Leaf Morphological Feature Extraction of Digital Image Anthocephalus Cadamba

Abstract

This research implemented an image feature extraction method using morphological techniques. The goal of this proccess is detecting objects that exist in the image. The image is converted into a grayscale image format. Then, grayscale image is processed with tresholding method to get initial segmentation. Furthermore, image from segmentation results are calculated using morphological methods to find the mapping of the original features into the new features. This process is done to get better class separation. Research conducted on two Antocephalus cadamba (Jabon) leaf diseased seedlings data set image that contained leaf spot disease and leaf blight. The results obtained morphological features such as rectangularity, roundness, compactness, solidity, convexity, elongation, and eccentricity able to represent the characteristic shape of the symptoms of the disease. All properties form the symptoms can be quantitatively explained by the features form. So it can be used to represent type of symptoms of two diseases in Antocephalus cadamba (Jabon).