Study On Suppression Of Speckle Noise Based On Adaptive Gaussian Filtering

Ultrasonic imaging technology is widely applied in clinical treatment, but because of imaging mechanism, it’s easy to contain speckle in ultrasound images. The existence of speckle makes the quality of image relatively poor, which is difficult for image to be processed. Because the image denoising method can suppress speckle, retaining edge feature and improve the image contrast.it can also make the clinicians identify and analyze lesion accurately. There fore, the research of ultrasound image denoising methods has the vital significance.There are two basic approaches to speckle reduction-compounding approach and post image formation filtering approach. Among them, the compounding approach includes frequency and space domain compounding technology. Because its simplicity and efficiency, Post image formation filtering approach have recently received increasing interest.In view of the complex method, this paper firstly obtains the conclusion that the SNR value of the image can increase and if there are N pieces of image to be compounded, the SNR value of the obtained image is the sqrt of N times of the original image.In order to prove this conclusion, we use field Ⅱ software system simulation cyst image and process the image with frequency composite way.The experimental results show that the signal to noise ratio of the composite image is increased, and the quality of the image is enhanced. In addition, the relationship between the noise particle size and the center frequency is also discussed.According to the filtering method, this paper presents an adaptive Gauss filtering algorithm for ultrasonic speckle noise reduction. This method firstly uses the local feature matching to detect the spots and the Rayleish distribution of the spots. According to the independent signal model, the similarity between speckle region and feature region is obtained. Finally, using the Gauss filter to filter the speckle, the similarity value is obtained as the standard deviation of the filter, so that the parameters of the filter can be guaranteed with the change of the image area.Comparative experimental results show that the algorithm is superior to the classical SRAD algorithm, DPAD algorithm and Gaussian algorithm. The evaluation results show that the new method of denoising have a very good performance in terms of noise suppression and edge retention, and efficiency of the algorithm is better.