DVIO: An Optimization-Based Tightly Coupled Direct Visual-Inertial Odometry

In this article, we propose a novel optimization-based tightly coupled Direct Visual-Inertial Odometry (DVIO), which fuses the visual and inertial measurements to provide real-time full state estimation. Different from existing frameworks, the key novelty of the proposed method is to integrate the data association, state estimation, and outlier detection into a nonlinear optimization framework in a tightly coupled way. Specifically, by jointly minimizing the preintegration error of the inertial measurement unit and the photometric error of the camera, the data association is tightly coupled with the process of state estimation. Then, an iterative selection strategy is design to reject outliers during the data association and establish more visual constraints in the optimization. In addition, a hybrid weighting method is proposed to tightly integrate the iterative selection into the optimization by dynamically weighting residual terms. As a consequence, the proposed method aligns the image patches, estimates the motion and removes the outliers synchronously. Comparative experiments on the public dataset and extensive real-world experiments show that DVIO outperforms state-of-the-art visual-inertial odometries in terms of both the accuracy and the robustness. Thus, DVIO is highly applicable to the navigation or the simultaneous localization and mapping of mobile devices or agile robots like micro air vehicles.

I. Introduction

Visual-Inertial navigation systems (VINS), which fuse the measurements from visual and inertial sensors, have gained significant attentions in the field of robotics [1]–[18]. Unlike the vision-only systems, the monocular VINS can recover the metric scale in addition to providing a more accurate and high-bandwidth estimation of the translational and rotational velocities. Accordingly, the monocular VINS has been widely applied in various robotic systems such as ground vehicles and aerial robots [19]–[22]. However, it is extremely challenging to design a real-time, accurate, and robust VINS due to the fact that existing frameworks have to make a compromise between the accuracy and robustness [6]–[8], [11], [23], which limits their applications in challenging environments. Thus, a real-time, accurate, and robust VINS is still expected to be further explored.