A Research On High-efficient Bayer Patterned Image Compression Algorithm In Deep Space Exploration Applications

Bayer patterned image has been widely used in deep space exploration due to its low storage space requirement and its ability to reflect color information.With the development of remote sensing technology,the amount of image data that needs to be transmitted on the satellite is increasing,and the downlink channel bandwidth of the satellite is limited.This requires the compression coding of the Bayer patterned image on the satellite.How to improve the compression performance of Bayer image by optimizing algorithm under the constrained hardware resource on the star has important research value and application prospect.This paper focuses on the research of Bayer patterned image compression in deep space exploration.Taking into account the reduction of hardware resources overhead and the quality of image compression recovery,a high-performance compression method for Bayer images based on the JPEG2000 standard is proposed for deep space exploration and compression processing with few hardware resources.The main research work is as follows:(1)By analyzing the data correlation between R-G0-G1-B components in Bayer images,an encoding method to remove Bayer image component correlation is proposed.This method uses 1D 4-point integer DCT to reduce the redundant information among 4 components.Then the transformed DCT coefficients are used to perform the wavelet transform,T1 encoding and T2 encoding independently in order using the standard JPEG2000 algorithm.On the basis of ensuring that the hardware is easily implemented,the quality of the restored image is improved.(2)Aiming at the problem of codestream interception in T2 coding for Bayer image component decorrelation coding method,a joint rate-distortion optimization interception method is proposed.After the four components are completely coded,the joint rate-distortion optimization interception is performed using a uniform threshold slope,and the code stream interception results of all the coded code blocks of each component are sequentially obtained,and an independent code stream organization output is performed.Experimental results show that,for Bayer patterned image,compared with the JPEG2000 algorithm to compress four color components respectively,the Bayer patterned image compression method based on the JPEG2000 standard proposed in this paper can effectively improve the quality of the decompression recovery image on the basis of precise control of the code rate,the PSNR increase by 1.4dB~1.8dB,the CPSNR increase by 1.9 dB~2.4dB.Satisfy the deep space exploration in the hardware resources,decompression to restore the high quality of the image and the precise rate control requirements.