Research On Estimation Method Of Point Spread Function For Ultrasound Image

At present,the ultrasound imaging system is an important medical detection method.The point spread function(PSF)has spatial variability,and its estimation method has attracted wide attention both at home and abroad.In the absence of a suitable model for describing the PSF of the ultrasound imaging system,the homomorphic filtering method is used to estimate the PSF.This method will inevitably encounter speckle noise and phase illegibility problems.In this thesis,the characteristics of PSF of B-type ultrasound imaging system are discussed by using Field ? model system.Multi-input/output inverse theorem(MINT)model is introduced,and a method of estimating PSF with high precision is found.The main work is as follows:Firstly,the thesis introduces the research background and its significance.Using the estimated PSF for recovery to obtain a clear ultrasound image,and thus assist the doctor on the lesion of the organs to make an accurate diagnosis.By improving the quality of ultrasound imaging,ultrasound diagnosis can be included in the diagnosis of specific disease staging.Secondly,the concept of ultrasound imaging degradation and its restoration is introduced.The degradation process of ultrasound imaging is essentially a convolution process.On the contrary,the recovery process is the process of deconvolution.At the same time,the method of segment estimation is adopted for the space transfer characteristics of PSF.The classical method homomorphic filtering method for PSF estimation of ultrasonic imaging and the speckle noise and phase blur problem of the method are analyzed.Then,the MINT model is used to estimate the PSF.The single input multiple output(SIMO)model is utilized to assist MINT.Using the MINT model to estimate the PSF requires tissue refection function(TRF),the SIMO model was chosen to estimate the TRF?The radio frequency echo data(RF-data)is divided into multiple parts to solve the spatial variability of PSF.Finally,the PSF estimation algorithm is verified by Field ? model and real RF-data.Firstly,the PSF produced by Field ? model is applied for the standard to evaluate the accuracy of PSF estimation algorithm based on homomorphic filtering and that of PSF estimation algorithm based on MINT-model.Then,for a real blur neonatal necrotizing enterocolitis(NEC)RF-data,homomorphic filtering method and new method are used to estimate the PSFs respectively.The results obtained by the two estimation methods are used to restore the image by use of the L-R algorithm and the qualities of the restored image were compared.The results show that the estimated PSF based on the MINT-model method is closer to the real PSF than that based on the homomorphic filtering method.