In this paper a novel Sparse Coding for Classification of Synthetic Aperture Radar (SAR) Images is proposed. Features utilized in SC (Sparse Coding) are extracted from the multisize patches around each pixel to precisely describe the complex terrains. In these one or two level thresholds techniques are introduced in the sparse coding for classifier of SAR Images to the restrict the range of reconstruction residual, which classifies the consistent classified points, and the rest of the pixels are considered as the indecisive ones in the original SAR image. Compared with habitual SC of SAR image classification and support vector machines (SVM) in a number of fixed-size patches. Later, the performance measure parameters like accuracy, kappa coefficient and time are calculated and compared for all RISAT-1 images sizes. The above mentioned classifications techniques are development and performance parameters are calculated using MATLAB 2014a software.
CITATION STYLE
Balnarsaiah, B., Rajitha, G., & Penta, B. (2019). Classifications of SAR Images Using Sparse Coding. In Springer Series in Geomechanics and Geoengineering (pp. 761–769). Springer Verlag. https://doi.org/10.1007/978-3-319-77276-9_69
Mendeley helps you to discover research relevant for your work.