基于预测和JPEG2000的红外图像无损压缩方法

Lossless Compression of Infrared Image Based on Prediction and JPEG2000

为了解决空间遥感凝视型探测器获取的数据量巨大的问题,研究了红外图像的无损压缩。经比较,JPEG2000国际标准压缩算法具有较好的压缩性能且有相应的ASIC(Application Specific Integrated Circuit)芯片易于系统实现。本文对连续的红外图像序列的空间和时间相关性进行了分析,以此为基础提出了一种基于预测和JPEG2000算法相结合的无损压缩方案,并以ASIC和FPGA为平台实现了一个复杂度低、性能高的实时无损压缩系统。对不同复杂程度的源图像进行的压缩实验表明,该系统能够实现连续红外图像平均无损压缩比4.516,与JPEG2000标准算法相比,无损压缩比提高了59%以上。

Among the various IR image compression methods, the standard compression algorithm JPEG2000 has better compression effect. A new compression scheme based on prediction and JPEG2000 is presented by analyzing the correlation of continuous IR image sequences. A low-complexity, high-performance, real-time compression system is realized based on FPGA and ASIC. The experimental results show that the average lossless compression ratio of continuous IR image sequences is 4.516, which is improved by more than 59% compared with JPEG2000.