星上多通道TDICCD相机内部的图像实时合成压缩

为了解决星上图像整合电路工作频率过高而造成的成像系统工作不稳定的问题，在相机的结构上进行创新，提出了一种基于FPGA平台实现的星上TDICCD遥感图像实时合成压缩方法。首先，介绍了星上多通道遥感图像实时合成压缩的原理；然后，针对TDICCD线阵推扫成像模式的特点，采用基于行的2D-DWT整型提升算法，提出了比特平面和编码过程全并行的EBCOT结构，实现了在对图像数据进行整合的同时进行图像的近无损压缩；最后，搭建试验平台进行验证，结果表明，在无损压缩和2倍压缩时，系统主时钟从原来的100 MHz分别降低为80 MHz和64 MHz，系统稳定，恢复图像的PSNR可以达到84 dB，满足大于80 dB的成像质量要求。处理每行的平均时间61.2 s小于相机的最小行周期63 s。算法稳定可靠，实时性好，实现了相机内部的合成压缩，满足实际工程需求。

Real-time composing and compression of image within satellite multi-channel TDICCD camera

In order to solve the problem of imaging system instability caused by high work frequency of image composing circuit on the satellite, the camera's structure was improved, a composing and compression method of TDICCD based on FPGA platform for satellite remote sensing images was proposed. First, the principle of real-time composing and compression for multi-channel satellite remote sensing image was introduced. Then, based on TDICCD linear array push-broom imaging mode characteristics, line-based 2D-DWT integer lifting scheme was used. EBCOT of bit-plane and coding parallel structure was proposed. Composing and compression of image data was simultaneously achieved. At last, experiment platform for the validation was established. Experimental results indicate that the system clock reduce to 80 MHz and 64 MHz respectively from the original 100 MHz in the lossless and 2 times compression. Reconstruct image's PSNR is 84 dB that is bigger than 80 dB. The average processing time of one line image is 61.2 s that is smaller than camera's smallest line period. The proposed algorithm is stable and reliable. It achieves the composing and compression within camera, it satisfy engineering requirements.