摘要
为了对比不同融合算法在提取试验地地类信息的效果,采用Brovey、HSV、PC、Gram-Schmit、和Wavelet融合算法对安塞县和永寿地区的2幅ETM+遥感影像进行融合,并采用信息熵、标准差、偏差指数和相关系数进行了融合效果评价。结果表明:就空间信息量而言,经过HSV变换的图像具有最大的空间信息,而经过Brovey变换的图像,其空间信息最低,其他融合算法则在空间信息量上差别不大;就原光谱信息的保真能力而言,经过HSV变化的图像,其光谱保真能力最差,其他算法之间的光谱保真能力相差不大。
Brovey,HSV,PC,Gram-Schmit and Wavelet fusion methods were adopted to compare the effects of different fusion methods on feature extraction in the experimental areas.Standard variation,information entropy,bias index and correlation coefficient were selected to evaluate the fusion results.In terms of the spatial information,the images with HSV transformation had the most spatial information,whereas the images with Brovey transformation had the lowest spatial information.For the other fusion methods,there were not significantly different.In terms of spectral fidelity,the images with HSV transformation had the lowest value and no significant differences were observed among the other fusion methods.
出处
《西北林学院学报》
CSCD
北大核心
2010年第5期152-156,共5页
Journal of Northwest Forestry University
基金
中国科学院知识创新工程"西北半干旱区水土保持林状态及环境因子地面观测"(KZCX1-YW-08-02-05)
关键词
融合算法
信息熵
标准差
偏差指数
相关系数
fusion methods
information entropy
standard variation
bias index
correlation coefficient