Research On Corneal Endothelial Cell Image Analysis Embedded System

The field of medical imaging is undergoing a profound transformation with the rapid development of computer technology and digital image processing technology.The counting instrument of corneal endothelial cells,with which corneal endothelial cell images are processed and analysed on the basis of computer,is one of medical image devices,and it has great significance to assist doctors' work and improve the diagnostic accuracy.Facing with a growing number of ophthalmic patients,clinicians make a diagnosis of corneal pathological changes quickly and accurately using the counting instrument of corneal endothelial cells.Existing products at home and abroad are expensive,bulky and not easy to carry.Therefore,this paper developed an embedded analysis system of corneal endothelial cells image,which could not only overcome the shortcomings above,but also has the advantages of easily handling,fast speed and high precision.This paper researched the methods of image analysis according to the characteristics of the corneal endothelial cell images.In the pre-processing stage,the paper compared and researched image enhancement,segmentation,feature extraction and other classic algorithms.In the analysis phase,the thesis proposed an improved angular and point recognition method based on Freeman chain code,effectively removed a lot of camouflage angles,and improved the recognition efficiency.In the system development process,the system demanding analysis was carried on.Firstly,this paper developed a simple and convenient man-machine interface through building a Linux system in a virtual machine and configuring the graphical user interface development environment of Qt.The camera collecting and controlling functions were achieved through the system of controls.Secondly,the paper achieved image processing and analysis algorithms using C++ programming,combined with the image processing interface of Qt to complete image analysis,stored images and analysis results in a SQLite database to achieve the functions of increasing,deleting,modifying,querying and so on.At last,the system was transplanted into the Tiny4412 platform and was tested.The results shows that the embedded corneal endothelial cell image analysis system has the advantages of high precision,strong real-time,high stability and reliability,as well as a strong practical value and broad market prospects.