Research On Real-time Image Processing Algorithm Based On The Medical Ultrasonic Images

As being nondestructive,painless,cheap and convenient,the B-mode ultrasonic diagnosis is extensively used in the medical areas.Yet for its own characteristics,the image quality should still be improved by the image processing technology.This article is based on the real developmental item,the purpose of which is to find a real-time image processing algorithm,to eliminate the speckle noise due to the acoustic transmission and to enhance the quality of the image.This article makes a study of the methods of the image enhancement,as Gaussian filters,sorting and ranking filters,self-adaptive filters,anisotropic diffusion and first order and second order filters,researches the theories,the features the spheres of application and the effects of these methods,and through a lot of simulation experiments analyzes their functions and applicability in the area of the ultrasonic imaging.Then considering the characters of the medical ultrasonic images,the article gives a simplification of the anisotropic diffusion,to make it able to be applied in the engineering application.And then a new method of the ultrasonic image processing is proposed,which is based on the parallel processing of simplified anisotropic diffusion and the isotropic laplacian filter,and could efficiently improve the quality of the medical ultrasonic images.As well as eliminating the effect of the speckle noise,the method can effectually enhance the edge features and keep the details and the texture of the images.As the method has a strong applicability,it can be used for the images of different parts of the body.The parameters easy to adjust,the amount of calculation not so large,this method can be used in the engineering applications.To further improve the functions of the method and to meet the need of different requests,by introducing the ideas of Sobel filter and the serial processing,this article also proposed several corrective methods,to make the method get better operations at the enhancement of the edge features and the eliminations of the speckle noise.