Inspired by enormous success of fully convolutional network (FCN) in semantic segmentation, as well as the similarity between semantic segmentation and pixel-by-pixel polarimetric synthetic aperture radar (PolSAR) image classification, exploring how to effectively combine the unique polarimetric properties with FCN is a promising attempt at PolSAR image classification. Moreover, recent research shows that sparse and low-rank representations can convey valuable information for classification purposes. Therefore, this paper presents an effective PolSAR image classification scheme, which integrates deep spatial patterns learned automatically by FCN with sparse and low-rank subspace features: (1) a shallow subspace learning based on sparse and low-rank graph embedding is firstly introduced to capture the local and global structures of high-dimensional polarimetric data; (2) a pre-trained deep FCN-8s model is transferred to extract the nonlinear deep multi-scale spatial information of PolSAR image; and (3) the shallow sparse and low-rank subspace features are integrated to boost the discrimination of deep spatial features. Then, the integrated hierarchical subspace features are used for subsequent classification combined with a discriminative model. Extensive experiments on three pieces of real PolSAR data indicate that the proposed method can achieve competitive performance, particularly in the case where the available training samples are limited.

Synthetic aperture radar (SAR) images inherently present random and complex spatial patterns, which makes the land-cover classification from SAR images a challenging task. A convolutional neural network (CNN) has been applied to the land-cover classification. However, the statistical properties of an SAR image have not yet been explicitly considered by CNN for feature extraction. To address this problem, this letter presents a statistical CNN (SCNN) for land-cover classification from SAR images, which enables the representation of learning and statistical analysis to be implemented with a unified framework. In the proposed SCNN, the distribution of mid-level primitive features, extracted by representation learning, is characterized by their first- and second-order statistics. These statistics are used to fit the land-cover representations, which encode the statistical properties of the SAR image in the feature space. Experiments on the TerraSAR-X data demonstrate that the SCNN is effective and efficient for the land-cover classification from SAR images.

Due to the development of deep convolutional neural networks (CNNs), great progress has been made in semantic segmentation recently. In this paper, we present an end-to-end Bayesian segmentation network based on generative adversarial networks (GANs) for remote sensing images. First, fully convolutional networks (FCNs) and GANs are utilized to realize the derivation of the prior probability and the likelihood to the posterior probability in Bayesian theory. Second, the cross-entropy loss in the FCN serves as an a priori to guide the training of the GAN, so as to avoid the problem of mode collapse during the training process. Third, the generator of the GAN is used as a teachable spatial filter to construct the spatial relationship between each label. Some experiments were performed on two remote sensing datasets, and the results demonstrate that the training of the proposed method is more stable than other GAN based models. The average accuracy and mean intersection (MIoU) of the two datasets were 0.0465 and 0.0821, and 0.0772 and 0.1708 higher than FCN, respectively.

Contradictions in spatial resolution and temporal coverage emerge from earth observation remote sensing images due to limitations in technology and cost. Therefore, how to combine remote sensing images with low spatial yet high temporal resolution as well as those with high spatial yet low temporal resolution to construct images with both high spatial resolution and high temporal coverage has become an important problem called spatio-temporal fusion problem in both research and practice. A Multi-Dictionary Bayesian Spatio-Temporal Reflectance Fusion Model (MDBFM) has been proposed in this paper. First, multiple dictionaries from regions of different classes are trained. Second, a Bayesian framework is constructed to solve the dictionary selection problem. A pixel-dictionary likehood function and a dictionary-dictionary prior function are constructed under the Bayesian framework. Third, remote sensing images before and after the middle moment are combined to predict images at the middle moment. Diverse shapes and textures information is learned from different landscapes in multi-dictionary learning to help dictionaries capture the distinctions between regions. The Bayesian framework makes full use of the priori information while the input image is classified. The experiments with one simulated dataset and two satellite datasets validate that the MDBFM is highly effective in both subjective and objective evaluation indexes. The results of MDBFM show more precise details and have a higher similarity with real images when dealing with both type changes and phenology changes.