通过对元胞自动机理论的研究,提出元胞自动机的遥感影像混合像元分解模型。利用多波段遥感数据验证混合像元分解算法的可行性,并将结果与线性分解模型进行比较。结果表明,元胞自动机混合像元分解模型在分解的准确性方面,明显优于一般线性模型的精度。最后,将分类结果与传统的监督分类算法比较,得出元胞自动机的混合像元分解模型明显优于监督分类精度的结论。
By means of cellular automata, a new scheme for detection and classification of sub-pixel in remote sensing images is presented. In this study, experiments which the mixed pixels were assigned several land cover classes according to the new proposed algorithms by using multi-channel remote sensing images data is proved feasible. Comparison of the proposed scheme with Linear-model were performed. The results show that the accuracy of proposed scheme is obviously higher than general linear-model s. At last, t...