On the basis of the information entropy and equilibrium degree of urban land-use spatial structure, the paper analyzes the characteristics and rules of urban land-use spatial structure changes in Wuhan in 1990s, in wh...On the basis of the information entropy and equilibrium degree of urban land-use spatial structure, the paper analyzes the characteristics and rules of urban land-use spatial structure changes in Wuhan in 1990s, in which the types of land-use are shrinking and urban land-use changes are disequilibria. With PCA and GRA employed, the driving forces have also been analyzed. The driving force of city welfare and social structure, the towing force of city industrial structure transition, and the pressing force of city construction and reconstruction are main momentum factors. Moreover, the latter forces are more significant.展开更多
针对高光谱图像中光谱信息提取时高维特征向量由于部分邻域叠加造成数据缺损,以及图像局部区域像素点在空间结构信息中存在同谱异类现象和密度差异的问题,提出了一种基于空谱超像素融合核极限学习机(SSKELM)的高光谱图像分类算法.对光...针对高光谱图像中光谱信息提取时高维特征向量由于部分邻域叠加造成数据缺损,以及图像局部区域像素点在空间结构信息中存在同谱异类现象和密度差异的问题,提出了一种基于空谱超像素融合核极限学习机(SSKELM)的高光谱图像分类算法.对光谱空间第一主成分分量进行超像素分割,每个超像素被看作一个形状自适应区域。利用空间信息、超像素内及像元间的核权重融合,获取像素点类别标签;同时,借助核函数在高维超平面数据中线性可分能力、极限学习机随机隐藏层输出矩阵及其优化算法的限制条件少等优势,将空谱像素点融合训练并形成新的矩阵样本输出.使用University of Pavia和Indian Pines两个数据集进行实验,总体准确率OA值较其他算法分别提高了1.76%和2.80%,有效验证本文提出方法在图像分类中具有一定价值.展开更多
基金Supported by Knowledge Innovation Program of Chinese Academy of Sciences(No.KZCX2-SW-415).
文摘On the basis of the information entropy and equilibrium degree of urban land-use spatial structure, the paper analyzes the characteristics and rules of urban land-use spatial structure changes in Wuhan in 1990s, in which the types of land-use are shrinking and urban land-use changes are disequilibria. With PCA and GRA employed, the driving forces have also been analyzed. The driving force of city welfare and social structure, the towing force of city industrial structure transition, and the pressing force of city construction and reconstruction are main momentum factors. Moreover, the latter forces are more significant.
文摘针对高光谱图像中光谱信息提取时高维特征向量由于部分邻域叠加造成数据缺损,以及图像局部区域像素点在空间结构信息中存在同谱异类现象和密度差异的问题,提出了一种基于空谱超像素融合核极限学习机(SSKELM)的高光谱图像分类算法.对光谱空间第一主成分分量进行超像素分割,每个超像素被看作一个形状自适应区域。利用空间信息、超像素内及像元间的核权重融合,获取像素点类别标签;同时,借助核函数在高维超平面数据中线性可分能力、极限学习机随机隐藏层输出矩阵及其优化算法的限制条件少等优势,将空谱像素点融合训练并形成新的矩阵样本输出.使用University of Pavia和Indian Pines两个数据集进行实验,总体准确率OA值较其他算法分别提高了1.76%和2.80%,有效验证本文提出方法在图像分类中具有一定价值.
