红外压缩图像的质量评价研究

Quality evaluation of impressed infrared images

红外图像作为一种特殊图像,为了更好地传输和存储,有时候必须进行压缩。传统的均方误差(MSE)图像质量客观评价方法,有时其评价结果与人们的主观视觉不一致。为此提出了一种建立在人眼视觉模型(HVS)基础之上的边缘均方误差评价方法,它能更有效地反映出图像压缩失真在视觉上的感知。采用快速傅里叶变换、离散余弦变换、哈尔变换、小波变换、哈达马变换等算法,可对编解码后的图像效果进行评价,证实了基于人眼视觉特性的误差统计方法在对压缩图像进行像质评价时是比较令人满意的。

As a kind of special images,infrared images have to be compressed for proper transfer and storage.The traditional method,which is based on MSE to objectively evaluate the quality of impressed images,sometimes cannot conform to effect of human vision perception.Edge mean square error is defined according to human vision system and a better evaluation result conforming is provided to human vision.Different thermal image compression algorithms are described.It is improved that edge mean square error is more reasonable to evaluate the images.