基于SIFT特征匹配的活性污泥显微图像拼接方法  

Microscopic Image Stitching Method of Activated Sludge Based on SIFT Feature Matching

作　　者：赵立杰 吴志豪 黄明忠 ZHAO Lijie;WU Zhihao;HUANG Mingzhong(Shenyang University of Chemical Technology,Shenyang 110142,China)

机构地区：[1]沈阳化工大学信息工程学院,辽宁沈阳110142

出　　处：《沈阳化工大学学报》2023年第1期74-79,86,共7页Journal of Shenyang University of Chemical Technology

基　　金：国家重点研发计划项目(2018YFB1700200);2020年辽宁省高等学校创新人才支持计划;2021年度高等学校基本科研项目重点项目(LJKZ0442)。

摘　　要：针对相差显微图像视野有限,单张图像难以全面展示样品信息的问题,提出一种基于SIFT特征匹配的活性污泥图像自动拼接方法.该方法通过SIFT特征提取算法提取污泥显微图像的特征,运用最近邻算法进行匹配.使用投影变换模型进行空间几何变换,将待拼接的活性污泥显微图像叠加在一起.采用渐入渐出融合算法对叠加后的图像进行融合,解决图像拼接过程中拼接缝明显的问题.某污水处理厂活性污泥相差显微图像拼接实验结果表明该方法在拼接质量和信息完整性方面表现良好.In response to the problem of limited field of view in phase contrast microscopy images,which makes it difficult for a single image to fully display sample information.An automatic image stitching method based on SIFT feature matching for activated sludge is proposed.The method extracts the features of the activated sludge microscopic images using the SIFT feature extraction algorithm and matches them using the nearest neighbor algorithm.Using a projection transformation model for spatial geometric transformation by stacking the activated sludge microscopic images to be stitched.A fade-in and fade-out blending algorithm is used to merge the stacked images,addressing the visible seams in the image stitching process.The experimental results of stitching phase-contrast microscopic images of activated sludge from a wastewater treatment plant show that our method performs well in terms of stitching quality and information integrity.

关 键 词：活性污泥显微图像 SIFT特征匹配 渐入渐出融合 图像拼接 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]