基于列率内插差值重采样分层编码的图像压缩技术-信号与信息处理专业论文.docxVIP

摘 摘 要 在当今的信息社会中，图像信息占全部信息的比重将越来越大；运用图像压 缩编码技术使图像信息的存储、传输占用更少的空间而所恢复的图像质量又保持 比较高的视觉效果，是当今图像压缩编码技术的关键。列率内插差值重采样分层 编码图像压缩技术结合了传统预测编码和变换编码的特点，既充分减小数据相关 性，又使得运算量不会太大。本论文尝试将列率滤波内插与图像分层的编码方法 结合起来，可实现ATM传输时的丢色优先权，从而实现多码率的、不同质量的 重建图像，满足不同情况下对图像的质量要求。 本方案首先采用基于二维IDCT变换的列率内插滤波器对亚采样图像进行列 率内插，然后将原图像与列率内插后的图像进行相减，就得到差值图像。再根据 内插差值的统计特性，编码误差对重建图像主观质量的影响，对内插差值进行最 佳量化。对原图像亚采样一次，然后列率内插一次，由前面提到的相减生成一层 残差；在亚采样的基础上再进行亚采样，然后再进行列率内插，就又生成一层残 差：如此进行下去就能生成类似于金字塔似的残差数据结构。最后对最佳量化后 的残差数据进行自适应算术编码。实验表明，当所用测试图像的数码率为 0．167bit／pel时，图相信噪比可达30．06dB。本系统在微机上用c语言编程实现。 关键词： 图像压缩编码 列率内插滤波 亚采样 重建图像 数码率 ABSTRACTIn ABSTRACT In the modem society,there iS more and more image information in the all the information．And it is the key of image compression coding to make the need of image memory and transmission bandwidth less and to reconstruct good image quality from human vision by image compression coding technology．A image compression technology based on layered coding with sequency interpolation， difference and subsample is combined witll the trait of traditional prediction and transform coding．It both minishes relativity among datum and does not make the time of computing more．TIlis paper is trying to use sequency interpolation filter on still image compression coding，and combine it with image compression method based on layered coding，such can implement a priority that makes some color datum put aside，that iS，are not transited at ATM，at t11e same time，reach different data rates and images having different quality．So difference quality requirement can be satisfied under the different conditions． In this scheme．firstly we use 2一D sequency interpolation filter based on IDCT to interpolate the subsampled image and get the difference the original image minus the interpolated image．Then the difference is quantized in terms of the statistical characteristics ofthe difference and coding error impacting on reconstructed image．A layer of difference is generated by subsampling original image once and interpolating su