针对模糊效应的卫星图像质量评价模型研究  

Research on evaluation models of satellite image blur

作　　者：唐中奇[1] 孙继银[1] 刘浩 

机构地区：[1]第二炮兵工程学院计算机与指挥自动化系,西安710025 [2]第二炮兵通信总站,北京100085

出　　处：《计算机应用研究》2011年第11期4394-4397,共4页Application Research of Computers

基　　金：国防"十一五"科研资助项目

摘　　要：为了实现景象匹配制导应用中对图像数据源质量的客观评价,针对卫星图像的模糊效应,研究了影响图像质量的清晰度评价指标,通过分析比较其对图像匹配过程的影响,挑选出有效指标作为评价因子;通过对大量实验数据特性的统计,结合线性回归分析,提出了基于假设检验的权值拟合算法,完成了各指标贡献率(即权值)的求解;进而建立了面向匹配制导应用的图像质量评价模型。实验表明,该模型具有良好的单调性和稳健性,在匹配制导中能对图像质量进行有效评定,对把握卫星基准图像源质量具有重要现实意义。In order to realize an objective evaluation on image quality in the application of image matching guidance,and aiming to the blur effect of satellite image,this paper researched into the clarity evaluation parameters which affected the quality of image.And picked certain effective parameters through analyzing and comparing the effects they exerted on image matching process.After taking the statistics of massive experiment datum into consideration,and referring to linear recursive analysis,this paper proposed a weights fitting method based on hypothesis test,and accomplished the calculation of the weights of relative parameters.Then set up an image quality evaluation model aiming at matching guidance.The tests show that this model is robust and of probabilistic monotonicity,and is meaningful in improving the evaluation of image quality in matching area,therefore this model is practical in ensuring the satellite image quality.

关 键 词：匹配制导 图像模糊效应 权值拟合算法 图像质量评价模型 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]