A Pre Processing Approach for Accurate Identification of Plant Diseases in leaves

摘要

Detection of diseases in plant leaves is very crucial in agriculture. If the diseases in leaves are not identified at the initial stage it may affect the productivity. Pre processing the images is an important step in any image processing applications. Several algorithms are used to detect the diseases in plant leaves. But, when these algorithms are used for the noisy images or images taken under poor light conditions, they do not tend to generate expected result. Pre processing of these images becomes necessary because of noise, difference in illumination, camera orientation. Pre processing of images assists the further processing steps. In the proposed method the image is initially sharpened, and then Median filter is applied for denoising. The image is then segmented using k-means clustering. Combining these methods, better quality image was obtained.