An Analysis of Texture Measures in PCA-Based Unsupervised Classification of SAR Images

摘要: In single-band single-polarized SAR images, intensity and texture are the information source available for unsupervised land cover classification. Every textural feature measure identifies patterns by different approaches. For efficient classification, measures have to be chosen suitably. Therefore, in this letter, role of various is analyzed their discriminative ability image classification into types like water, urban, vegetation areas. To make algorithm adaptable, these features fused using principal component analysis (PCA), components used purposes. highlight effectiveness PCA, difference between PCA- non-PCA-based classifications also analyzed. Analysis real-world data with application PCA presented letter. The how every individual contributes process presented, then, a set according improving accuracy. By analysis, it observed that comprising mean, variance, wavelet components, semivariogram, lacunarity, weighted rank fill ratio provides good accuracy up 90.4% than measures, increased justifies complexity involved process.