论文部分内容阅读
针对影像分割所需区域内部满足一致性、区域间互不相交的要求,鉴于高光谱影像地物在尺寸、形状、光谱上的异质性,开展高光谱影像分割研究。应用视觉皮层细胞神经元之间存在的局部兴奋全局抑制振荡网络对视觉影像信息进行深入提取和处理,在此基础上,结合小世界神经网络高群集系数、短特征路长的特点,研究了LEGION神经元振荡器所具有的小世界神经网络的同步性能,从而构建了耦合小世界网络的LEGION分割算法。进一步采用直观参数设置,简化高微分方程的计算复杂,减少迭代次数。实验表明:耦合小世界网络的LEGION分割算法,可有效地把高光谱影像中同质地物分割在一起,达到信息提取的目的。
In order to meet the requirements of intra-region consistency and inter-region disjointness in image segmentation, hyperspectral image segmentation is studied in view of the heterogeneity of hyperspectral image features in size, shape and spectrum. Based on the characteristics of high global clustering coefficient and short feature length of small-world neural network, the local excitatory global inhibitory oscillating network exists between visual cortex cell neurons to extract and process the visual image information. Neuron oscillator with the small-world neural network synchronization performance, thus building a coupled small-world network LEGION segmentation algorithm. Further use of visual parameters to simplify the calculation of high differential equations complex and reduce the number of iterations. Experiments show that the LEGION segmentation algorithm coupled to a small-world network can effectively segment the homogenous objects in the hyperspectral image to achieve the purpose of information extraction.