Flying to Image-Specified Objects: 3D Quadrotor Navigation via Cross-Graph Memory and Viewpoint Planning

2026-06-29

AI Summary

This work addresses the challenge of enabling quadrotor drones to navigate precisely to a specific object instance in 3D environments using only a target image, under constraints such as continuous control, limited field of view, and safety requirements. The authors propose a hierarchical navigation framework that decouples high-level decision-making from low-level motion control. By introducing a viewpoint-aware action node mechanism, the method generates semantically guided action candidates around frontier regions and potential targets, while a lightweight cross-map semantic memory module facilitates effective propagation of semantic information to decision nodes. This approach jointly optimizes object recognition and navigation policy and integrates a dynamically feasible 3D trajectory planner. Extensive simulations demonstrate significant improvements over strong baselines, and real-world flight experiments confirm the system’s practicality and robustness.

Abstract

Instance-Specific Image-Goal Navigation (InstanceImageNav) requires a robot to navigate toward the exact object instance depicted in a query image. Extending this task to quadrotors is challenging due to continuous 3D control, limited field of view (FOV), and safety constraints, which make successful navigation highly dependent on selecting informative viewpoints. We propose a hierarchical navigation framework for quadrotor InstanceImageNav that separates high-level decision making from low-level motion execution. Instead of navigating directly to spatial locations, the system generates viewpoint-aware action nodes around frontier regions and potential target objects, enabling the robot to explore while maintaining informative viewpoints for detecting the target instance. A lightweight semantic memory maintains object-level and observation-level context, allowing semantic cues to propagate to candidate action nodes for decision making. A learning-based policy selects the most promising action node, and a trajectory planner generates dynamically feasible 3D flight paths for safe execution. Experiments in simulation demonstrate consistent improvements over strong baselines, and real-world quadrotor flights validate the practicality and robustness of the proposed framework.

Problem

• InstanceImageNav
• quadrotor navigation
• 3D navigation
• viewpoint planning
• image-specified objects

Innovation

• InstanceImageNav
• viewpoint planning
• cross-graph memory
• hierarchical navigation
• quadrotor navigation