论文部分内容阅读
针对传统模糊聚类的遥感影像分析方法的不足,重点研究基于模糊ISODATA聚类的遥感影像分析。通过Matlab软件编程实现基于迭代自组织数据分析技术、模糊C均值聚类、模糊ISODATA算法对合成图像、纹理图像及真实遥感影像的分类,并对其分类结果进行讨论。通过实验数据对比,评价FISODATA算法的优越性。实验结果表明:ISODATA算法及FISODATA算法都能够实现变类,而FCM算法只能在固定聚类数下进行分类,但是,ISODATA算法分类机制不稳定,不能每次都确定正确聚类数。在迭代过程中,将FISODATA算法引入模糊集理论,便能够快速准确的实现聚类数的确定。
Aiming at the shortcomings of the traditional fuzzy clustering method for remote sensing image analysis, the remote sensing image analysis based on fuzzy ISODATA clustering is mainly studied. By Matlab software programming based on iterative self-organizing data analysis technology, fuzzy C-means clustering, fuzzy ISODATA algorithm for synthetic images, texture images and real remote sensing images classification, and its classification results are discussed. By comparing the experimental data, evaluate the superiority of FISODATA algorithm. The experimental results show that both the ISODATA algorithm and the FISODATA algorithm can achieve variable classes, while the FCM algorithm can only classify the data under a fixed number of clusters. However, the classification mechanism of the ISODATA algorithm is not stable and the correct number of clusters can not be determined each time. In the process of iteration, the FISODATA algorithm is introduced into the fuzzy set theory, which can quickly and accurately determine the number of clusters.