Search

Wheat is the primary source of nutrition for more than two-thirds of the world's population. Wheat can be adversely affected by a number of diseases, resulting in lower yields and, in severe cases, famine. Plants are negatively impacted by illnesses in leaves, a type of sickness. Rapid and precise recognition methods must be used in practice in order to minimize losses. This research study focuses on the segmentation of afflicted regions in order to diagnose the plant disease rapidly and enhance crop quality, therefore solving this problem. As a result, the existing fuzzy c-means clustering (FCM) procedure is highly sensitive to noise, and local spatial information is frequently introduced, resulting in a high computational complexity, which arises from an iterative calculation of the distance between pixels in local spatial neighbors and clustering centers. This study proposes a faster and more reliable FCM technique called marker and mask based membership filtering (M3FCM) to overcome this problem. As a first step, morphological reconstruction is used to incorporate images' local spatial information into M3FCM, ensuring noise protection and image part preservation. Second, local membership filtering relies solely on the spatial neighbors of the membership partition to replace the modification of membership divider based on the distance among pixels inside local spatial neighbors and clustering centers. Kaggle images serve as training and testing for the CNN classification system. The suggested segmentation model achieved 98.06 percent accuracy, while the CNN model scored 94.91 percent accuracy in the experiment.