JPEG2000压缩模拟的残余条纹成因分析

发布时间：2018-05-27 11:14

本文选题：图像压缩 + 压缩比　；参考：《航天返回与遥感》2016年01期

【摘要】：基于某些光学遥感卫星图像相对辐射校正效果不理想,存在残余细条纹现象,文章以JPEG2000压缩为例,分析了图像压缩对相对辐射校正的不利影响。首先对相对辐射校正残余条纹的成因进行了傅里叶分析,然后利用JPEG2000压缩器,对样本图像进行了模拟压缩——解压模拟实验和信息损失分析,实验结果证实了傅里叶分析的结论:在图像压缩比过大的情况下,相对辐射校正难以取得满意的效果,残余条纹现象可能产生。最后利用频率域空间特性量化分析残余条纹和图像信息的方法,定量分析了图像压缩对相对辐射校正的不利影响,结果表明:在采用有损压缩的前提下,降低图像压缩比并不能从根本上解决残余条纹问题,而星上相对辐射校正是彻底的解决相对辐射校正残余条纹的更优方案。
[Abstract]:Based on the fact that the relative radiation correction effect of some optical remote sensing satellite images is not ideal and there are residual fine stripes, this paper takes JPEG2000 compression as an example to analyze the adverse effects of image compression on the relative radiation correction. In this paper, the causes of residual fringes for relative radiation correction are analyzed by Fourier transform, and then simulated compression-decompression simulation experiments and information loss analysis are carried out on the sample images by using JPEG2000 compressors. The experimental results confirm the conclusion of Fourier analysis: when the image compression ratio is too large, the relative radiometric correction is difficult to obtain satisfactory results, and the residual fringe phenomenon may occur. Finally, the negative effects of image compression on relative radiation correction are quantitatively analyzed by using the method of quantifying residual fringe and image information in frequency domain. The results show that: under the premise of lossy compression, Reducing image compression ratio can not fundamentally solve the residual fringe problem, but on-board relative radiation correction is a better solution to solve the residual fringe problem.
【作者单位】：中国天绘卫星中心;
【分类号】：TP751

【相似文献】