A novel pipeline based on line-structured features and cross-camera perspectives for coarse aggregate point cloud registration

Yihui Shan, Wei Li, Ming Yang, Jiangang Ding, Yuanlin Zhao, Aojia Tian

Abstract: Highlights•Single perspective data cannot meet aggregate grading requirements.•Dual view data registration meets the requirements but requires high precision and efficiency.•Propose an efficient overlap recognition method to improve registration accuracy.•Propose an innovative line data approach to achieve accurate registration.•Experimentally proven methodology is balanced and suitable for real-time industrial use.