基于分类配准与全景像素坐标系的显微图像快速拼接算法  

Fast Stitching Algorithm of Microscopic Images Based on Classification Registration and Panoramic Pixel Coordinate System

作　　者：刘伟强 吴杰[2,3] 王红强 朱华庆 LIU Weiqiang;WU Jie;WANG Hongqiang;ZHU Huaqing(School of Biomedical Engineering,Anhui Medical University,Hefei Anhui 230032,China;Hefei Institutes of Physical Science,Chinese Academy of Sciences,Hefei Anhui 230031,China;Science Island Branch,Graduate School of USTC,Hefei Anhui 230026,China)

机构地区：[1]安徽医科大学生物医学工程学院,安徽合肥230032 [2]中国科学院合肥物质科学研究院,安徽合肥230031 [3]中国科学技术大学研究生院科学岛分院,安徽合肥230026

出　　处：《中国医疗设备》2024年第5期1-8,47,共9页China Medical Devices

基　　金：国家自然科学基金(61973295);安徽省重点研发项目(201904a07020092)。

摘　　要：目的提出一种基于分类配准与全景像素坐标系的显微图像实时拼接算法,以解决现有图像拼接算法精度低、计算复杂度高等问题。方法依据单视野图像的质量特征,利用评价函数对图像进行分类配准,适应不同的配准模型;构建全景像素坐标系,依据配准结果在此坐标系下进行整体拼接与缝隙融合;基于自主搭建的尿液样本显微扫描系统进行实验,以验证算法的有效性。结果实验结果表明,相比现有的图像拼接方法,本文基于分类配准与全景像素坐标系的显微图像快速拼接算法更有利于拼接缝隙的融合,且更稳定,拼接效果显著优于加速鲁棒特征和快速傅里叶变换单独拼接;在拼接时间方面,基于分类配准与全景像素坐标系的显微图像实时拼接算法的拼接时间大大缩短,拼接速度提升了1倍,且对于矩形扫描区域拼接的大视野图像,无明显拼接缝隙痕迹,可完成高质量的实时拼接。结论基于分类配准与全景像素坐标系的显微图像快速拼接算法能够满足快速高分辨率的大视野图像拼接需求,为显微图像拼接提供了一种更优的方法,便于相关人员进一步分析和应用。Objective Based on classification registration and panoramic pixel coordinate system,to propose a real-time microscopic image mosaic algorithm to solve the problems of low accuracy and high computational complexity of existing image mosaic algorithms.Methods According to the quality characteristics of the single field image,the evaluation function was utilized to classify and align the images to adapt to different alignment models.A panoramic pixel coordinate system was constructed,and overall splicing and seam fusion were carried out under this coordinate system according to the alignment results.Experiments were carried out based on the self-built urine sample micro-scanning system to verify the effectiveness of the algorithm.Results The experimental results showed that compared with the existing image mosaic methods,the rapid microscopic image mosaic algorithm based on classification registration and panoramic pixel coordinate system was more conducive to the fusion of stitching gaps,and was more stable,and the stitching effect was significantly better than that of speeded up robust features and fast Fourier transform alone.In terms of stitching time,the real-time stitching algorithm of microscopic images based on classification registration and panoramic pixel coordinate system greatly reduced the splicing time,and the splicing speed was doubled.Moreover,for large-field images stitched in rectangular scanning area,there was no obvious stitching gap trace,and high-quality real-time stitching can be completed.Conclusion The rapid microscopic image mosaic algorithm based on classification registration and panoramic pixel coordinate system can meet the needs of fast and high-resolution large-field image stitching,and can provide a more favorable method for the microscopic image mosaic method,facilitating the further analysis by the related personnel.

关 键 词：显微图像 特征点检测 图像拼接 快速傅里叶变换 

分 类 号：R197.39[医药卫生—卫生事业管理]