An Enhanced Image Segmentation Approach for Detection of Diseases in Fruit

The progress in the realm of image segmentation has helped farmers to use nominal inputs for higher production within limited time. Preliminary identification of diseases on fruits is limited to naked eyes since the majority of these symptoms can only be identified by microscopic visuals. Image segmentation plays a vital part in distinguishing their infected parts from the disinfected ones. In this paper, clustering is used as an approach in image segmentation to cautiously discover the affected parts of the fruits by segmenting the affected areas from the non-affected parts. Four clustering techniques—IS-KM, IS-FEKM, IS-MKM, and IS-FECA—were employed for this purpose. The quality of segmentation was evaluated using few performance measures like SC, RMSE, MSE, MAE, NAE, and PSNR. The result obtained using IS-FECA is more reasonable compared to the other methods. Roughly each value of performance parameters confers better results for IS-FECA-based image segmentation method, which means proper separation of diseased parts in fruits from their un-affected ones is attainable.