睑板腺图像处理与分析的研究-测试计量技术及仪器专业论文.docxVIP

AbstractMcibomian Abstract Mcibomian Gland Dysfunction(MGD)is a common chronic ocular surface disease．In recent years，the effective observation and analysis of mcibomian gland mo印hology is the objective basis on smdying and evaluating meibomian gland， which is also a hot issue in the clinical research of ophthalmology．Imaging area irregularities of tarsal plate surface，uneven illumination and other factors lead to uneven distribution of gray in mcibomian gland infrared image，low contrast of gland and non-gland area，which makes it difficult to extract the gland area efficiently． Related research has been smdied in the paper on the issues of uneven illumination and low contrast of mcibomian gland infrared image，which provides an effective and objective basis for the clinical diagnosis of MGD．The main contents include： 1，A local standard deviation Mask dodging improved algorithm for image preprocessing is proposed．Based on the Mask dodging principle，dodging image is obtained by the subtraction between original image and the image by Gaussian blur． The ratio of global standard deviation and local standard deviation in dodging image is used as a scaling factor to enhance the contrast of the original image adaptively． And the pre-processed image quality is evaluated．The results show that local standard deviation Mask dodging algorithm Can effectively eliminate uneven illumination of meibomian gland image and greatly enhance the image contrast．Thus，this algorithm Can overcome the shortcomings of traditional dodging algorithms． 2，The effect to the contrast of meibomian gland image by light source wavelength is smdied and enhance contrast of meibomian gland image using dual-wavelength image fusion technology is also studied．Based on the principles of tissue optics，Three-level equivalent optical model of eyelid tissue is established including conjunctiva，meibomian gland and muscle．The meibomian gland imaging device of 850nm and 940nm dual-wavdength is designed and the quality