基于方向角预测三维小波变换的干涉多光谱图像压缩

根据大孔径静态干涉多光谱图像的成像特点,提出一种基于方向角预测的三维小波变换.这种三维小波变换的新颖之处在于它将方向预测结合到三维提升小波中.这样每次提升小波中的预测可在相关性最强的方向上进行而不需总是局限在变换的方向上.实验证明,对于干涉多光谱图像这种方向性很强的立体图像序列,基于方向角预测的三维小波变换比原始三维提升小波变换有明显改进,在相同的量化编码下,基于方向角预测的三维小波变换比原始三维提升小波变换提高1 dB左右.经该种方法压缩的图像的光谱特性也得到明显改善.     

An adaptive OPD and dislocation prediction used characteristic of interference pattern for interference hyperspectral image compression

According to the imaging principle and characteristic of LASIS (Large Aperture Static Interference Imaging Spectrometer), we discovered that the 3D (three dimensional) image sequences formed by different interference pattern frames, which were formed in the imaging process of LASIS Interference hyperspectral image, had much stronger correlation than the original interference hyperspectral image sequences, either in 2D (two dimensional) spatial domain or in the spectral domain. We put this characteristic into image compression and proposed an adaptive OPD (optical path difference) and dislocation prediction algorithm for interference hyperspectral image compression. Compared the new algorithm proposed in this paper with Dual-Direction Prediction [1] proposed in 2009, lots of experimental results showed that the prediction error entropy of the new algorithm was much smaller. In the prediction step of lifting wavelet transform, this characteristic would also reduce the entropy of coefficients in high frequency significantly, which would be more advantageous for quantification coding [2].     

Optical image compression based on adaptive directional prediction discrete wavelet transform

The traditional lifting wavelet transform cannot effectively reconstruct the nonhorizontal and nonvertical high-frequency information of an image. In this paper, we present a new image compression method based on adaptive directional prediction discrete wavelet transform (ADP-DWT). We first design a directional prediction model to obtain the optimal transform direction of the lifting wavelet. Then, we execute the directional lifting transform along the optimal transform direction. The edge and texture energy can be reduced in the nonhorizontal and nonvertical directions of the high-frequency sub-bands. Finally, the wavelet coefficients are coded with the set partitioning in hierarchical trees (SPIHT) algorithm. The new method holds the advantages of both adaptive directional lifting (ADL) and direction-adaptive discrete wavelet transform (DA-DWT), and the computational complexity is far lower than that in these methods. For the images containing regular and fine textures or edges, the coding preformance of ADP-DWT is better than that of ADL and DA-DWT.     

基于非冗余轮廓波的图像质量可伸缩编码算法

提出一种用非冗余轮廓波的中低比特率图像质量可伸缩编码算法。该算法采用双正交小波分解和方向滤波器组(DFB)实现图像的非冗余稀疏表示, 不但具有轮廓波对图像中线状奇异性边缘和纹理细节的稀疏表示特点, 而且克服了轮廓波变换系数4/3冗余的缺点。算法中对图像非冗余轮廓波系数各子带系数分布进行统计分析, 通过对变换系数的重新组合, 构造了有利于图像编码的空间方向树结构, 并统计验证了其零树特性, 采用分级树集合分裂和阈值量化达到图像质量可伸缩的嵌入式编码。实验结果表明,其解码算法在中低比特率压缩情况下, 压缩后重构图像的感知质量明显优于小波域SPIHT,JPEG2000编码标准, 峰值信噪比PSNR值与JPEG2000相当, 而图像纹理和边缘细节的视觉效果优于JPEG2000和小波域SPIHT算法。     

基于双边滤波和双树复小波的图像去噪算法

提出了基于空域双边滤波和双树复小波变换的图像去噪算法.该算法使用双树复小波变换对含噪图像进行多尺度和多方向的分解,对各个高频方向子带使用带有方向窗的局部维纳滤波算法进行去噪.在重构过程中,对每一个尺度上重构得到的低通图像使用空域的双边滤波算法进一步的去除噪声.实验结果表明本文提出的图像去噪算法获得了明显的去噪性能改善.     

An adaptive VQ method used on interferential multi-spectral image lossless compression

In this paper, a high efficient adaptive LBG VQ (vector quantization) method is proposed, which is an improved version of the expansion LBG VQ method [1] (proposed in 2009)and exploits the shortest distance from a point to line segment. Moreover, the proposed method is used on the lossless compression of LASIS (Large Aperture Static Imaging Spectrometer) interferential multi-spectral images. Compared with the expansion LBG VQ method proposed in 2009, the experimental results showed that the adaptive LBG VQ method improved the compression performance obviously. And the experiment also showed that when using these LBG VQ methods above without Dual-Direction Prediction [2], the entropy generated by expansion LBG VQ method [1] sometimes will be larger than that generated by traditional LBG VQ method [3], and when using these LBG VQ methods above after Dual-Direction Prediction [2], CR (compression ratio) generated by expansion LBG VQ method [1] sometimes will be less than that generated by traditional LBG VQ method [3], while the entropy generated by the method proposed in this paper will be much less than that generated by either traditional LBG VQ method [3] or the expansion LBG VQ method [1], and CR generated by the method proposed in this paper will be much larger than that generated by either traditional LBG VQ method [3] or the expansion LBG VQ method [1].     

大孔径静态干涉成像光谱仪数据源选择对压缩效果影响的研究

针对实际工程应用中存在的大孔径静态干涉成像光谱仪(large aperture static imaging spectrometer,LASIS)的数据压缩问题,根据LASIS的成像特点,在详细分析其数据特性之后,提出了基于光程差维数据源的压缩,并与以往的LASIS和LAMIS数据源压缩进行对比研究。选择不同的数据源获得的压缩效果不同,为了更合理的选择LASIS数据源,降低压缩算法带来的图像和光谱损失,实验以LASIS原理样机推扫获得的短波红外数据作为研究对象,在详细介绍了三种数据的数据特点以及提取方法之后,变换得到上述三种形式的数据源,采用工程中成熟应用的JPEG和JPEG2000算法对三种数据源分别进行逐帧压缩和重建,然后对压缩效果在图像维、干涉维和光谱维以及压缩比进行了详细的对比分析,并提取视场中三种目标物质的光谱曲线,对压缩前后它们的光谱信息损失进行了度量。实验结果表明,相较于采用LASIS和LAMIS数据源作为压缩对象,以LASIS光程差维数据作为压缩对象,无论是在图像维、干涉维还是光谱维的压缩评价标准下都表现出明显的优越性：同样的压缩方案下,基于光程差维数据源的压缩能在获得较高压缩比的同时获得更低的均方误差和更高的峰值信噪比,且对光谱信息的损失也是三种数据源中最小的。     

Magnetic resonance image reconstruction using trained geometric directions in 2D redundant wavelets domain and non-convex optimization

Reducing scanning time is significantly important for MRI. Compressed sensing has shown promising results by undersampling the k-space data to speed up imaging. Sparsity of an image plays an important role in compressed sensing MRI to reduce the image artifacts. Recently, the method of patch-based directional wavelets (PBDW) which trains geometric directions from undersampled data has been proposed. It has better performance in preserving image edges than conventional sparsifying transforms. However, obvious artifacts are presented in the smooth region when the data are highly undersampled. In addition, the original PBDW-based method does not hold obvious improvement for radial and fully 2D random sampling patterns. In this paper, the PBDW-based MRI reconstruction is improved from two aspects: 1) An efficient non-convex minimization algorithm is modified to enhance image quality; 2) PBDW are extended into shift-invariant discrete wavelet domain to enhance the ability of transform on sparsifying piecewise smooth image features. Numerical simulation results on vivo magnetic resonance images demonstrate that the proposed method outperforms the original PBDW in terms of removing artifacts and preserving edges.     

基于运动补偿和码率预分配的干涉多光谱图像压缩算法

提出了一种基于运动补偿的三维小波变换和基于码块预测的码率预分配的图像压缩算法.利用干涉多光谱图像成像推扫平移特性,在小波变换中使用运动补偿来减少帧间相关性,并对图像组中各个图像小波变换和量化后EBCOT编码码块的有效比特平面进行独立的熵估计.以图像估计熵总和指导整个图像组码率预分配,以解决平均分配码率对重建图像质量带来的影响.实验结果表明:该算法在8倍压缩时,图像序列的平均峰值信噪比比3D-SPIHT提高了0.85～1.25 dB,比单帧JPEG2000提高了1.91～4.25 dB, 算法复杂度低,易于硬件实现.     

基于多方向多尺度变换的图像压缩算法

针对小波变换方向选择性差的局限,提出了一种多方向多尺度的的图像变换。圆对称滤波器组首先将图像分解为高频子带和低频子带,然后利用方向滤波器组将高频子带分解为多个方向子带,而对低频子带进行小波变换。多方向多尺度变换能以更稀疏的方式表示图像的边缘和纹理等几何特征,有利于图像压缩。在该变换基础上,结合迭代量化、嵌入式块截断编码(EBCOT)和集合分裂嵌入式块编码(SPECK)构建一种压缩算法。实验结果表明,对于纹理和边缘丰富的图像,压缩算法的性能相对于JPEG2000有明显地提高。     

Acoustic time-frequency domain imaging

The axial resolution of conventional acoustic micro imaging is limited by the wavelength of acoustic waves. Acoustic time-frequency domain imaging was recently proposed to overcome the wavelength limit [Zhang et al., J. Acoust. Soc. Am. 118, 3706-3720 (2005)]. A continuous wavelet transform based acoustic time-frequency domain imaging technique is investigated in this paper. Experiments are performed on real 3D data collected from microelectronic packages. Results demonstrate the proposed technique reveals more image details and enhances the image contrast in comparison with conventional time domain imaging.     

结合小波域变换和空间域变换的图像增强方法

提出了一种结合了小波域和空间域处理方法的图像增强算法. 该算法首先对小波域中的高频系数进行修正,使图像具有更好的局部对比度和更丰富的细节,由于双树复小波变换（Dual-tree Complex Wavelet Transform,DT CWT）具有更好的方向选择性,在小波变换的过程中选用了这一方法;然后,通过空间域中的非线性变换,调整图像的整体对比度. 该算法可根据图像本身的特性实现参数的自动选择. 经过本文方法的处理所得的图像,无论在视觉效果上还是在统计上,都优于前人工作的结果.     

基于峰值信噪比和小波方向特性的图像奇异值去噪技术

提出一种利用小波变换子图像不同的方向特性和峰值信噪比进行奇异值分解的图像去噪算法。由于图像经过小波变换后,低频子图像集中了原图像的大部分能量噪声,故仅作简单维纳滤波;而噪声则主要集中在小波域中的三个不同方向的高频子图中,且系数较小,因此可以利用奇异值分解进行去噪处理,即用较大的奇异值和对应的特征向量重构出去噪图像,然而由于奇异值分解固有的行列方向性,对于高频对角线子图重构出的图像去噪效果不理想,故采取旋转至行列方向后再进行常用的奇异值滤波;最后将去噪后的低频和高频子图进行小波反变换重构出最终的去噪图像,其中重构所需的奇异值个数由图像的峰值信噪比确定。 实验结果表明,该方法在有效去噪的同时较好的保留了原有的高频细节信息。     

一种基于非下采样Contourlet变换多聚焦图像融合算法

针对现有小波类图像融合算法的不足,提出了一种基于非下采样Contourlet变换多聚焦图像融合算法,并在Contourlet域中引入了局部区域可见度以及局部方向能量的概念.针对低频子带系数和各带通方向子带系数分别提出了基于局部区域可见度以及基于局部方向能量的系数选择方案.通过对多聚焦图像融合的仿真实验,表明该算法相对于传统的基于离散小波变换和离散小波框架变换融合算法能够有效减少有用信息的丢失以及虚假信息的引入,同时能够从源图像中提取更多的有用信息并注入到融合图像中, 得到更好视觉效果和更优量化指标的融合图像.     

光谱去相关技术在高光谱图像小波压缩中的应用

随着数据量的不断增长,如何有效压缩高光谱图像成为影响其普及应用的一个关键问题。近年来,小波压缩技术已经被证明是高光谱图像压缩方法中很有发展前景的一个,但由于其对高光谱图像特性的利用较为有限而使其性能的进一步提升受到了限制。文章根据高光谱图像的光谱特征,提出了一种基于光谱去相关的高光谱图像小波压缩方法,设计了分块预测方法来同时去除光谱间相关性和空间相关性,并将其应用于小波压缩方法之中。首先,将高光谱图像分为几个具有高谱间相关性的图像块。然后推导出各块中波段的近似成比例的特性,并在各块分别进行基于这一特性和超光谱图像其他特性设计波段预测编码。最后,将预测用的参考波段和预测后获得的偏差数据,通过小波编码技术进行压缩。实验结果表明,所设计的方法与目前先进的超光谱压缩技术相比其性能有显著的提升。与AT-3DSPIHT算法比较,最高PSNR或SNR提升幅度均能达到4.2 dB左右。此外,此方法在低比特率下的优势也十分突出。     

Image fusion using non-separable wavelet frame

In this paper, an image fusion method is proposed based on the non-separable wavelet frame (NWF) for merging a high-resolution panchromatic image and a low-resolution multispectral image. The low-frequency part of the panchromatic image is directly substituted by multispectral image. As a result, the multispectral information of the multispectral image can be preserved effectively in the fused image. Due to multiscale method for enhancing the high-frequency parts of the panchromatic image, spatial information of the fused image can be improved. Experimental results indicate that the proposed method outperforms the intensity-hue-saturation (IHS) transform, discrete wavelet transform and separable wavelet frame in preserving spectral and spatial information.     

基于旋转、平移和尺度不变的平稳小波图像增强

针对传统的离散正交小波变换对信号的起始位置比较敏感的特点,提出了具有旋转、平移和尺度不变的平稳小波变换,将图像变换到极坐标,采用方向能量函数确定图像主轴方位,并将图像主轴旋转到水平方向得到方向归一化的图像。然后通过对图像的重整和小波基的位移、伸缩、旋转,来消除位移和尺度的影响,得到具有旋转、平移和尺度不变的平稳小波变换。结合非线性增强算子以及Otsu自适应最优阈值估计方法,对图像进行增强。实验结果表明,提出的方法能够在增强图像的同时很好地保留图像的边缘信息,是一种有效的图像增强方法。     

基于图像二维小波多尺度分解傅里叶变换轮廓术

应用傅里叶变换轮廓术测量物体三维面形时,当被测物体形状复杂或是被噪声严重污染时,导致频谱分布展宽,发生频谱混叠现象,基频提取困难,无法准确恢复物体的三维面型.提出了基于小波分解的傅里叶变换轮廓术,采用小波变换的方法对变形条纹图进行二维多尺度分解,重构被测物的背景图像,滤出图像的零频成分,得到相对变形条纹.运用小波变换与傅里叶变换轮廓术相结合的方法,只需拍摄一幅变形条纹图,将被测物体与背景分离,不受背景成分的影响,且易于基频信息的提取,降低了对滤波器的要求.实验证明该方法较好地防止了频谱的混叠问题,提高了测量范围与解相精度.     

基于直方图变换的多光谱图像3D SPIHT压缩编码算法

提出了新的多光谱图像压缩方案,直方图变换的三维分层树的集划分(3D SPIHT)压缩编码算法。基于多光谱图像的成像特点,在去相关之前,提出一种可逆的直方图变换方法对多光谱各波段图像灰度值进行调整,来提高各波段间的相关性,然后再对变换后的图像利用K_L和二维小波变换去除谱间冗余和空间冗余。小波编码采用两种编码方案:3D_SPIHT以及对它进行改进后的三维位平面的SPIHT算法,并对两者进行了比较。实验表明,采用直方图变换的这两种方法都获得了良好的效果,比没变换前有更好的图像质量和压缩性能。     

Three dimension double inversion recovery gray matter imaging using compressed sensing

The double inversion recovery (DIR) imaging technique has various applications such as black blood magnetic resonance imaging and gray/white matter imaging. Recent clinical studies show the promise of DIR for high resolution three dimensional (3D) gray matter imaging. One drawback in this case however is the long data acquisition time needed to obtain the fully sampled 3D spatial frequency domain (k-space) data. In this paper, we propose a method to solve this problem using the compressed sensing (CS) algorithm with contourlet transform. The contourlet transform is an effective sparsifying transform especially for images with smooth contours. Therefore, we applied this algorithm to undersampled DIR images and compared with a CS algorithm using wavelet transform by evaluating the reconstruction performance of each algorithm for undersampled k-space data. The results show that the proposed CS algorithm achieves a more accurate reconstruction in terms of the mean structural similarity index and root mean square error than the CS algorithm using wavelet transform.