基于Lanczos核的实时图像插值算法

doi:10.11959/j.issn.1000-436x.2017115

通信学报 ›› 2017, Vol. 38 ›› Issue (6): 142-147.doi: 10.11959/j.issn.1000-436x.2017115

基于Lanczos核的实时图像插值算法

郭莹,李伦,王鹏

武汉数字工程研究所，湖北武汉 430074

修回日期:2017-03-31 出版日期:2017-06-25 发布日期:2017-06-30
作者简介:郭莹（1986-），女，山东枣庄人，武汉数字工程研究所工程师，主要研究方向为数字图像处理、目标识别技术。|李伦（1986-），男，湖北武汉人，武汉数字工程研究所高级工程师，主要研究方向为图像处理、辅助决策。|王鹏（1980-），男，湖北枝江人，武汉数字工程研究所高级工程师，主要研究方向为图像处理、数据融合技术。
基金资助:
船舶基金资助项目(11J3.7.2)

Real time interpolation algorithm based on Lanczos kernel

Ying GUO,Lun LI,Peng WANG

Wuhan Digital Engineering Institute,Wuhan 430074,China

Revised:2017-03-31 Online:2017-06-25 Published:2017-06-30
Supported by:
The Shipping Foundation of China(11J3.7.2)

摘要/Abstract

摘要：

针对现有彩色图像插值算法的实时性及可靠性不能兼备的问题，提出了一种高效的实时彩色图像缩放算法，算法基于Lanczos核生成可应用于整个目标图像的核查找表，并且目标图像所有像素的插值仅涉及定点小数运算，计算复杂度低；同时算法采取了优化的光栅扫描方法，内存需求低、效率高。实验结果表明，该算法不仅获得了较好的图像质量，而且计算复杂度远低于经典的图像缩放算法，能有效实现彩色图像甚至彩色视频的任意倍数实时缩放。

关键词: 实时彩色图像缩放, Lanczos核, 查找表, 定点小数运算

Abstract:

In order to improve the timeliness and reliability of color image scaling algorithms,a high effective real-time color image scaling algorithm was proposed,which first generated the kernel look up table that could be used for whole target image based on Lanczos kernel,and interpolation of all target image pixels had only fixed point operation which had low computational complexity.Meanwhile,an optimized method for rasterizing pixels of target image was proposed,which needed little memory space but embraced high efficiency of memory utilization.The experiment results show that the proposed algorithm achieves good image quality,and the computational complexity is far more less than classic image scaling algorithms.So the proposed algorithm can effectively implement real-time resizing of color images and color videos in any scaling factors.

Key words: real-time color image resizing, Lanczos kernel, look up table, fixed point operation

中图分类号:

TP391

郭莹,李伦,王鹏. 基于Lanczos核的实时图像插值算法[J]. 通信学报, 2017, 38(6): 142-147.

Ying GUO,Lun LI,Peng WANG. Real time interpolation algorithm based on Lanczos kernel[J]. Journal on Communications, 2017, 38(6): 142-147.

图/表 7

图1

图2

图3

表1

图4

表2

表3

参考文献 12

[1]	JAINA . Fundamentals of digital image processing[M]. Prentice-Hall,Inc Upper Saddle River,NJ,USA, 1989.
[2]	KEYS R . Cubic convolution interpolation for digital image processing[J]. IEEE Trans Acoust,Speech,Signal Processing, 1981,29(6): 1153-1160.
[3]	XIAO J P , ZOU X C , LIU Z L . Adaptive interpolation algorithm for real-time image resizing[C]// Proceedings of the 1st International Conference on Innovative Computing,Information and Control. 2006.
[4]	PARKKINEN J , NENONEN P . A fast method for scaling color images[C]// 17th European Signal Processing Conference. 2009: 2032-2036.
[5]	刘晓松, 杨新, 汪进 . 基于统计特征的彩色图像快速插值方法[J]. 电子学报, 2004,32(1): 29-33.
	LIU X S , YANG X , WANG J . Statistical feature based fast color image interpolation[J]. Acta Electronica Sinica, 2004,32( 1): 29-33.
[6]	郎文辉, 江朝晖, 周荷琴 ,等. 区域指导的彩色图像插值[J]. 电子学报, 2006,34(1): 35-39.
	LANG W H , JIANG Z H , ZHOU H Q ,et al. Area-directed interpolation of color image[J]. Acta Electronica Sinica, 2006,34(1): 35-39.
[7]	FU S J , RUAN Q Q , WANG W Q . Region based shock-diffusion equation for adaptive image enhancement[C]// Advances in Machine Vision,Image Processing,and Pattern Analysis. Berlin:Springer, 2006: 387-395.
[8]	付树军, 阮秋琦, 穆成坡 ,等. 基于双向耦合扩散的保持特片的边缘锐化和图像增强[J]. 计算机学报, 2008,31(3): 529-536.
	FU S J , RUAN Q Q , MU C P ,et al. Feature preserving coupled bidirectional flow for edge sharpening and image enhancement[J]. Chinese Journal of Computers, 2008,31(3): 529-536.
[9]	肖进胜, 饶天宇, 贾茜 ,等. 改进的自适应冲击滤波图像超分辨率插值算法[J]. 计算机学报, 2015,38(6): 1131-1139.
	XIAO J S , RAO T Y , JIA Q ,et al. Interpolation algorithm based on improved adaptive shock filter in image super-resolution[J]. Chinese Journal of Computers, 2015,38(6): 1131-1139.
[10]	席志红, 海涛 . 基于非线性复扩散耦合冲激滤波器的图像放大算法研究[J]. 通信学报, 2014,35(2): 1-8.
	XI Z H , HAI T . Image enlargement research based on anisotropic complex diffusion coupling to shock filter[J]. Journal on Communications, 2014,35(2): 1-8.
[11]	TURKOWSKI K , GABRIEL S . Filters for common resampling tasks[M]. Academic Press, 1990: 147-165.
[12]	CLAUDEE . Lanczos filtering in one and two dimensions[J]. Journal of Applied Meteorology, 1979,18(8): 1016-1022.

算法	PSNR_R/dB	PSNR_G/dB	PSNR_B/dB	平均值/dB
双线性插值	42.361 42	43.292 04	42.365 66	42.673 04
双三次插值	47.974 21	47.979 19	47.968 03	47.973 81
文献[4]算法	43.664 90	43.536 90	43.832 30	43.678 03
本文算法	47.950 00	47.923 70	47.961 90	47.945 20

性能评估	本文算法	双线性插值	双三次插值	文献[4]算法
计算时间/ms	172	515	671	261
加法次数	8	16	22	9
乘法次数	9	18	29	12

性能评估	本文算法	文献[4]的缩小算法	文献[4]的放大算法
加法次数	8	8	9
乘法次数	9	15	12
插值权重总数	306	Ws×2+Hs×2	Wd×2+Hd×2
像素索引总数	0	Ws+Hs	Wd+Hd
亚像素索引总数	0	Ws+Hs	0

基于Lanczos核的实时图像插值算法

Real time interpolation algorithm based on Lanczos kernel

在线阅读

PDF下载

可视化

被引次数

摘要/Abstract

引用本文

使用本文

图/表 7

参考文献 12

相关文章 1

Metrics

推荐阅读 0