在小波包变换的基础上提出了一种新的基于嵌入式零树编码的数据压缩方法,并将其应用于电力系统故障数据压缩中.采用小波包变换方法降低小波系数之间的相关性,再利用树形结构表示小波系数并进行逐次累计量化,进一步减小小波系数之间的相关性.在编/解码过程中数据流根据重要性依次排列,在任一点停止编/解码仍然能够重构出原始信号,适合于实时处理.仿真实验表明,改进后的嵌入式零树编码方法对电力系统故障数据压缩十分有效,并且其嵌入式的特性提高了故障录波器的性能.
A new embedded zerotree wavelet based data compression method was proposed and applied to power system malfunction data compression. The method used wavelet packet to reduce the correlation of wavelet transform coefficients, then zerotree coding and successive approximation provided a compact representation of the wavelet coefficients. Using the embedded coding algorithm, an encode can terminate the encoding at any point without much distortion, which is suitable for real time processing. Test results show that the modified embedded zerotree coding method is effective for power system malfunction data compression, and that the embedded property is very helpful for improving fault record performance.