The Study Of High Resolution Plane-wave Emission Ultrasound Imaging Algorithm

Ultrasound imaging is one of the most important medical technologies. It uses high frequency sound waves and their echoes to form an image of tissue to help diagnose the diseases in clinical applications. Its unique characteristics such as cost effectiveness, convenience, and high safety make it become one of the four mainstream medical imaging techniques. Although ultrasound imaging has found wide application medically, some problems of the current ultrasound imaging techniques are worth seriously studying due to the constraint of the conventional ultrasound imaging principle. There are three main problems. Firstly, the resolution is too low to distinguish between the close targets. Secondly, the quality of the image is degraded greatly by the noise. Finally, the temporal resolution is not high enough to diagnose the organ movements in realtime. In practice, these problems have a strong relationship between each other. Thus how to settle the problem effectively is significant to design a high-powered ultrasound imaging system in medical application.The thesis introduces the basis ultrasound imaging principle and the structure of ultrasound system firstly, and analyzes the main factors that influence the frame-rate and the image quality and the corresponding solutions. It helps us to understand the ultrasound system deeply.The thesis proposed a novel plane-wave ultrasound imaging method based on sparse representation at the first time. How to increase the imaging quality by single plane wave insonifying is the key problem to be settled in the high-frame-rate plane wave emission imaging area. This thesis applied the over-complete dictionary sparse representation theory to the linear ultrasound imaging model and derived the ultrasound imaging model based on sparse representation which is able to obtain the high resolution and high contrast ultrasound image directly by the single acquired echo data.The thesis acquired the simulation data for test by the well-known ultrasound simulation software Field-II. The simulation results show that the proposed method is able to enhance the resolution and contrast of the ultrasound image, and performs well against noise.Speckle noise is the inherent problem of ultrasound image. This thesis introduces the formation mechanism of speckle noise and its impact on the image quality. To effectively preserve ultrasound image edges while filtering, an adaptive denoising algorithm based on variations and Euler’s elastic energy model is proposed. And in order to solve the conflict between iteration times and step sizes, an iteration-stopping criterion and an adaptive step-size iteration speckle removal scheme are proposed. To evaluate the proposed algorithm, both synthetically introduced speckle images and real medical ultrasound are used. The simulation results show that the proposed method not only preserves image details effectively while filtering, but also reduces the iteration times.