Fully automated structured light scanning for high-fidelity 3D reconstruction via graph optimization.

Convenient and high-fidelity 3D model reconstruction is crucial for industries like manufacturing, medicine and archaeology. Current scanning approaches struggle with high manual costs and the accumulation of errors in large-scale modeling. This paper is dedicated to achieving industrial-grade seamless and high-fidelity 3D reconstruction with minimal manual intervention. The innovative method proposed transforms the multi-frame registration into a graph optimization problem, addressing the issue of error accumulation encountered in frame-by-frame registration. Initially, a global consistency cost is established based on point cloud cross-multipath registration, followed by using the geometric and color differences of corresponding points as dynamic nonlinear weights. Finally, the iteratively reweighted least squares (IRLS) method is adopted to perform the bundle adjustment (BA) optimization of all poses. Significantly enhances registration accuracy and robustness under the premise of maintaining near real-time efficiency. Additionally, for generating watertight, seamless surface models, a local-to-global transitioning strategy for multiframe fusion is introduced. This method facilitates efficient correction of normal vector consistency, addressing mesh discontinuities in surface reconstruction resulting from normal flips. To validate our algorithm, we designed a 3D reconstruction platform enabling spatial viewpoint transformations. We collected extensive real and simulated model data. These datasets were rigorously evaluated against advanced methods, roving the effectiveness of our approach. Our data and implementation is made available on GitHub for community development.

Extrinsic calibration for multi-LiDAR systems involving heterogeneous laser scanning models.

In the realm of autonomous driving, there is a pressing demand for heightened perceptual capabilities, giving rise to a plethora of multisensory solutions. Among these, multi-LiDAR systems have gained significant popularity. Within the spectrum of available combinations, the integration of repetitive and non-repetitive LiDAR configurations emerges as a balanced approach, offering a favorable trade-off between sensing range and cost. However, the calibration of such systems remains a challenge due to the diverse nature of point clouds, low-common-view, and distinct densities. This study proposed a novel targetless calibration algorithm for extrinsic calibration between Hybrid-Solid-State-LiDAR(SSL) and Mechanical-LiDAR systems, each employing different scanning modes. The algorithm harnesses planar features within the scene to construct matching costs, while proposing the adoption of the Gaussian Mixture Model (GMM) to address outliers, thereby mitigating the issue of overlapping points. Dynamic trust-region-based optimization is incorporated during iterative processes to enhance nonlinear convergence speed. Comprehensive evaluations across diverse simulated and real-world scenarios affirm the robustness and precision of our algorithm, outperforming current state-of-the-art methods.