Efficient Compression Algorithm for Improving Visual Quality ofWeak Targets in High Precision Images

In order to effectively compress high precision images, and ensure desirable fidelity of weak target area of the reconstructed image, a high precision image compression algorithm based on error optimization coding is proposed to improve the quality of weak target area. First, the JPEG-LS (joint photographic experts group lossless) compression algorithm is used to compress the image data, and the error coding data is selected in an adaptive way during the run-length encoding process. Then, the non-uniform quantization based on visual quality is also carried out, and the quantized values are decomposed to remove the correlation between them. Finally, the entropy coding of the decomposed data is carried out by MQ arithmetic encoder. In the reconstruction, the inverse quantization value is reconstructed according to the quantized interval, and the inverse quantization optimization and filtering optimization are carried out. The performance of this algorithm is also compared with those of JPEG-LS and JPEG2000 (joint photographic experts group 2000) algorithms. The experimental results show that this algorithm can achieve high-efficiency and high-precision compression of image data. This algorithm is low in complexity and easy for hardware implementation. Although the algorithm incorporates error data optimization and coding process, the amount of data encoded is small, and it is equivalent to the JPEG-LS in compression speed. The compression speed of the JPEG2000 algorithm is about 4.47 times higher. At the same time, it effectively reduces the information loss caused by the conventional algorithms. The peak signal-to-noise ratio of the reconstructed image of the proposed algorithm is equivalent to or slightly lower than that of JPEG-LS and JPEG2000, but it has better visual quality and fidelity of the weak target area.