Study On Related Technology Of Super Depth Of Field In Leucorrhea Microscopy

Leucorrhea testing is a routine test to determine the health of women,abnormal symptoms of leucorrhea indicates that the patient may be infected with gynecological diseases.With the progress of science and technology,intelligence and automation have become the hot topics that people are discussing in the field of science and technology.How to use these technologies to bring more convenience to people's lives is the theme for many years to come.In the rapid development of intelligent robots,smart home,automatic driving and other technologies,automated medical detection technology is also more and more mature.The emergence of fully automatic blood analyzers,stool detectors,etc.,greatly improves the efficiency of medical testing.However,there are still many problems in the process of automating medical equipment.The microscope sacrifices the microscope's depth of field for greater magnification.During the microscopic examination of the cells,the smear layer of the sample solution on the slide has a certain thickness,so that the cells can not be clearly displayed in the same picture at the same focal length.As a result,the algorithm needs to process multiple images when identifying the cells,which increases the time and complexity of the detection.This thesis is based on the project background of automatic leucorrhea detector,on the one hand through the optimization of mining equipment,on the other hand through the image registration technology and image fusion technology,the images under different focal lengths are processed to get a clear image,so as to achieve an effect of the ultra depth.This thesis is mainly produced from the following directions:In the image acquisition,this thesis uses a metal needle to absorb leucorrhea solution,drip a small amount on the slide,the uniform smear after warm bath to the body temperature,and allowed to stand for some time,through the grayscale camera to capture some pictures of the same field of view at different focal lengths,select three pictures from these pictures as the material pictures of the follow-up process.In the image registration,we choose the correct size,the right position and the right contrast area as the area of the offset correction,and then calculate the difference by the image after the image is panned,and select the minimum translation distance as the offset correction distance.In the selection of image fusion algorithm,according to the characteristics of leucorrhea microscopic pictures,this thesis analyzes and compares the methods of weighted fusion,principal component analysis,Laplacian pyramid image fusion,contrast pyramid image fusion method,gradient pyramid fusion method and so on,and proposed an ultra-depth algorithm for leucorrhea microscopic image.In this thesis,by improving the mining algorithm,the images under different focal lengths and the same field of vision are corrected and fused,which provides the accurate material images for the subsequent recognition algorithms.The recognition rate of the formed components in the leucorrhea microenvironment increased from 91.4% to 94.2%.Experimental results show that the improved algorithm reduces the false detection rate of recognition and improves the accuracy of the recognition algorithm.