Efficient Feature-Free Initialization for Monocular Visual-Inertial Systems Using a Feed-Forward 3D Model

2026-05-17

AI Summary

This work addresses the challenge of slow and unreliable initialization in monocular visual-inertial navigation systems under texture-poor or visually degraded conditions, where conventional methods rely heavily on feature matching and typically require 3–4 seconds with high failure rates. The paper presents the first feature-free initialization framework, which leverages a feedforward 3D model to directly predict scale-ambiguous point clouds from a single image. By integrating IMU preintegration, scale alignment, and state estimation into an end-to-end pipeline, the proposed method eliminates dependence on visual features. It achieves initialization in under 1.2 seconds with over 90% success rate, substantially improving both speed and robustness—particularly in environments where visual cues are scarce or unreliable.

Abstract

Fast and reliable initialization is critical for monocular visual-inertial navigation systems (VINS), as it establishes the starting conditions for subsequent state estimation. Despite steady progress, most existing methods heavily rely on visual feature correspondences and require 3-4 seconds of sensory data for successful initialization, which limits their applicability and efficiency. With the advent of feed-forward 3D models that can directly predict point clouds from images, we revisit the visual-inertial initialization problem from a concise perspective. In this work, we propose a feature-free initialization framework that leverages up-to-scale point clouds predicted by a feed-forward 3D model, thereby obviating the need for visual feature tracking and estimation. This design substantially reduces system complexity and improves the reliability of initialization. Experiments on public datasets demonstrate that the proposed feature-free initialization method achieves the highest success rate, exceeding 90%, and significantly reduces the data duration required for successful initialization, typically to under 1.2 s. We further validate our method on a self-collected dataset covering various indoor and outdoor scenarios, demonstrating robust performance, particularly in visually degraded environments where existing methods often fail. The code and dataset are available at https://github.com/Yuantai-Z/FF-VIO-Init.

Problem

• monocular visual-inertial navigation
• initialization
• feature-free
• visual-inertial systems
• fast initialization

Innovation

• feature-free initialization
• feed-forward 3D model
• monocular visual-inertial navigation
• point cloud prediction
• fast initialization