摘要
针对遥感图像质量评价问题,提出了基于Contourlet变换的结构相似性(SSIM)评价的视觉模型。首先,通过25位遥感专业人员对经过处理的200幅高斯模糊图像、200幅椒盐噪声图像、500幅压缩失真图像进行评价,建立主观评分库;然后对经过Contourlet变换后的IKONOS图像进行C-SSIM质量评价;最后将CSSIM评价结果回归到主观评价空间,与均方误差、峰值信噪比、SSIM评价结果相比,本文方法与主观评价数据库较为一致,并优于其他质量评价模型。
To assess the remote sensing image quality, a novel human vision system model is proposed based on the SSIM of Contourlet transform. Firstly, a subjective image quality assessment database is established with 200 Gaussian-noise images, 200 Salt＆ Pepper noise images, and 500 compression distortion images, processed in matlab7 and then evaluated by 25 remote sensing experts. Secondly, IKONOS images after Contourlet transform are assessed by C-SSIM model. Lastly, experiments show that the C-SSIM model performs better than others, in contrast to MSE, PSNR, SSIM image quality assessment results, is the contourlet-SIIM Model result are consistent with the subjective database.
出处
《武汉大学学报（信息科学版）》
EI
CSCD
北大核心
2014年第1期12-16,共5页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金资助项目(41271456)~~
