Remote Sensing Single-Image Resolution Improvement Using A Deep Gradient-Aware Network with Image-Specific Enhancement

Mengjiao Qin (1) , Sébastien Mavromatis (2) , Linshu Hu (1) , Feng Zhang (1) , Renyi Liu (1) , Jean Sequeira (2) , Zhenhong Du (1)

1 ZJU - School of Earth Sciences, Zhejiang University
2 GMOD - MODélisation Géométrique

Super-resolution (SR) is able to improve the spatial resolution of remote sensing images, which is critical for many practical applications such as fine urban monitoring. In this paper, a new single-image SR method, deep gradient-aware network with image-specific enhancement (DGANet-ISE) was proposed to improve the spatial resolution of remote sensing images. First, DGANet was proposed to model the complex relationship between low-and high-resolution images. A new gradient-aware loss was designed in the training phase to preserve more gradient details in super-resolved remote sensing images. Then, the ISE approach was proposed in the testing phase to further improve the SR performance. By using the specific features of each test image, ISE can further boost the generalization capability and adaptability of our method on inexperienced datasets. Finally, three datasets were used to verify the effectiveness of our method. The results indicate that DGANet-ISE outperforms the other 14 methods in the remote sensing image SR, and the cross-database test results demonstrate that our method exhibits satisfactory generalization performance in adapting to new data.

Mots clés: super-resolution; CNN; remote sensing; deep gradient-aware network; image-specific enhancement

Domaines: Vision par ordinateur et reconnaissance de formes [cs.CV]; Géométrie algorithmique [cs.CG]; Synthèse d'image et réalité virtuelle [cs.GR]; Traitement des images [eess.IV]

Mengjiao Qin, Sébastien Mavromatis, Linshu Hu, Feng Zhang, Renyi Liu, et al.. Remote Sensing Single-Image Resolution Improvement Using A Deep Gradient-Aware Network with Image-Specific Enhancement. Remote Sensing, 2020, ⟨10.3390/rs12050758⟩. ⟨hal-02492404⟩