基于分类矢量量化器的超声检测数据压缩方法

Ultrasonic detecting data compression method based on a classified vector quantizer

研究了基于分类矢量量化技术的超声数据压缩算法,提出了基于峰值数的超声数据分类方法以及基于峰值距离失真测度的码字搜索算法。本文提出的算法充分利用了超声信号的特征,能够很好地保存超声信号的峰值信息。实验表明,相对于普通矢量量化器,本文提出的分类矢量量化器能够在保持相同压缩比的情况下提高重构信号的信噪比。将此矢量量化器与霍夫曼编码相结合,其压缩比可达1:50,高于基于小波变换的压缩算法。

The research of ultrasonic data compression method based on classified vector quantization was carried out. An ultrasonic data classified method based on the peak number and a code-searching method based on the peak distance distortion measure were presented. The proposed algorithm in this paper takes the full advantage of the ultrasonic signal＇ s characteristic to keep the peak information significantly. Compared to the normal vector quantizer, experimental results showed that this classified vector quantizer could effectively improve the SNR （signal-to-noise ratio） of the reconstructed signals under the same compression ratio. When the vector quantizer was combined with Huffman could be up to 1：50, which was higher than the compression ratio of the method based on Coding, its compression ratio wavelet transform.