Hybrid Approach to Medicinal Plant Leaf Image Enhancement Using Contrast Limited Adaptive Histogram Equalization and Homomorphic Filter

Abstract

Image pre-processing methods are intended to enhance some image features important for further processing including edge detection or object recognition. Preprocessing is very crucial in image processing applications for noise removal and also for suppressing or highlighting other attributes. This study seeks to preprocess medicinal plant leaf images by enhancing the quality of the images using a combined Homomorphic filter and Contrast Limited Adaptive Histogram Equalization. The study also segments the image from its background using the OTSU global thresholding algorithm. Image enhancement and segmentation performance metrics are further used to determine the efficiency of the proposed method. It was realized from the experiment using MATLAB that the proposed method outperformed the Contrast Limited Adaptive Histogram Equalization in terms of the Peak Signal to Noise ratio, Mean Square Error, Mean Absolute Error for the image enhancement and Jaccard Index, DICE Index and Matthews Correlation Coefficient metric for the image segmentation. � 2021 IEEE.