Segmentation of crop disease leaf images using fuzzy C---means clustering algorithm

For improvement on segmentation precision of crop disease images, an adaptive segmentation method of crop disease images was proposed based on fuzzy C-mean clustering algorithm (FCM), according to the properties of crop disease images. The segmentation algorithm used the pixel gray and mean of neighborhood pixels as input features, and modified the membership function of FCM which contains local neighborhood information of image. The optimal cluster number and the degree of fuzziness of FCM were chosen through cluster validity and experiments respectively. The optimal cluster number is 4, and the degree of fuzziness is 2. The adapted segmentation method was used to segment cotton disease leaf images. The result shows that the method of segmentation is satisfactory to separate disease part from normal part of leaves. The mean segmentation errors ≤5%. It is very effective in segmenting and processing crop disease images.