Image Processing And Analysis Of Single Radial Immunodiffusion Samples In Medical Laboratory

Single radial immunodiffusion plays an important role in basic technical means of medical laboratory,and it has an indispensable position in clinical medicine research and drug supervision industry.For this kind of universal and mature inspection technology,it is usually measured by manual using vernier calipers or scanned images.These methods have large detection errors and are time consuming,and seriously affects the efficiency of drug development and production.This paper uses machine vision and image measurement technology to measure single radial immunodiffusion samples,which solves the problem that the fuzzy edge of images can not be detected at present.The accurate measurement of image is realized to ensure the accuracy of quantitative analysis results of single radial immunodiffusion.The contents of this paper mainly are as follows:(1)Aiming at the problem that it is very difficult to extract shallow target in the image of single radial immunodiffusion sample,an image denoising enhancement algorithm based on wavelet and semi-soft threshold is proposed in this paper.Firstly,the sample gray image is decomposed by wavelet.When three high frequency images are de-noised by semi-soft threshold,the low-frequency image is enhanced by CLAHE algorithm and then the wavelet fusion is carried out.Comparing and analyzing the traditional enhancement denoising method and combined with two-step calibration method to correct the distorted sample image,the algorithm can enhance the detail to improve contrast,and realize preprocessing and deformity correction of the sample images.(2)According to the target position feature of single radial immunodiffusion sample images,this paper proposes a fast segmentation algorithm based on K-means clustering for desegregated points,which combines bisecting K-Means clustering to perform preliminary fast segmentation and LOF algorithm to cut off the residual background.Compared with other segmentation algorithms,such as threshold-based,region-based growth,k-means clustering and so on,this method can obtain complete target image region with less background interference and excellent segmentation effect.(3)Aiming at the continuous diffusion phenomenon of the edge of single radial immunodiffusion samples,we proposes a fuzzy edge segmentation information extraction algorithm,and improve circle fitting algorithm.Through contrast experiments,the edge information extraction algorithm proposed in this paper can extract the target sediment edge for different experimental samples and different targets,and the edge is continuous with few discontinuities and less interference.The measurement algorithm can improve the measurement efficiency and accuracy.The degree is closest to the theoretical value and the reliability is high.(4)Through the joint development of MATLAB and Visual Studio software,weembed the research algorithm into the host computer software,the designed software system can not only complete measurement of single radial immunodiffusion image's datas,but also can draw the specific curve of antigen and antibody,and it has the function of testing concentration of unknown antigen.This subject focuses on measurement scheme of single radial immunodiffusion sample.Through researching the main algorithm,we finally realized obtaining the target sediment measurement results directly from the host compute.This paper contains 54 charts,12 tables and 65 references.