基于模板匹配算法的图像拼接算法  

Image Stitching Algorithm Based on Template Matching Algorithm

作　　者：彭仁夔 熊志勇[2] 申媛 PENG Ren-kui;XIONG Zhi-yong;SHEN Yuan(Jiangxi Jinlu Technology Development Co.,Ltd.,Nanchang 330000,China;School of Mathematics and Computer Science,Nanchang University,Nanchang 330031,China)

机构地区：[1]江西锦路科技开发有限公司,江西南昌330000 [2]南昌大学数学与计算机学院,江西南昌330031

出　　处：《电脑与信息技术》2024年第6期56-60,共5页Computer and Information Technology

基　　金：国家自然科学基金资助项目(项目编号:62102174)。

摘　　要：针对多帧图像进行全景图像拼接中的拼接时间与拼接速度等相关问题,提出了3种对图片进行无效图片过滤的方法,分别为检测图像边缘信息进行过滤、对图像的信息熵进行计算过滤以及图像方差数据过滤。由于图像拼接过程中可能产生上下偏移的问题,为原始图像添加边框,使得图像在上下位置匹配更加精准。最后对拼接完成的图像结果,使用特征点检测算法检测图像中的特征点,并结合直接拟合方法,获取图像倾斜角度,对图像添加y轴偏移量进行倾斜校正。实验结果有效地证明了改进的算法满足图像拼接的实时性要求,在OCR识别结果相似性上有所提升。To address the issues related to stitching time and stitching speed in panoramic image stitching of multiple frames,three methods of ineffective image filtering were proposed,including detecting image edge information for filtering,calculating image entropy for filtering,and filtering image variance data.Due to the possibility of upper and lower offset during the image stitching process,a border was added to the original image to make the image matching more accurate in the upper and lower positions.Finally,for the completed image results,feature point detection algorithm was used to detect feature points in the image,and combined with direct fitting method,the image tilt angle is obtained and the image was corrected for tilt.The experimental results effectively demonstrate that the improved algorithm meets the real-time requirements and improves the similarity of OCR recognition results.

关 键 词：模板匹配 图像拼接 倾斜校正 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]