Building instance segmentation;
roof plane instance segmentation;
building vectorization;
semantic segmentation;
3D reconstruction;
D O I:
10.5194/isprs-annals-X-1-W1-2023-971-2023
中图分类号:
K85 [文物考古];
学科分类号:
0601 ;
摘要:
Deep learning is a powerful tool to extract both individual building and roof plane polygons. But deep learning requires a large amount of labeled data. Hence, publicly available level of detail (LoD)-2 datasets are a natural choice to train fully convolutional neural networks (FCNs) models for both building section and roof plane instance segmentation. Since publicly available datasets are often automatically derived, e.g. based on laser scanning, they lack on annotation accuracy. To complement such a dataset, we introduce manually annotated and synthetically generated data. Manually annotated data is domain-specific and has a high annotation quality but is expensive to obtain. Synthetically generated data has high-quality annotations by definition, but lacks domain-specificity. Moreover, we not only detect individual building section instances, but also roof plane instances. We predict separations not only between individual buildings, but also by a class that describes the line which separates roof planes. The predicted building and roof plane instances are polygonized by a simple tree search algorithm. To achieve more regular polygons, we utilize the Douglas-Peucker polygon simplification algorithm. We describe our dataset in detail to allow comparability between successive methods. To facilitate future works in building and roof plane prediction, our Roof3D dataset is accessible at https://github.com/dlrPHS/GPUB.
机构:
Amirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, IranAmirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, Iran
Karbalaali, Haleh
Javaheriani, Abdolrahim
论文数: 0引用数: 0
h-index: 0
机构:
Amirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, Iran
Univ Tehran, Inst Geophys, Tehran 14359444111, IranAmirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, Iran
Javaheriani, Abdolrahim
Dahlke, Stephan
论文数: 0引用数: 0
h-index: 0
机构:
Philipps Univ, Dept Math & Comp Sci, D-35032 Marburg, GermanyAmirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, Iran
Dahlke, Stephan
Reisenhofer, Rafael
论文数: 0引用数: 0
h-index: 0
机构:
Univ Bremen, AG Computat Data Anal, D-28359 Bremen, GermanyAmirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, Iran
Reisenhofer, Rafael
Torabi, Siyavash
论文数: 0引用数: 0
h-index: 0
机构:
Dana Geophys Co, Seism Data Proc Dept, Tehran 1919935331, IranAmirkabir Univ Technol, Dept Petr Engn, Tehran 158754413, Iran