A segmentation method for disease spot images incorporating chrominance in Comprehensive Color Feature and Region Growing

Abstract

In this research work, a segmentation method for disease spotted leaves using Advanced Comprehensive Color Feature (ACCF) and Region Growing method captured under real field condition is proposed. The captured diseased leaves have two main challenges one is clutter background and then uneven illumination, this issue makes the robust segmentation lower. Two methods namely Advanced Comprehensive Color Features (ACCF) and Region Growing method are used in this process for segmentation of disease spots to overcome those challenges. The Advanced Comprehensive Color Feature detection consists of different color spaces, color indexes and color to grayscale conversation using Singular Value Decomposition (SVD) which makes more powerful discrimination of disease spots from uneven illumination. In disease spot segmentation, region growing method is used for eliminating clutter background by interactively selecting growing seeds in ACCF map. The morphological operation is applied in the resultant region growing method. Under real field condition, the proposed method gives an average accuracy of 87% in segmentation.