A Robust and Efficient Loop Closure Detection Approach for Hybrid Ground/Aerial Vehicles

Abstract

Frequent and dramatic viewpoint changes make loop closure detection of hybrid ground/aerial vehicles extremely challenging. To address this issue, we present a robust and efficient loop closure detection approach based on the state-of-the-art simultaneous localization and mapping (SLAM) framework and pre-trained deep learning models. First, the outputs of the SuperPoint network are processed to extract both tracking features and additional features used in loop closure. Next, binary-encoded SuperPoint descriptors are applied with a method based on Bag of VisualWords (BoVW) to detect loop candidates efficiently. Finally, the combination of SuperGlue and SuperPointdescriptors provides correspondences of keypoints to verify loop candidates and calculate relative poses. The system is evaluated on the public datasets and a real-world hybrid ground/aerial vehicles dataset. The proposed approach enables reliable loop detection, even when the relative translation between two viewpoints exceeds 7m or one of the Euler angles is above 50°.

Real-Time Rubber Quality Model Based on Cnn-Lstm Deep Learning Theory

“…1) Monocular-inertial tests: In the EuRoc dataset tests, we selected several SOTA SLAM methods as contrast objects, including traditional visual-inertial SLAM methods like VI-DSO [24], VI-ORB [25], OKVIS [26], VINS-Mono [11] , ORB-SLAM3 [12], the hybrid SLAM algorithm SP-Loop [7], and the SOTA end-to-end deep learning SLAM system DVI-SLAM [5]. We compared the trajectories from both quantitative and qualitative perspectives.…”

“…For instance, DX-Net arXiv:2405.03413v3 [cs.RO] 4 Jun 2024 [6] simply replaces ORB features with deep features but relies on traditional methods for tracking, leading to incoherence in the deep feature information. Similarly, SP-Loop [7] integrates deep learning features only into the loop closure module, retaining traditional feature extraction methods elsewhere. Consequently, there is a need for a versatile hybrid SLAM method that effectively integrates deep learning technology to address complex environmental challenges comprehensively.…”