| According to the progress of the disease,varicose veins are divided into three stages: early,middle and late stage.In the late stage,the leg veins protrude from the surface of the skin,accompanying with the symptoms such as bleeding and ulcers.And the procedure is irreversible.On the other hand,in the early stages,the symptoms can be effectively alleviated by simple methods.Therefore,it is important for venous diagnosis in early stage.However,the symptom is hard to be observed directly in early stage.Hence,a near-infrared imaging system was employed to obtain more vein details by means of infrared image.It will facilitate to diagnose leg varices.In this thesis,the image of the leg vein was obtained by using the near-infrared imaging system.The obtained leg vein image was filtered,enhanced and segmented for analysis.In order to extract the feature of the varicose leg-veins,an adaptive median filter was employed to eliminate noise and recognize the image details.Due to the low contrast of the filtered image,an improved contrast limited adaptive histogram equalization(CLAHE)enhanced algorithm was developed.We used Laplace operator and an adaptive filter to sharpen the leg vein image after CLAHE enhancement algorithm.The sharpened image was filtered again to further enhance the leg vein image and increase the contrast of vein with the rest areas.The threshold binarization was unable to segment the leg vein image effectively.This thesis proposed an adaptive threshold binarization to segment the enhanced leg vein image.And we studied how to determine the dynamic threshold for the characteristics of the pixel gray value in the neighborhood.The performance of the proposed algorithm was adequate for the leg vein image segmentation.In conclusion,a low contrast leg vein image could be better segmented through the proposed image processing algorithm.The segmented images provided a basis for the varicose veins diagnosis in the early stage. |
PDF Full Text Request