Fully Polar Coordinate Object Detection: A Constraint-Based Polar Bounding Box Approach for LiDAR and Scanning Radar
ShuHeng Lin, Chieh-Chih Wang, Wen-Chieh Lin
AI summary
Problem
Existing polar-based detection methods struggle with feature distortion and rely on unsuccessful undistortion techniques, while Cartesian grids ignore the natural non-uniform sparsity of range sensor data.
Approach
The authors propose a fully polar detection pipeline that replaces Cartesian ground truth with a novel constraint-based polar bounding box representation, incorporates polar-specific data augmentation, and applies an azimuthal shift to maintain continuity during training and evaluation.
Key results
- Constraint-based polar bounding box representation enabling direct polar training and evaluation
- +2.88% AP30 improvement over state-of-the-art polar detector on LiDAR
- +2.17% AP30 gain over Cartesian-based detection methods on LiDAR
- +13.11% AP30 improvement over Cartesian methods on scanning radar
Why it matters
Validates fully polar pipelines for autonomous vehicle perception, offering robust long-range detection and better computational efficiency for LiDAR and radar systems.
Abstract
Polar coordinates are widely used in segmentation tasks for range sensors such as LiDAR and radar, owing to their ability to naturally align with point cloud sparsity and distribution. However, their use in detection is limited by feature distortion. Existing polar-based detection works focused on undistorting features from the polar coordinates back to canonical Cartesian representations, but their results remain unsuccessful. In this work, we propose fully polar coordinate object detection, performing training and evaluation entirely in polar coordinates without relying on Cartesian metrics. To achieve this, we design a constraint-based polar bounding box representation, that enables the direct conversion of Cartesian bounding boxes via a constrained minimum bounding rectangle (MBR). Using the state-of-the-art polar-based detector as our baseline, we conduct experiments on the Boreas dataset. The results demonstrate that our approach improves the LiDAR detection AP30 metric by 2.88%, and yields a 2.17% gain over Cartesian-based detection methods. On more challenging scanning radar detection experiments, our method achieves an 13.11% improvement in AP30 compared to Cartesian-based detection methods. These findings validate the feasibility of fully polar coordinate object detection and demonstrate its robustness and generalizability across multiple range sensor modalities.