectrum · 30 sept. 2026
SAM-V: Geometry-Aware Segment Anything for Multi-View Instance Segmentation
arXiv · 23 sept. 2026