Method on Simulating Remote Sensing Image Band by Using Groundobject Spectral Features Study

Cheng, Xi; Shen, Zhanfeng; Luo, Jiancheng; Shen, Jinxiang; Hu, Xiaodong; Zhu, Changming

doi:10.3724/sp.j.1010.2010.00045

Cited by 2 publications

(2 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, it is still difficult to model the complex and random atmosphere and clouds. When the input condition is insufficient for simulating the images of earth background, based on the correlation between the spectral domains, the known spectral images can be used to achieve target spectral image synthesis [9][10][11]. However, the correlation between the spectral domains is implicit and non-linear.…”

Section: Introductionmentioning

confidence: 99%

SDTGAN: Generation Adversarial Network for Spectral Domain Translation of Remote Sensing Images of the Earth Background Based on Shared Latent Domain

Wang

Zhu²,

Guo

et al. 2022

Remote Sensing

View full text Add to dashboard Cite

The synthesis of spectral remote sensing images of the Earth’s background is affected by various factors such as the atmosphere, illumination and terrain, which makes it difficult to simulate random disturbance and real textures. Based on the shared latent domain hypothesis and generation adversarial network, this paper proposes the SDTGAN method to mine the correlation between the spectrum and directly generate target spectral remote sensing images of the Earth’s background according to the source spectral images. The introduction of shared latent domain allows multi-spectral domains connect to each other without the need to build a one-to-one model. Meanwhile, additional feature maps are introduced to fill in the lack of information in the spectrum and improve the geographic accuracy. Through supervised training with a paired dataset, cycle consistency loss, and perceptual loss, the uniqueness of the output result is guaranteed. Finally, the experiments on the Fengyun satellite observation data show that the proposed SDTGAN method performs better than the baseline models in remote sensing image spectrum translation.

show abstract

Section: Introductionmentioning

confidence: 99%

SDTGAN: Generation Adversarial Network for Spectral Domain Translation of Remote Sensing Images of the Earth Background Based on Shared Latent Domain

Wang

Zhu²,

Guo

et al. 2022

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Because too simple describes the relationship of the remote sensing bands, the linear model is difficult to obtain good simulation results [2][3][4][5]. Even nonlinear models also were used by some scholars to fit the relationship between remote sensing bands and then simulate blue band, however, due to the samples used to train the model often are inconsistent with the target image in imaging time, the trained model is also difficult to be used for target image's band fitting [6][7][8][9][10].…”

Section: Introductionmentioning

confidence: 99%

Using MOD09 Data to Produce a Natural-color Image from the Blue-lacked Multispectral Remote Sensing Data

Shen¹,

Zhang²,

Qu³

2015

J Remote Sensing & GIS

View full text Add to dashboard Cite

Because of atmospheric effects, some satellite sensors cover only two visual spectral bands (green and red bands) in addition to bands in the near-infrared to thermal-infrared regions, and lack a blue band. As a result, a natural-color image cannot be obtained, as the blue band is necessary in combining red, green, and blue to produce natural color. This greatly affects the application of remote sensing in many areas such as virtual reality, terrain simulation, and visual interpretation. In this study, the MODIS land surface product (MOD09) was used as reference imagery from which to select pixel samples, and a non-linear regression analysis model-a back-propagation artificial neural network (BPN)was used to fit the spectral reflectance relationship among the blue band and red, green, and near-infrared bands. Landsat TM/MSS, ZY1-02C and SPOT blue bands were then simulated with the trained fitting model, and a natural-color image was output. The experiment result shows that the MOD09 samples trained BPN model well simulated the blue band of a multispectral image and even more informative blue band, more importantly; it can eliminate the influence of the atmospheric for the blue band to some degree. With the simulated blue band, a more realistic and informative natural-color image was acquired.

show abstract

Method on Simulating Remote Sensing Image Band by Using Groundobject Spectral Features Study

Cited by 2 publications

References 3 publications

SDTGAN: Generation Adversarial Network for Spectral Domain Translation of Remote Sensing Images of the Earth Background Based on Shared Latent Domain

SDTGAN: Generation Adversarial Network for Spectral Domain Translation of Remote Sensing Images of the Earth Background Based on Shared Latent Domain

Using MOD09 Data to Produce a Natural-color Image from the Blue-lacked Multispectral Remote Sensing Data

Contact Info

Product

Resources

About