基于深度神经网络的遥感图像彩色化方法

Translated title of the contribution: Remote Sensing Image Colorization Based on Deep Neural Networks with Multi-Scale Residual Receptive Filed

Jianan Feng, Qian Jiang*, Xin Jin, Shin Jye Lee, Shanshan Huang, Shaowen Yao

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

3 Scopus citations

Abstract

To solve the problems of mistaken coloring and color bleeding in the current colorization methods, an end-to-end deep neural network is proposed to achieve remote sensing image colorization. First, the multi-scale residual receptive filed net is introduced to extract the key features of source image. Second, a color information recovery network is con-structed by using U-Net, complex residual structure, attention mechanism, sequeeze-and-excitation and pixel-shuffle blocks to obtain color result. NWPU-RESISC45 dataset is chosen for model training and validation. Compared with other color methods, the PSNR value of the proposed method is increased by 6-10 dB on average and the SSIM value is increased by 0.05-0.11. In addition, the proposed method also achieves satisfactory color results on RSSCN7 and AID datasets.

Translated title of the contributionRemote Sensing Image Colorization Based on Deep Neural Networks with Multi-Scale Residual Receptive Filed
Original languageChinese (Traditional)
Pages (from-to)1658-1667
Number of pages10
JournalJisuanji Fuzhu Sheji Yu Tuxingxue Xuebao/Journal of Computer-Aided Design and Computer Graphics
Volume33
Issue number11
DOIs
StatePublished - 20 Nov 2021

Fingerprint

Dive into the research topics of 'Remote Sensing Image Colorization Based on Deep Neural Networks with Multi-Scale Residual Receptive Filed'. Together they form a unique fingerprint.

Cite this