Rice Crop Digital Image Compression Technology Based on Discrete Fourier Transform

基于离散傅立叶变换的水稻作物数字图像压缩技术研究(英文)

[Objective] To study the digital image compression technology in rice monitoring system. [Method] A digital image compression technology program based on the discrete Fourier transform was proposed, and simulation experiments were carried out to compress the image at different compression ratios. [Result] When com- pression ratios were less than 30, the compression ratio, image entropy, average codeword length, coding efficiency and redundancy which reflected the quality of the coding, and the parameter PSNR which estimated the fidelity of the compressed im- age were all achieved good results that human eye could barely percept the differ- ence between the original image and decompressed image; and when the compres- sion ratios were more than 30, there was a certain distortion in the decompressed image. And when the compression ratio was 91.516 3, although the image had some distortion, the PSNR was still achieved to 21.528 2, and human eye could accept the decompressed image intuitively within the acceptable error range. [Conclusion] The results show that the proposed image compression program is a viable, effective, and better image compression technology which can satisfy the requirements of the crop monitoring system on image storage, transforming and transporting.

[目的]研究水稻监控系统中的数字图像压缩技术。[方法]提出了一种基于离散傅立叶变换的数字图像压缩技术方案,并通过仿真试验对图像进行了不同压缩比的压缩。[结果]对于压缩比在30以内的数字图像的压缩,衡量编码压缩质量的5个参数(压缩比、图像的熵、平均码字长度、编码效率、冗余度)和评价压缩图像逼真度的参数(PSNR)均达到较好效果,人眼几乎不能分辨出解压缩图像和原始图像之间的差异;在压缩比超过30的数字图像的压缩上,解压缩图像存在一定的失真。当压缩比为91.5163时,虽然图像有一定的失真,但PSNR仍能达到21.5282,从直观上人眼在一定的误差范围内仍能接受解压缩图像。[结论]该研究结果表明提出的图像压缩方案是一种可行的、有效的、较好的图像压缩技术,完全可以满足农作物监控系统中图像存储、交换与传输的要求。