科研之友
微信
新浪微博
Facebook
分享链接
摘要
This paper presents a fuzzy C-means clustering based segmentation method for segmenting common fish diseases in China. To introduce a new effective segmentation method, three steps are involved: In the first step, RGB components are extraction for cluster space, and then hyper tangent function was used for objective function by replacing original Euclidean distance on feature space, meanwhile, instead of pixel numbers, image gray level was obtained for computing. Finally, by ways of clustering, image region of cyprinids common diseases were clustering. The paper compares the results with results of other general segmentation methods include standard fuzzy C-means algorithm. The experimental results showed that proposed method, with strong robustness and quick segmentation ability, could segment color images of cyprinids common diseases precisely by 92%.
