基于小波变换的激光干扰图像质量评价

Quality Assessment of Laser Disturbing Images Based on Wavelet Transform

分析了激光对CCD探测器的干扰损伤机理及损伤图像的特点,总结了常用图像质量评价流程,分析了激光干扰图像质量评价要考虑的因素。通过实验对采集的连续两帧激光干扰图像运用客观质量评价方法进行评价,得到了误差、信噪比及相似度的评价结果。实验结果表明对于两帧连续采集的激光干扰图像,其干扰效果差别不明显,这与主观质量评价结果是一致的,表现为其原始图像、差值图像及直方图没有明显变化,两帧图像误差评价结果均小于0.058%,峰值信噪比参数(PSNR)为51.0793dB,相似度评价参数相关系数为0.9999。针对两幅图像差别细微的特点,采用小波变换的方法提取了差值图像的水平、垂直、近似及对角线等细节分量,结果表明对差值图像进行小波变换所得结果可以较好地体现原始图像的细节信息。

Both the vulnerability and laser-induced damaged image character of CCD detector are studied. The commonly used image quality assessment process is summed up. The error, signal-to-noise ratio and the similarity of two frame of series laser disturbing images are assessed through objective assessment methods. The results show that the disturbing effect of two frame of series laser disturbing images is not obvious and it is consistent with the subjective assessment, which are demonstrated in four aspects, the change of original images and the difference image along with their histograms are not obvious, the gray value difference of both images is less than 0. 058%, the peak signal-to-noise ratio is 51. 0793 dB and the correlativity is 0. 9999. According to the nuance of two images, the vertical, horizontal, approximate and diagonal details of difference image is picked up by wavelet transform and results can reflect the original image details better.