基于梯度的CCSDS压缩码流控制算法研究

为了优化CCSDS压缩标准的率失真特性,提出了一种基于梯度的CCSDS压缩码流控制算法。该算法采用梯度模值来描述图像各编码段的纹理复杂度,进而用模值之和表示各段所包含的信息量。首先计算出各编码段梯度模值之和,并将其相加得到整幅图像的梯度模值之和,根据前者在后者中所占比例,将码流容量分配给各编码段。采用MATLAB对新算法、CCSDS标准算法及Jin算法进行仿真测试,比较三者图像的恢复质量。同时,测试新算法和Jin算法分配码流所需时间,并进行对比。实验结果表明,在图像恢复质量方面,新算法优于CCSDS标准算法,稍逊于Jin算法,但是新算法的实时性比Jin算法好。因此,基于梯度的码流控制算法可应用于CCSDS标准中。

A Rate Control Algorithm for CCSDS Based on Gradient

To improve the rate-distortion performance of CCSDS, a rate control algorithm based on gradient was proposed. The gradient magnitude was used to describe the texture complexity of image, and the amount of information contained by each segment was related to the magnitude sum of the segment. After calculating all of the segment magnitude sums, it added them to obtain the gradient magnitude sum of the whole image. And then, according to the proportion of the former in the latter, the rate volume was allocated to each segment. After simulating the proposed algorithm, CCSDS and Jin algorithm with MATLAB, it compared the restored image quality of the three algorithm. At the same time, both of the proposed algorithm and Jin algorithm were implemented to compare the time cost by each algorithm for allocating the rate. The experiment shows that, the restored image quality of the proposed algorithm is better than that of CCSDS, while it is a little worse than Jin algorithm. But the real-time of the proposed algorithm is better than Jin algorithm. Therefore, the rate control algorithm can be applied to CCSDS.