International Journal on Smart Sensing and Intelligent Systems
Because of not needing training samples and performing the classification just according to the inherent similarity of data in the multidimensional space, unsupervised classification method now gets more and more attention by remote sensing data analyst. Duo to the special growth environment of mangroves, field measurements is difficult to be done to obtain training samples. Therefore unsupervised method provides a good adjunct way for the classification of mangroves remote sensing image. This paper presents an immune network based unsupervised classification method, which is not necessary to define complex objective function.