Dear colleagues and friends, X. TUFUAB Technical Symposium was held in Aksaray on 25-27, April 2019. The Symposium was carried out by the organizations of Aksaray University and General Directorate of Mapping. As a international symposium in the field of photogrammetry and remote sensing, X.TUFUAB Technical Symposium 2019 is devoted to promote the advancement of knowledge, research, development, education and training in Geographical Information Sciences, Information Technology, Environmental Management and Resources, Sustainable Agriculture, Surveying, Photogrammetry and Remote Sensing, their integration and applications, as to contribute to the well-being of humanity and the sustainability of the environment. 425 participants and scientists from 7 countries were attended to this symposium. 125 oral presentations and 10 poster presentations were presented during the symposium. 135 presentations take place in 25 sessions in two days.
In recent years, deep learning methods have come to the forefront in many areas that require remote sensing, from medicine to agriculture, from defense industry to space research; and these methods have achieved tremendous success as compared to traditional methods. Together with substantial growth in available data with high-quality labels and computational resources, these deep neural network architectures and techniques have seen remarkable developments. The major difference between deep learning and classical recognition methods is that deep learning methods consider an end-to-end learning scheme which gives rise to learning features from raw data. Better regularization techniques and robust optimization algorithms introduced with state-of-the-art deep learning models are other factors leading this difference. In this paper, we discuss the remote sensing problems and how deep learning can be used to solve these problems with a special focus on medical and remote sensing applications. In particular, we briefly review outperforming architectures within the deep learning literature and their use cases.
Sparse unmixing (SU) aims to express the observed image signatures as a linear combination of pure spectra known a priori and has become a very popular technique with promising results in analyzing hyperspectral images (HSI) over the past ten years. In SU, utilizing the spatial-contextual information allows for more realistic abundance estimation. To make full use of the spatial-spectral information, in this letter, we propose a pointwise mutual information (PMI) based graph Laplacian regularization for SU. Specifically, we construct the affinity matrices via PMI by modeling the association between neighboring image features through a statistical framework, and then we use them in the graph Laplacian regularizer. We also adopt a double reweighted $\ell_{1}$ norm minimization scheme to promote the sparsity of fractional abundances. Experimental results on simulated and real data sets prove the effectiveness of the proposed method and its superiority over competing algorithms in the literature.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.