Cloud removal Using Atmosphere Model
This work addresses cloud removal for remote sensing data analysis, offering incremental improvements in speed and accuracy with a new simulation benchmark.
The authors tackled cloud removal in remote sensing by proposing a scattering model within low-rank and sparse frameworks, developing a faster and more accurate variant, and creating a semi-realistic simulation method for objective evaluation. They achieved improved accuracy compared to state-of-the-art methods, including deep learning models, and provided theoretical analysis for regularization parameters.
Cloud removal is an essential task in remote sensing data analysis. As the image sensors are distant from the earth ground, it is likely that part of the area of interests is covered by cloud. Moreover, the atmosphere in between creates a constant haze layer upon the acquired images. To recover the ground image, we propose to use scattering model for temporal sequence of images of any scene in the framework of low rank and sparse models. We further develop its variant, which is much faster and yet more accurate. To measure the performance of different methods {\em objectively}, we develop a semi-realistic simulation method to produce cloud cover so that various methods can be quantitatively analysed, which enables detailed study of many aspects of cloud removal algorithms, including verifying the effectiveness of proposed models in comparison with the state-of-the-arts, including deep learning models, and addressing the long standing problem of the determination of regularisation parameters. The latter is companioned with theoretic analysis on the range of the sparsity regularisation parameter and verified numerically.