RANSAC
PulseAugur coverage of RANSAC — every cluster mentioning RANSAC across labs, papers, and developer communities, ranked by signal.
-
新方法提高点云配准的准确性和效率
研究人员开发了一种新的点云配准算法，该算法利用概率性自更新局部对应关系和线矢量集来提高准确性和效率。该方法采用双RANSAC交互模型和全局提前终止条件来平衡性能。评估显示，与现有技术相比，均方根误差和时间效率有了显著提高，并附带C++源代码。
-
BIMStruct3D pipeline automates building information model generation from 3D scans
Researchers have developed BIMStruct3D, a novel pipeline that automates the creation of Building Information Models (BIM) from 3D point cloud data. This hybrid approach integrates learning-based semantic segmentation wi…
-
NONSAC framework offers scalable, robust model estimation for large datasets
Researchers have developed NONSAC, a novel framework designed for robust and scalable model estimation from extremely large datasets that contain noise and outliers. This method involves sampling non-minimal data subset…
-
高斯混合描述符改进了用于对象重建的3D片段匹配
研究人员引入了一种名为高斯混合描述符（GMD）的新方法，用于对象重建任务中的3D片段匹配。该描述符利用高斯混合模型分析和描述断裂表面上的点分布，从而能够更准确地识别相邻片段。该方法包括分割局部表面块，估计凹陷和凸起区域的GMM参数，然后合并这些区域描述符。GMD之间的相似性使用L2距离进行测量，片段对齐由RANSAC和ICP算法实现。