Building-PCC
收藏资源简介:
Building-PCC数据集是由荷兰代尔夫特理工大学创建,专注于城市建筑点云完成的基准数据集。该数据集包含从海牙和鹿特丹两城市收集的50,000个建筑实例,每个实例关联AHN3和AHN4两套机载点云数据及手工重建的3D建筑模型作为基准。数据集旨在评估深度学习方法在处理因遮挡、信号吸收等因素导致的点云不完整性方面的性能,特别适用于3D地理信息应用,如3D重建、语义理解和自动驾驶等领域的研究。
The Building-PCC dataset was created by Delft University of Technology in the Netherlands, serving as a benchmark dataset focused on urban building point cloud completion. It contains 50,000 building instances collected from The Hague and Rotterdam. Each instance is associated with two airborne point cloud datasets (AHN3 and AHN4) and manually reconstructed 3D building models as the benchmark. The dataset aims to evaluate the performance of deep learning methods in handling point cloud incompleteness caused by factors such as occlusion and signal absorption, and it is particularly suitable for research in 3D geospatial applications including 3D reconstruction, semantic understanding, autonomous driving and other fields.

- 1Building-PCC: Building Point Cloud Completion Benchmarks代尔夫特理工大学, 荷兰 · 2024年



