基于去噪的SAR复图像SPIHT压缩算法

提出一种对合成孔径雷达(SAR,Synthetic Aperture Radar)复图像压缩的改进的分层树集合分割(SPIHT,Set Partitioning in Hierarchical Trees)算法,高效完成SAR复图像压缩,既有效地保留了SAR复图像的相位信息,又抑制了斑点噪声.首先根据SAR复图像的统计特征进行预处理,然后进行小波变换,对不同子带的小波变换系数采取不同的门限,进行软门限斑点噪声抑制,再经过SPIHT量化编码,得到压缩后的码流.仿真实验表明,在64倍高压缩比的情况下,通过对重建SAR复图像数据的相位误差、等效视数、平均空域相关系数等指标的测量,验证了该算法仍然具有很好的性能. 

Approach based on speckle reduction using SPIHT for data compression of complex SAR image

A method based on speckle reduction using SPIHT(set partitioning in hierarchical trees) for data compression of complex SAR(synthetic aperture radar) image was presented, which could achieve both speckle reduction and image compression. According to the statistic characteristic of complex SAR image, complex SAR image was processed before wavelet transform. Soft-threshold speckle reduction and SPIHT image coding were applied to image compressing in wavelet domain. Wavelet coefficients in different subbands adopted different threshold in soft-threshold speckle reduction. The improved approach preserves the phase of the raw complex SAR image well and has a good effect of speckle reduction. It proves the efficiency of the improved approach for data compression of complex SAR image through evaluating phase error, ENL(equivalent number of looks) and complex spatial correlation efficient, comparing the intensity images of raw data and decompressed data, even at high compression ratio (64∶ 1).