Model-based Simulation And Reconstruction Of The Frequency Domian Photoacoustic Imaging

Photoacoustic tomography(PAT)based on photoacoustic effect is currently the most promising technique of noninvasive imaging.It can combine the high contrast in optical imaging and the good spatial resolution in ultrasound imaging,so it has been widely used in the field of biomedical imaging,materials science and industrial inspection.A continuous wave laser is applied which makes frequency domain photoacoustic imaging more economic and portable,and it has the advantages of compact structure and low cost.However,the back projection algorithm requires the full-view of photoacoustic data to obtain an more accurate reconstructed image in the frequency domain photoacoustic imaging,which will take a long time in the actual signal acquisition,and sometimes it is not achievable for special tissues such as breasts.In this thesis,we focus on the study of a frequency domain model based reconstruction algorithms,the purpose of which is to improve the quality of the reconstruction results and to solve the limited angle scanning problems.The main work and contributions are as follows:Firstly,we derived the discretized version of the forward model of the frequency-domain photoacoustic imaging modality in the framework of the photoacoustic wave equation,based on which the inverse reconstruction scheme was presented based on the regularization strategy.In order to improve the accuracy of the model-based reconstruction algorithm,the main parameters are discussed through simulation experiments without the influence of outside interference factors,including regularization parameter selection method of the L-curve and the generalized cross validation(GCV),the Tikhonov and TSVD regularization methods,the number of discrete frequency components,and the number of projection positions.Simulation results show that the L-curve method and Tikhonov regularized inverse method are more suitable for frequency-domain photoacoustic imaging problems,and the qulity of the reconstructed image basically satisfies the linear relationship of the number of projection positions and the discrete frequency components.Secondly,system noise and measurement errors in real experimental are studied.We developed a frequency-domain photoacoustic imaging experiment platform that uses linear frequency modulation to periodically modulate the optical signal output from a continuous-wave laser to excite the photoacoustic signal and verify the performance of a model-based reconstruction algorithm.The experimental results show that the model-based reconstruction algorithm can not only reduce the artifacts in the reconstructed image,but also can only use a small amount of photoacoustic amplitude and phase information to reconstruct high contrast images compared with the back projection reconstruction algorithm.Although reducing the experimental data acquisition amount acquisition time,it can achieve limited view scanning,high-quality photoacoustic imaging.