基于微生物图像识别的水质监测方法研究

项目名称： 基于微生物图像识别的水质监测方法研究

项目编号： No.61501522

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 无线电电子学、电信技术

项目作者： 张洁玉

作者单位： 中国药科大学

项目金额： 19万元

中文摘要： 随着经济的发展，我国水体污染问题日益严重。实施水质监测对掌握水资源污染程度及选取合适的治污方法至关重要，其中水体指示性微生物检测是水质监测的重要手段。在微生物检测方法中，利用图像识别技术对微生物进行识别分类和数量统计的方法，具有快捷、高效、适合大样本水体监测等优点，有很大的应用价值和发展前景。本项目立足于水资源监测的重大需求，开发基于图像识别的微生物分析技术，根据水体中微生物图像边缘模糊、微生物重叠及背景复杂等特点针对性地设计图像分割算法及有效特征的提取及匹配方法，结合模式分类手段完成水体中常见指示性微生物的识别分类和数量统计并实现对水质的评估，提出面向自然水体状况和污水处理后水体质量的快速自动监测技术，为水资源保护及治理提供依据。本项目的研究成果可进一步与实时显微图像采集系统结合，实现对水体质量的实时监测与评估，为后续水体治理及制作一定范围内的水质分布图奠定基础。

中文关键词： 水质监测；图像分割；特征提取；图像识别

英文摘要： With the rapid development of economy, the pollution of environment especially water is increasingly serious. It's vital to carry out some effective monitoring over water to get their quality information and then adopt reasonable measures for pollution control. Among them, the detection of some aquatic microorganisms which are highly sensitive to water pollution is served as an important tool for water-quality monitoring. Comparing with the traditional physical or chemical analysis, this novel method for microbial identification and quantitative analysis, which is based on image recognition with convenient process and high detection efficiency, exhibits great application value and prospect for water monitoring. Aiming to develop such an analysis technique of aquatic microorganisms, in this project some algorithms of image segmentation, feature extraction and matching will be specially designed and studied according to the feature of the micrographs of microorganisms, such as dim border, overlapping of microorganisms, and complex background, followed by pattern classification to distinguish the marker microorganisms and perform quantity statistic, based on which the water quality can be evaluated. By working on this research project, we want to propose a fast and automatic method for water-quality monitoring of nature or treated water which can provide strong support for protection of water resources. The research finding is expected to further combine with the real time micrograph taking system, which can be used for real time water monitoring and lays a foundation for water governance and further building water-quality distribution map.

英文关键词： water quality monitoring;image segmentation;feature extraction;image recognition