Towards Large-Scale City Reconstruction from Satellites

Duan, Liuyun; Lafarge, Florent

doi:10.1007/978-3-319-46454-1_6

Cited by 44 publications

(26 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These approaches start by detecting edge fragments and find an optimal cycle that links the edges into a coherent region. In [7], the authors propose a method that produces superpixels in the from of small polygons which they combine into object regions with the aim to label aerial images. In our work, we predict the polygon around the object directly, using a carefully designed RNN.…”

Section: Related Workmentioning

confidence: 99%

Annotating Object Instances with a Polygon-RNN

Castrejón¹,

Kundu²,

Urtasun³

et al. 2017

2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

304

270

View full text Add to dashboard Cite

We propose an approach for semi-automatic annotation of object instances. While most current methods treat object segmentation as a pixel-labeling problem, we here cast it as a polygon prediction task, mimicking how most current datasets have been annotated. In particular, our approach takes as input an image crop and sequentially produces vertices of the polygon outlining the object. This allows a human annotator to interfere at any time and correct a vertex if needed, producing as accurate segmentation as desired by the annotator. We show that our approach speeds up the annotation process by a factor of 4.7 across all classes in Cityscapes, while achieving 78.4% agreement in IoU with original ground-truth, matching the typical agreement between human annotators. For cars, our speed-up factor is 7.3 for an agreement of 82.2%. We further show generalization capabilities of our approach to unseen datasets.

show abstract

Section: Related Workmentioning

confidence: 99%

Annotating Object Instances with a Polygon-RNN

Castrejón¹,

Kundu²,

Urtasun³

et al. 2017

2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

304

270

View full text Add to dashboard Cite

show abstract

“…A few recent research works have proposed automated methodologies for urban scene reconstruction in LOD 1 from stereo pairs of high-resolution satellite images. The method of (Duan , Lafarge, 2016) has used a semi-global matching technique (Hirschmuller, 2008) to find correspondences in a stereo pair of epipolar images, followed by a joint classification using image radiometry coupled with estimated elevation information to retrieve 3D city models. Even though this method offered a solution for 3D urban reconstruction at a large scale, small geometries could not be captured precisely.…”

Section: Related Workmentioning

confidence: 99%

“…2. The method described in (Duan , Lafarge, 2016), see Section 2 for a summary. Figure 8 shows the results of these different methods using the same test image over Nairobi as in Section 3.…”

Section: 21mentioning

confidence: 99%

Automated Chain for Large-Scale 3d Reconstruction of Urban Scenes From Satellite Images

Tripodi¹,

Duan²,

Trastour³

et al. 2019

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Automatic city modeling from satellite imagery is a popular yet challenging topic in remote sensing, driven by numerous applications such as telecommunications, defence and urban mamagement. In this paper, we present an automated chain for large-scale 3D reconstruction of urban scenes with a Level of Detail 1 from satellite images. The proposed framework relies on two key ingredient. First, from a stereo pair of images, we estimate a digital terrain model and a digital height model, by using a novel set of feature descriptors based on multiscale morphological analysis. Second, inspired by recent works in machine learning, we extract in an automatic way contour polygons of buildings, by adopting a fully convolutional network U-Net followed by a polygonization of the predicted mask of buildings. We demonstrate the potential of our chain by reconstructing in an automated way different areas of the world.

show abstract

“…With the exception of satellite platforms (Duan & Lafarge, 2016;Toth & Jóźków, 2016), researchers regularly report submeter accuracy of 3D acquisition techniques (Jarząbek-Rychard & Borkowski, 2016;Kabolizade, Ebadi, & Mohammadzadeh, 2012;Mårtensson & Reshetyuk, 2016;Rottensteiner et al, 2014;Wang, Kutterer, & Fang, 2016). Hence, we consider positional accuracy in the range of 0-1 m in 10 cm increments resulting in 11 error classes.…”

Section: Perturbation and Grades Of Accuracymentioning

confidence: 99%

The effect of acquisition error and level of detail on the accuracy of spatial analyses

Biljecki

Heuvelink

Ledoux

et al. 2017

Cartography and Geographic Information Science

View full text Add to dashboard Cite

There has been a great deal of research about errors in geographic information and how they affect spatial analyses. A typical GIS process introduces various types of errors at different stages, and such errors usually propagate into errors in the result of a spatial analysis. However, most studies consider only a single error type thus preventing the understanding of the interaction and relative contributions of different types of errors. We focus on the level of detail (LOD) and positional error, and perform a multiple error propagation analysis combining both types of error. We experiment with three spatial analyses (computing gross volume, envelope area, and solar irradiation of buildings) performed with procedurally generated 3D city models to decouple and demonstrate the magnitude of the two types of error, and to show how they individually and jointly propagate to the output of the employed spatial analysis. The most notable result is that in the considered spatial analyses the positional error has a much higher impact than the LOD. As a consequence, we suggest that it is pointless to acquire geoinformation of a fine LOD if the acquisition method is not accurate, and instead we advise focusing on the accuracy of the data. ARTICLE HISTORY

show abstract

Towards Large-Scale City Reconstruction from Satellites

Cited by 44 publications

References 29 publications

Annotating Object Instances with a Polygon-RNN

Annotating Object Instances with a Polygon-RNN

Automated Chain for Large-Scale 3d Reconstruction of Urban Scenes From Satellite Images

The effect of acquisition error and level of detail on the accuracy of spatial analyses

Contact Info

Product

Resources

About