New method of remote sensing image classification based on Ant Colony

For enhancing the accuracy and the versatility of automatic image classification, on the basis of the traditional Ant Colony Algorithm, we try to analysis the images from a new angle, and introduce the model of Ant Colony Algorithm, and take the exploratory research. According to the biological characteristic of the ant colony´s community intelligence´s, this method achieves the classification of remote sensing images. The method is applied to TM image of Hubei Province. Comparing with unsupervised classification, it achieves better results.