Visual Computing Research Center
Shenzhen University, China
: We present UrbanScene3D, a large-scale data platform for research of urban scene perception and reconstruction. UrbanScene3D contains over 128k high-resolution images covering 16 scenes including large-scale real urban regions and synthetic cities with 136 km2 area in total. The dataset also contains high-precision LiDAR scans and hundreds of image sets with different observation patterns, which provide a comprehensive benchmark to design and evaluate aerial path planning and 3D reconstruction algorithms. In addition, the dataset, which is built on Unreal Engine and Airsim simulator together with the manually annotated unique instance label for each building in the dataset, enables the generation of all kinds of data, e.g., 2D depth maps, 2D/3D bounding boxes, and 3D point cloud/mesh segmentations, etc. The simulator with physical engine and lighting system not only produce a variety of data but also enable users to simulate cars and drones in the proposed urban environment for future research.
Download

paper data code supp

UrbanScene3D (1.43TB) is publicly accessible for non-commercial uses only. Permission is granted to use the data only if you agree:
  • - The dataset is provided "AS IS". Despite our best efforts to assure accuracy, we disclaim all liability for any mistakes or omissions;
  • - All works that utilize this dataset including any partial use must cite our paper provided here;
  • - You refrain from disseminating this dataset or any altered variations;
  • - You are not permitted to utilize this dataset or any derivative work for any commercial endeavors;
  • - We reserve all rights that are not explicitly granted to you.

Choose the way of data downloading (Google Drive by default)

Simulator

63GB
Suzhou
New York
Shanghai
San Francisco
Shenzhen
Chicago
campus
residence
square
hospital

Benchmark

virtual scenes

School

Town

Bridge

Castle

Planning Method Overlapping Rate Proxy Path Image Reconstruction
49MB
964KB
390GB
16GB
Oblique Photography N/A N/A
69.2KB
7GB
140MB
Smith et al. 2018 90% box
184KB
71GB
4GB
coarse
inter
fine
70% box
185KB
65GB
4GB
coarse
inter
fine
Zhou et al. 2020 90% box
169KB
99GB
3GB
coarse
inter
fine
70% box
80KB
47GB
3GB
coarse
inter
fine
Zhang et al. 2021 90% box
170KB
64GB
1GB
coarse
inter
fine
70% box
106KB
38GB
1GB
coarse
inter
fine
Planning Method Overlapping Rate Proxy Path Image Reconstruction
28MB
576KB
205GB
9GB
Oblique Photography N/A N/A
46KB
2GB
41GB
Smith et al. 2018 90% box
141KB
51GB
3GB
coarse
inter
fine
70% box
55KB
20GB
2GB
coarse
inter
fine
Zhou et al. 2020 90% box
105KB
55GB
2GB
coarse
inter
fine
70% box
44KB
22GB
2GB
coarse
inter
fine
Zhang et al. 2021 90% box
127KB
39GB
537MB
coarse
inter
fine
70% box
58KB
16GB
427MB
coarse
inter
fine
Planning Method Overlapping Rate Proxy Path Image Reconstruction
31MB
848KB
293GB
13GB
Oblique Photography N/A N/A
69KB
5GB
113MB
Smith et al. 2018 90% box
150KB
50GB
3GB
coarse
inter
fine
70% box
150KB
48GB
3GB
coarse
inter
fine
Zhou et al. 2020 90% box
131KB
64GB
2GB
coarse
inter
fine
70% box
77KB
39GB
2GB
coarse
inter
fine
Zhang et al. 2021 90% box
165KB
49GB
1GB
coarse
inter
fine
70% box
103KB
33GB
1GB
coarse
inter
fine
Planning Method Overlapping Rate Proxy Path Image Reconstruction
17M
472KB
182GB
8GB
Oblique Photography N/A N/A
51KB
3GB
32MB
Smith et al. 2020 90% box
86KB
32GB
2GB
coarse
inter
fine
70% box
86KB
28GB
2GB
coarse
inter
fine
Smith et al. 2018 90% box
64KB
39GB
2GB
coarse
inter
fine
70% box
45KB
32GB
2GB
coarse
inter
fine
Zhou et al. 2020 90% box
80KB
29GB
427MB
coarse
inter
fine
70% box
61KB
20GB
411MB
coarse
inter
fine

real scenes

PolyTech

ArtSci

Planning Method Overlapping Rate Proxy Path + image Reconstruction
27MB
86GB
2GB
Oblique Photography N/A N/A
24GB
169MB
Smith et al. 2018 N/A box
15GB
468MB
coarse
inter
fine
Zhou et al. 2020 N/A box
29GB
944MB
coarse
inter
fine
Zhang et al. 2021 N/A box
17GB
465MB
coarse
inter
fine
Planning Method Overlapping Rate Proxy Path + image Reconstruction
8M
186GB
16GB
Oblique Photography N/A N/A
35GB
497MB
Smith et al. 2018 N/A box
38GB
2GB
coarse
inter
fine
Zhou et al. 2020 N/A box
43GB
7GB
coarse
inter
fine
Zhang et al. 2021 N/A box
70GB
6GB
coarse
inter
fine

evaluation program

190MB

demo