ENHANCING RICE LEAF IMAGES: AN OVERVIEW OF IMAGE DENOISING TECHNIQUES

Abstract

The systematic and meticulous handling and processing of digital images through use of advanced computer algorithms is popularly known as the digital image processing. It has received significant attention in both academic and practical fields. Image enhancement serves as a crucial preprocessing stage in each of the image-processing chain. It enhances the quality of the image and emphasizes its features, making all subsequent tasks (segmentation, feature extraction, classification) more reliable and accurate. Image enhancement is also essential for rice leaf analysis, particularly for disease detection, nutrient deficiency evaluation, and growth analysis. Denoising followed by contrast enhancement of the images are the primary steps of image enhancement and image preprocessing. Image filters are generally employed as image denoising techniques. Image filtering operations are designed to transform or enhance the visual characteristics of an image. These include properties such as brightness, contrast, color balance, and sharpness. Thus, they play a very significant and crucial utility function in enhancement of the overall image quality and enabling the extraction of useful information from the visual data. In the current work, to provide an extensive comparative study of some of the well known image-denoising methods combined with a popular histogram equalization technique CLAHE (Contrast Limited Adaptive Histogram Equalization) for efficient denoising of rice leaves image. The experimental part of this work was performed on a rice leaf image dataset to ensure that the data used in the study is relevant and representative. The results of these experiments were then closely examined using a variety of different metrics to ensure that the image enhancement methods are tested thoroughly and comprehensively. This approach not only provides a strong basis for assessing the effectiveness of various methodologies in the digital image processing but also reveals certain insights that might be useful for adaptation in the future towards agricultural research, and other varied domains.