激光与光电子学进展2024,Vol.61Issue(6) :331-337.DOI:10.3788/LOP232600

快照式显微光谱成像系统及水藻分类识别研究

Research on Snapshot Microspectral Imaging System and Algae Classification Recognition

李栋梁 蔡红星 任玉 李霜 花扬扬 王婷婷 周建伟 曲冠男 王朔 曹洋铭 张桁源
激光与光电子学进展2024,Vol.61Issue(6) :331-337.DOI:10.3788/LOP232600

快照式显微光谱成像系统及水藻分类识别研究

Research on Snapshot Microspectral Imaging System and Algae Classification Recognition

李栋梁 1蔡红星 1任玉 1李霜 1花扬扬 1王婷婷 1周建伟 1曲冠男 1王朔 2曹洋铭 2张桁源1
扫码查看

作者信息

  • 1. 长春理工大学物理学院,吉林长春 130022;长春理工大学吉林省光谱探测科学与技术重点实验室,吉林长春 130022
  • 2. 吉林求是光谱数据科技有限公司,吉林长春 130000
  • 折叠

摘要

目前的显微光谱成像系统的探测模块主要以推扫型光谱成像仪为主,无法进行动态微观样本的观测.基于超材料宽谱调制型光谱成像技术体制,使用该原理研制的快照式光谱相机作为探测模块,其与显微镜模块形成新型的快照式显微光谱成像系统.该系统可实时获取样本的光谱曲线与光谱图像信息.同时利用该系统获取不同藻类的吸收光谱曲线,进一步使用基于支持向量机的图像分割识别算法,对水中的动态藻类样本进行识别.共测试样本80个,预测结果准确率为100%,召回率为65.52%,为快照式光谱成像技术在显微领域的应用奠定基础.

Abstract

Current detection modules in microspectral imaging systems primarily comprise push-broom spectral imagers,which cannot observe dynamic microscopic samples.In this study,leveraging the capabilities of metamaterial broad-spectrum-modulated spectroscopic imaging technology,a snapshot spectroscopic camera was developed and used as a detector module to form a novel snapshot microscopic spectroscopic imaging system with a microscope module.This system enables real-time acquisition of spectral curves and image information from samples.Additionally,we used the developed system to obtain the absorption spectral curves of different algae and further used a image segmentation recognition algorithm based on a support vector machine to recognize dynamic algae samples in water.A total of 80 samples were tested in this experiment,yielding 100%accuracy and 65.52%recall for the prediction results.Thus this result forms the foundation for the application of snapshot spectral imaging technology in the field of microscopy.

关键词

显微镜/光谱成像/快照式/吸收光谱/目标识别

Key words

microscope/spectral imaging/snapshot/absorption spectrum/target recognition

引用本文复制引用

基金项目

吉林省教育厅项目(JJKH20230795KJ)

出版年

2024
激光与光电子学进展
中国科学院上海光学精密机械研究所

激光与光电子学进展

CSTPCDCSCD北大核心
影响因子:1.153
ISSN:1006-4125
参考文献量9
段落导航相关论文