论文部分内容阅读
针对智能交通流数据量大、无固定模式、精度要求不同的特点,提出了一种根据具体数据格式自适应切换压缩模式的编码算法。对于可进行有损压缩的数据,将时间序列预测思想应用于编码算法中,进行Contourlet逐级分解,直至精度达到要求;在该层应用ARMA建模,传送时只传递其ARMA参数,译码时先以ARMA模型构造Contourlet系数,然后重构原始数据;对于需进行无损压缩的数据,先分析其数据特性,结合传统算法,设计了一种等长、变长编码相结合的压缩方法,在保证数据可完全译出的前提下,达到最佳压缩率。仿真结果表明:对于无损压缩,新算法压缩率可达0.329~0.62;对于有损压缩,压缩率可达0.055~0.29,相对误差为2.34%~4.32%。
Aiming at the characteristics of large traffic volume, no fixed mode and different precision requirements, an encoding algorithm for adaptively switching compressed modes according to specific data format is proposed. For data that can be lossy-compressed, the idea of time-series prediction is applied to the coding algorithm, and the Contourlet is decomposed step by step until the accuracy is met. At this layer, ARMA is used to model and transmit only its ARMA parameters. Firstly, the Contourlet coefficients are constructed by ARMA model, and then the original data is reconstructed. For the data to be losslessly compressed, the data characteristics are analyzed firstly. Combined with the traditional algorithm, a compression method combining equal length and variable length coding is designed. The data can be completely translated under the premise of achieving the best compression ratio. The simulation results show that the compression rate of the new algorithm can reach 0.329 ~ 0.62 for lossless compression and 0.055 ~ 0.29 for lossy compression with the relative error of 2.34% ~ 4.32%.