Abstract:Accurate and robust place recognition is essential for robot navigation, yet achieving full pose invariance and high performance across diverse scenes remains challenging. In this work, we propose a novel global and local combined descriptor named binary triangle combined (BTC) descriptor. We first extract the keypoints of a point cloud by projecting the points to planes extracted therein. Any three keypoints form a unique triangle, with the lengths of its sides constituting a triangle descriptor that captures the global appearance of the point cloud. Thanks to the distinct shape of a triangle given three side lengths, the similarity between two triangles and their vertices (i.e., keypoints) correspondence can be naturally determined from the side lengths of the triangle descriptors. The matched triangle pairs evaluate the appearance similarity between two point clouds, while the vertices' correspondence enables accurate estimation of their relative pose; both are crucial for the place recognition task. To enhance the accuracy of triangle matching, BTC introduces a binary descriptor, which describes the point distribution neighboring each keypoint. The local geometry information encoded by the binary descriptor augments descriptiveness and discriminativeness to the triangle descriptor. Collectively, the two descriptors achieve both global and local descriptions of the environment with high accuracy, efficiency, and robustness. We extensively compare the proposed BTC descriptor against state-of-the-art methods (e.g., Scan Context, LCD-Net) on a wide range of datasets collected using different types of LiDAR sensors (spinning LiDARs and nonrepetitive scanning LiDARs) in various environments (urban, campus, forest, park, and mountain). The quantitative results demonstrate that BTC exhibits greater adaptability and significant improvement in precision compared to its counterparts, especially in challenging cases with large viewpoint variations (e.g., reverse direction, large translation, and/or rotation).

BRoPH: An efficient and compact binary descriptor for 3D point clouds

BRoPH: A compact and efficient binary 3D feature descriptor

Rpbs: Rotational Projected Binary Structure For Point Cloud Representation

BTC: A Binary and Triangle Combined Descriptor for 3-D Place Recognition

An Efficient and Compact 3D Local Descriptor Based on the Weighted Height Image.

KDD: A kernel density based descriptor for 3D point clouds

3D Binary Signatures

Spacetarget Pose Estimation Based on Binary Rotational Projection Histogram

Efficient and accurate registration with BWPH descriptor for low-quality point clouds

A Hybrid Compression Framework for Color Attributes of Static 3D Point Clouds

PPTFH: Robust Local Descriptor Based on Point-Pair Transformation Features for 3D Surface Matching

QBB: Quantile-Based Binarization of 3D Point Cloud Descriptors

I-BRIEF: A Fast Feature Point Descriptor with More Robust Features

Novel 3D local feature descriptor of point clouds based on spatial voxel homogenization for feature matching

Embrace descriptors that use point pairs feature

CBDF: Compressed Binary Discriminative Feature.

Matching Algorithm for 3D Point Cloud Recognition and Registration Based on Multi-Statistics Histogram Descriptors

Robust Kernel-based Feature Representation for 3D Point Cloud Analysis via Circular Convolutional Network

A Comprehensive Performance Evaluation of 3D Local Feature Descriptors

HoPPF: A novel local surface descriptor for 3D object recognition

A fast and robust local descriptor for 3D point cloud registration