Jing Li

In this paper, we propose a deep learning framework, namely AFGL-Net to achieve building façade parsing, i.e., obtaining the semantics of small components of building façade, such as windows and doors. To this end, we present an autoencoder embedding position and direction encoding for local feature encoding. The autoencoder enhances the local feature aggregation and augments the representation of skeleton features of windows and doors. We also integrate the Transformer into AFGL-Net to infer the geometric shapes and structural arrangements of façade components and capture the global contextual features. These global features can help recognize inapparent windows/doors from the façade points corrupted with noise, outliers, occlusions, and irregularities. The attention-based feature fusion mechanism is finally employed to obtain more informative features by simultaneously considering local geometric details and the global contexts. The proposed AFGL-Net is comprehensively evaluated on Dublin and RueMonge2014 benchmarks, achieving 67.02% and 59.80% mIoU, respectively. We also demonstrate the superiority of the proposed AFGL-Net by comparing with the state-of-the-art methods and various ablation studies.

This paper proposes a building façade contouring method from LiDAR (Light Detection and Ranging) scans and photogrammetric point clouds. To this end, we calculate the confidence property at multiple scales for an individual point cloud to measure the point cloud’s quality. The confidence property is utilized in the definition of the gradient for each point. We encode the individual point gradient structure tensor, whose eigenvalues reflect the gradient variations in the local neighborhood areas. The critical point clouds representing the building façade and rooftop (if, of course, such rooftops exist) contours are then extracted by jointly analyzing dual-thresholds of the gradient and gradient structure tensor. Based on the requirements of compact representation, the initial obtained critical points are finally downsampled, thereby achieving a tradeoff between the accurate geometry and abstract representation at a reasonable level. Various experiments using representative buildings in Semantic3D benchmark and other ubiquitous point clouds from ALS DublinCity and Dutch AHN3 datasets, MLS TerraMobilita/iQmulus 3D urban analysis benchmark, UAV-based photogrammetric dataset, and GeoSLAM ZEB-HORIZON scans have shown that the proposed method generates building contours that are accurate, lightweight, and robust to ubiquitous point clouds. Two comparison experiments also prove the superiority of the proposed method in terms of topological correctness, geometric accuracy, and representation compactness.