Real-Time Branch-to-Tool Distance Estimation for Autonomous UAV Pruning: Benchmarking Five DEFOM-Stereo Variants from Simulation to Jetson Deployment

2026-03-27

AI Summary

This work addresses the need for real-time, high-precision distance estimation of fine branches in autonomous drone-based tree pruning by proposing DEFOM-PrunePlus, a lightweight stereo matching model. A high-fidelity synthetic dataset comprising 5,520 image pairs is constructed using Unreal Engine 5, incorporating ZED Mini virtual camera modeling and EXR-based dense depth supervision. The model is deployed on an NVIDIA Jetson Orin platform and achieves a depth mean absolute error (MAE) of 64.26 cm and a delta-1 accuracy of 87.59% at a typical operating distance of 2 meters, with an inference speed of 3.3 FPS. These results demonstrate a superior trade-off between accuracy and real-time performance compared to other lightweight models and Vision Transformer Small (ViT-S) variants, while also validating effective sim-to-real transfer capability.

Abstract

Autonomous tree pruning with unmanned aerial vehicles (UAVs) is a safety-critical real-world task: the onboard perception system must estimate the metric distance from a cutting tool to thin tree branches in real time so that the UAV can approach, align, and actuate the pruner without collision. We address this problem by training five variants of DEFOM-Stereo - a recent foundation-model-based stereo matcher - on a task-specific synthetic dataset and deploying the checkpoints on an NVIDIA Jetson Orin Super 16 GB. The training corpus is built in Unreal Engine 5 with a simulated ZED Mini stereo camera capturing 5,520 stereo pairs across 115 tree instances from three viewpoints at 2m distance; dense EXR depth maps provide exact, spatially complete supervision for thin branches. On the synthetic test set, DEFOM-Stereo ViT-S achieves the best depth-domain accuracy (EPE 1.74 px, D1-all 5.81%, delta-1 95.90%, depth MAE 23.40 cm) but its Jetson inference speed of ~2.2 FPS (~450 ms per frame) remains too slow for responsive closed-loop tool control. A newly introduced balanced variant, DEFOM-PrunePlus (~21M backbone, ~3.3 FPS on Jetson), offers the best deployable accuracy-speed trade-off (EPE 5.87 px, depth MAE 64.26 cm, delta-1 87.59%): its frame rate is sufficient for real-time guidance and its depth accuracy supports safe branch approach planning at the 2m operating range. The lightweight DEFOM-PruneStereo (~6.9 FPS) and DEFOM-PruneNano (~8.5 FPS) run fast but sacrifice substantial accuracy (depth MAE > 57 cm), making estimates too unreliable for safe actuation. Zero-shot inference on real photographs confirms that full-capacity models preserve branch geometry, validating the sim-to-real transfer. We conclude that DEFOM-PrunePlus provides the most practical accuracy-latency balance for onboard distance estimation, while ViT-S serves as the reference for future hardware.

Problem

• autonomous UAV pruning
• real-time distance estimation
• branch-to-tool proximity
• stereo depth perception
• safety-critical navigation

Innovation

• DEFOM-Stereo
• real-time depth estimation
• autonomous UAV pruning
• sim-to-real transfer
• embedded deployment