中文摘要：

为了提高在低比特率条件下的解码图像质量和视觉效果，根据交通图像背景和局部相似特点与原子参数量化特性，给出了基于原子参数预测和量化的交通图像压缩算法。该算法首先分解一批交通图像，以分解后的原子参数构建原子库，在此基础上，利用构建原子库对稀疏分解的原子参数进行预测和量化，然后对投影分量进行排序差分处理，采用变长编码对处理后的投影分量进行编码，根据投影分量的重排顺序，对经过预测和量化后的原子参数进行相应的重排，最后采用算术编码对重排后的原子参数进行编码。仿真试验结果表明，与已有文献中的方法相比，该算法能够更有效的实现交通图像的压缩，在相同压缩比下，解码图像有更高的峰值信噪比和主观图像质量。

英文摘要：

According to similar characteristics of background and local parts of traffic images, a traffic image compression algorithm based on forecast and quantization of atomic parameters was proposed to improve decoded image quality and vision impact at low bit-rates. The atomic dictionary was constructed by using the atomic parameters obtained from decomposition of traffic images. The forecast and quantization of atomic parameters was made by using the sparse decomposed atomic parameters. Then, projective components were arranged in order, subtracted and variable length coded. After that, atomic parameters were rearranged according to the rearranging order of projective components, and coded by arithmetic coding. Simulation results show that the algorithm can be more effective in the traffic image compression, and the image quality is better and has higher peak signal-to-noise ratio in the same compression ratio comparing with the method in the previous literatures.