摘要
针对基于图像处理的机械故障诊断和状态监测信息数据量大,不利于传输和存储等问题,提出了一种基于奇异值分解(SVD)和小波变换的图像压缩算法。该算法利用奇异值分解和小波变换对图像进行两级压缩。首先对图像进行奇异值分解,通过设定能量比阈值自适应地保留部分数值较大的奇异值,舍弃其它数值较小的奇异值,实现图像初步压缩,然后对保留的奇异值对应的奇异向量矩阵采用小波变换进一步压缩。将所提算法与霍夫曼编码结合,对火炮炮膛疵病图像进行了压缩。试验结果表明,本文算法在保证图像重建质量情况下,可以有效提高图像压缩比。
Aimed at the problems that the information data of mechanical fault diagnosis and state monitor based on image processing are great,these are not beneficial to information transmission and storage,a kind of image compression method based on singular value decomposition（SVD） and wavelet transform was put forward.The method made use of SVD and wavelet transform to carry out two-level compression of image.The images were decomposed by use of SVD at first,and a part of bigger singular values were reserved adaptively by setting energy ratio threshold while the rest singular values were abnegated.Thereby,first image compression was realized.Then,the singular vectors associated with the reserved singular values were further compressed.This method were combined with Huffman code and applied in bore flaw image compression.Experiment results showed that the proposed method can enhance image compression ratio while the image reconstruction quality is certified.
出处
《火炮发射与控制学报》
北大核心
2012年第4期38-42,共5页
Journal of Gun Launch & Control
关键词
故障诊断
奇异值分解
小波变换
图像压缩
霍夫曼编码
fault diagnosis
singular value decomposition
wavelet transformation
image compression
Huffman code
