Super-resolution reconstruction of remote sensing images based on convolutional neural networks
Xuhui Liu, Yan Xu, Yali Che, Jie Tang · 2022 5th International Conference on Data Science and Information Technology (DSIT) · 2022
Aiming at the problem that the current remote sensing image super-resolution reconstruction algorithm extracts insufficient image information and lacks high-frequency information such as edges and textures, this paper proposes a remote sensing image super-resolution reconstruction algorithm based on multi-scale extraction and attention mechanism. The algorithm uses the Inception-ResNet module to fuse information extracted at different levels to obtain richer image features; introduces an attention mechanism to obtain high-frequency information and enriches texture details; uses a residual network to supplement image information and alleviate network gradient problems. Experiments show that the peak signal-to-noise ratio and structural similarity of the proposed algorithm are better than those of SRCNN, VDSR, SRGAN, and other comparison algorithms, and the reconstruction effect has more detailed information and clearer edge texture than the comparison algorithms.