Classification of Textures in Microphotographies of Leaves Epidermis

Microscopic images of leaves, collected from Mona Island dry forest (which is located between Puerto Rico and the Dominican Republic), were analyzed. For each leaf side an image was obtained at two magnifications (200x and 400x). This resulted in four samples of images showing a wide variety of textures and stomata patterns. For each group of images we used the gray-level co-occurrence method to characterize the observed gray level patterns. From the GLCM matrix several texture features were calculated among others: the angular second moment (ASM), the contrast, correlation, inverse difference moment (ISM), and entropy. Visual inspection indicates the formation of three groups of images at 200x magnifications based on the observed patterns. The results of the GLCM analysis indicate consistency between the texture features and the isotropic and anisotropic patterns observed in the leaves