基于FPGA的九点最临近插值算法的设计与实现

打开文本图片集

中图分类号：TP391.4；TN402 文献标识码：A 文章编号：2096-4706（2025）12-0041-05

Design and Implementation of Nine-point Nearest Interpolation Algorithm Based on FPGA

GUAN Haoyang，LIU Dequan，MENG Xiangyi （SchoolofPhysicsand Electronic Information Engineering，Ningxia NormalUniversity， Guyuan756o99，China）

Abstract： InFPGAs，implementing the bilinear interpolation algorithm through pipelinesconsumes significantresources. This paper proposesanine-point nearestneighbor interpolationalgorithm.Byusing linearcalculation，thegrayscalevaluesof the nine pointsnearthemappedpositionof thepointtobemeasuredareobtained.Thepixelvalueofthepixel tobeinterpolatedisset equal tothe pixel valueofthe nearest pixel among the nine points tothe mapped position，thus completing the nine-poit nearest neighborinterpolationalgorith.ExperimentsshowthatinFPGAhardwareresources，thein-pointnearestneighbortepolation algorithm saves 21.2587% of the total block memory and 20.1413% of the total registers. In terms of image interpolation accuracy， the smoothness ratios for three different types of images are 109.632 3% 104.9660% and 106.5688% ，respectively，while the information entropyratiosare 99.9959% 0 99.946 1% ，and 99.933 2% ，respectively.Thenine-point nearest neighbor interpolation algorithm basicallymeets the requirements ofthe bilinear interpolation algorithm while reducing resource consumption.

Keywords： Field-Programmable Gate Array （FPGA）; nine-point nearest neighbor interpolation; average gradient amplitude; information entropy

0 引言

插值技术广泛应用于军事、航空、医学、消费电子、手持设备和多媒体设备等领域的图像处理中[1-2]，且现多采用FPGA实现视频图像的实时处理[3-]。（剩余5321字）