Research And Implementation Of Microfluidic Droplet Detection Method

The novel coronavirus pneumonia(SARS-CoV-2)erupted at the end of December 2019,and swept the world in 2020.It seriously endangered human public health safety and posed great challenges to the medical and health system of mankind.As sars-cov-2 is highly infectious and highly pathogenic,rapid and accurate diagnosis and control is very important.Because of its high sensitivity and specificity,nucleic acid detection has been widely used as an important detection method in this epidemic.As a nucleic acid detection method widely used in scientific research and clinical practice,polymerase chain reaction(PCR)technology has been iterated to the third generation.Among them,the first generation PCR is the traditional end-point PCR technology,which can realize the qualitative detection of nucleic acid.The second generation of PCR is real-time quantitative PCR(qPCR),which can realize the relative quantitative detection of nucleic acids.The third generation of PCR technology is digital PCR(dpcr).As a typical application case of droplet technology,digital PCR can amplify the nucleic acid of a single nucleic acid template in a micro droplet environment,so as to achieve high sensitivity and specificity of absolute quantitative nucleic acid detection.As a micro reactor of biochemical reaction,micro droplets have the advantages of high throughput,low reagent consumption,no cross contamination and homogeneous reaction.In the micro droplet technology,the method of directly counting the droplets is used for quantitative analysis of the results.Early droplet counting is often done manually,and its accuracy and efficiency are significantly limited.In order to achieve accurate and efficient droplet counting in various application scenarios,the research on new droplet detection and counting methods has important practical significance.In order to realize the automatic detection of droplets and improve the accuracy and efficiency of detection,based on computer vision,this paper studies a low complexity,low cost and high sensitivity droplet detection and recognition method,which can realize efficient and accurate droplet detection and recognition through image mosaic and recognition algorithm for bright field droplets and fluorescent droplets.Based on image analysis algorithm,an analysis system for droplet detection and counting is studied and implemented.Firstly,the image acquisition device for detecting droplets on microfluidic chip is studied and implemented.Combined with fluorescence excitation module,the whole image acquisition of droplets in microfluidic chip is realized;Secondly,the image mosaic algorithm is studied,and a variety of mosaic algorithms are compared and analyzed,especially the application of SIFT(scale invariant feature transform)algorithm in droplet image mosaic;Thirdly,the droplet image recognition algorithm is studied.According to the characteristics of droplet detection and recognition,the accuracy of droplet recognition is improved,and the recognition rate of fluorescent droplet and bright field droplet is more than 99.2%.In this paper,a simple,reliable and accurate image recognition algorithm is studied for droplet recognition and detection,which can achieve the same recognition accuracy as watershed algorithm,and its complexity is lower than watershed algorithm;Finally,aiming at the application of droplet detection,the software of droplet detection and analysis is researched and realized,and the automatic detection of droplets is realized by cooperating with the droplet detection device.The experimental results show that the microfluidic droplet detection system can realize the accurate and efficient detection,recognition and counting of bright field droplets and fluorescent droplets in various environments.