Research And Realization Of Digital PCR Droplet Recognition Method

Droplet digital PCR is the third generation technology based on the traditional PCR technology.This technology is mainly through the polymerase chain reaction(PCR),through the recognition and detection of droplet images to accurately quantify the accounting molecules in DNA,also known as absolute quantitative.But in the process of readout of droplet recognition results,under the condition of the previous artificial counted tend to statistical difficulties because of the extremely large number of nucleic acid,or under the condition of the reaction of accounting rarely prone to identify errors.In view of the above situation,this paper conducts correlation processing and algorithm optimization for digital PCR droplet image recognition,and realizes accurate droplet recognition through a series of image processing from droplet image acquisition to completion.In this paper,the accuracy of droplet image recognition is improved from three aspects.Firstly,the droplet image is preprocessed from the aspects of image grayscale,image smoothing denoising and image enhancement based on fractional order differentiation.The second aspect is the detection of threshold droplet image based on particle swarm optimization.Particle swarm optimization is combined with otsu method to perform droplet image and morphology optimization of binary image.The experimental results show that our algorithm can achieve very good results in edge enhancement in the detection and analysis of droplets.In the droplet image recognition,the number and size of droplet in the image can be accurately identified.The recognition rate of fluorescence and non-fluorescence reached 94.7%,It greatly improves the detection efficiency of PCR and shortens the recognition time and error rate,which is of great significance to the recognition and detection of digital PCR.