Filling SRTM Void Data Via Conditional Adversarial Networks

Dong, Guoshuai; Chen, Fang; Ren, Peng

doi:10.1109/igarss.2018.8518992

Cited by 13 publications

(7 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The baseline CGAN model has great potential in restoring SRTM data because of its effective representational power. Our preliminary study (Dong et al, 2018) has validated its effectiveness over the interpolation based methods. However, the baseline CGAN has limitations in void filling.…”

Section: Limitationsmentioning

confidence: 84%

See 1 more Smart Citation

A shadow constrained conditional generative adversarial net for SRTM data restoration

Dong

Huang

Smith

et al. 2020

Remote Sensing of Environment

Self Cite

View full text Add to dashboard Cite

The original data produced by the Shuttle Radar Topography Mission (SRTM) tend to have an abundance of voids in mountainous areas where the elevation measurements are missing. In this paper, deep learning models are investigated for restoring SRTM data. To this end, we explore generative adversarial nets, which represent one state of the art family of deep learning models. A conditional generative adversarial network (CGAN) is introduced as the baseline method for filling voids in incomplete SRTM data. The problem regarding shadow violation that possibly arises from the CGAN restored data is investigated. To address this deficiency, shadow geometric constraints based on shadow maps of satellite images are devised. In addition, a shadow constrained conditional generative adversarial network (SCGAN), which incorporates the shadow geometric constraints into the CGAN, is developed. Training the SCGAN model requires both the remote sensing observations (i.e., the original incomplete SRTM data and satellite images) and the ground truth data (i.e., the complete SRTM data, which are manually refined from the incomplete SRTM data with the reference of in-situ measurements). The integration of the multi-source training data enables the SCGAN model to be characterized by comprehensive information including both mountain shape variation and mountain shadow geometry. Experimental results validate the superiority of the SCGAN over the comparison methods, i.e., the interpolation, the convolutional neural network (CNN) and the baseline CGAN, in SRTM data restoration.

show abstract

Section: Limitationsmentioning

confidence: 84%

“…In addition, the complete SRTM data are taken as the target outputs of the generator of the CGAN and thus the generator and discriminator are adversarially trained in a supervised fashion. The void filling CGAN was previously investigated in our preliminary work (Dong et al, 2018). It is considered as one baseline method in this paper.…”

Section: Introductionmentioning

confidence: 99%

A shadow constrained conditional generative adversarial net for SRTM data restoration

Dong

Huang

Smith

et al. 2020

Remote Sensing of Environment

Self Cite

View full text Add to dashboard Cite

show abstract

“…Although the methods using interpolation techniques can fill DEM voids well, it heavily relies on the DEM's local features, such as void size, terrestrial slopes, etc. [30].…”

Section: Introductionmentioning

confidence: 99%

Voids Filling of DEM with Multiattention Generative Adversarial Network Model

et al. 2022

View full text Add to dashboard Cite

The digital elevation model (DEM) acquired through photogrammetry or LiDAR usually exposes voids due to phenomena such as instrumentation artifact, ground occlusion, etc. For this reason, this paper proposes a multiattention generative adversarial network model to fill the voids. In this model, a multiscale feature fusion generation network is proposed to initially fill the voids, and then a multiattention filling network is proposed to recover the detailed features of the terrain surrounding the void area, and the channel-spatial cropping attention mechanism module is proposed as an enhancement of the network. Spectral normalization is added to each convolution layer in the discriminator network. Finally, the training of the model by a combined loss function, including reconstruction loss and adversarial loss, is optimized. Three groups of experiments with four different types of terrains, hillsides, valleys, ridges and hills, are conducted for validation of the proposed model. The experimental results show that (1) the structural similarity surrounding terrestrial voids in the three types of terrains (i.e., hillside, valley, and ridge) can reach 80–90%, which implies that the DEM accuracy can be improved by at least 10% relative to the traditional interpolation methods (i.e., Kriging, IDW, and Spline), and can reach 57.4%, while other deep learning models (i.e., CE, GL and CR) only reach 43.2%, 17.1% and 11.4% in the hilly areas, respectively. Therefore, it can be concluded that the structural similarity surrounding the terrestrial voids filled using the model proposed in this paper can reach 60–90% upon the types of terrain, such as hillside, valley, ridge, and hill.

show abstract

“…Conditional GAN (CGAN) refers to a GAN with conditional constraints, which is used to guide network training given the corresponding label. The literature [21,22] used CGAN and the improved CGAN model to analyze the structural expression of spatial interpolation. However, CGAN still has the same defects as GAN, such as the disappearance of gradients.…”

Section: Introductionmentioning

confidence: 99%

DEM Void Filling Based on Context Attention Generation Model

Zhang

Shi

et al. 2020

IJGI

View full text Add to dashboard Cite

The digital elevation model (DEM) generates a digital simulation of ground terrain in a certain range with the usage of 3D point cloud data. It is an important source of spatial modeling information. Due to various reasons, however, the generated DEM has data holes. Based on the algorithm of deep learning, this paper aims to train a deep generation model (DGM) to complete the DEM void filling task. A certain amount of DEM data and a randomly generated mask are taken as network inputs, along which the reconstruction loss and generative adversarial network (GAN) loss are used to assist network training, so as to perceive the overall known elevation information, in combination with the contextual attention layer, and generate data with reliability to fill the void areas. The experimental results have managed to show that this method has good feature expression and reconstruction accuracy in DEM void filling, which has been proven to be better than that illustrated by the traditional interpolation method.

show abstract

Filling SRTM Void Data Via Conditional Adversarial Networks

Cited by 13 publications

References 5 publications

A shadow constrained conditional generative adversarial net for SRTM data restoration

A shadow constrained conditional generative adversarial net for SRTM data restoration

Voids Filling of DEM with Multiattention Generative Adversarial Network Model

DEM Void Filling Based on Context Attention Generation Model

Contact Info

Product

Resources

About