This paper aims to present a new algorithm to remove thin clouds and retain information in corrupted images without the use of auxiliary data. By injecting physical properties into the cycle consistent generative adversarial network (GAN), we were able to convert a cloudy multispectral image to a cloudless image. To recover information beneath clouds and shadows we create a synthetic multispectral space to obtain illumination invariant features. Multispectral vectors were transformed from Cartesian coordinates to Polar coordinates to obtain spectral angular distance (SAD) then we employed them as input to train the deep neural network (DNN). Afterward, the outputs of DNN were transformed to Cartesian coordinates to obtain shadow and cloud-free multispectral images. The proposed method, Hybrid GAN-SAD yields trustworthy reconstructed results because of exploiting transparent information from certain multispectral bands to recover uncorrupted images.

No abstract

Remote sensing image scene classification is a contentious research area, particularly in difficult-to-classify regions. The Danube delta is a constantly changing and difficult to categorize region. Machine learning methods have recently been used for scene classification because of their beneficial results. For many remote sensing applications, co-registration of Multi Spectral Images (MS) and Synthetic Aperture Radar (SAR) data is crucial. This paper focuses on both supervised and unsupervised novel machine learning methods, such as t-SNE, k-means, and SVM, applied to co-registered Sentinel-1 and Sentinel-2 data of the Danube delta. The outcome demonstrates that Sentinel-1 vertical-vertical (VV) is a better band for data training since it exhibits more details, and the learned SVM classifier using t-SNE can be applied to other days with a respectable level of accuracy.

No abstract

No abstract