Unsupervised Segmentation Method for Diseases of Soybean Color Image Based on Fuzzy Clustering
| Autor: | Jiangsheng Gui, Li Hao, Shusen Sen, Wenshu Li, Yanfei Liu |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2013 |
| Předmět: | |
| Zdroj: | Sensors & Transducers, Vol 159, Iss 11, Pp 32-38 (2013) |
| ISSN: | 1726-5479 2306-8515 |
| Popis: | The method of color image segmentation based on Fuzzy C-Means (FCM) clustering is simple, intuitive and is to be implemented. However, the clustering performance is affected by the center point of initialization and high computation and other issues. In this research, we propose a new color image unsupervised segmentation method based on fuzzy clustering. This method combines advantages of the fuzzy C-means algorithm and unsupervised clustering algorithm. Firstly, by gradually changing clusters c, and according to validity measurement, it can unsupervised search for optimal clusters c; then in order to achieve higher accuracy of clustering effect, the distance measurement scale was improved. In our experiments, this method was applied to color image segmentation for three kinds of soybean diseases. The results show that this method can more accurately segment the lesion area from the color image, and the segmentation processing of soybean disease is ideal, robustness, and have a high accuracy. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|