BEVLCD+: Real-Time and Rotation-Invariant Loop Closure Detection Based on BEV of Point Cloud

The closed-loop constraint provided by loop closure detection (LCD) plays an essential role in eliminating accumulated errors of simultaneous localization and mapping (SLAM) systems. BEVLCD+: a real-time and rotation-invariant LCD method based on a bird's eye view (BEV) of a point cloud has been proposed in this article. Projecting a 3-D point cloud to a 2-D BEV allows for a remarkable compression of data, resulting in improved processing efficiency using a convolutional neural network (CNN) and facilitating the integration of point-level verification of loops through spatial information efficiently. Traditional convolution and pooling modules are improved to extract rotation-invariant features, ensuring that our BEVLCD+ method is rotation-invariant. Global descriptors aggregated by NetVLAD are used for fast retrieval, and local features are used for accurate verification. Local feature matching and RANSAC-based singular value decomposition (SVD) provide the 3 degrees of freedom (DoF) pose of loop pairs, which include translation and yaw in the horizontal direction. Experiments on benchmark datasets prove the excellent performance of our method.