Research On The Coded Aperture Imaging Algorithm

At present,with the rapid development of nuclear radiation imaging technology,people are paying more and more attention to the position resolution,signal-to-noise ratio and real-time imaging of nuclear radiation imaging.Coded aperture imaging technology follows this trend,replaces traditional parallel hole and pinhole imaging,and has been widely used in nuclear radiation imaging.In this paper,Monte Carlo software was used to simulate the MURA coded aperture imaging.And the reconstruction effect of different decoding algorithms on the projected image was studied,and the decoding algorithm was further optimized to improve the quality of the reconstructed image.This paper focuses on the basic principle of coded aperture imaging,and theoretically deduces whether the coded aperture imaging would have different effects on the quality of reconstructed images when applied to far-field and near-field.According to the arrangement of coded aperture,the MURA coding array was selected as the simulation object,and the MURA coded hole imaging model was established by Geant4 software.Set different positions and difform y-radiation sources which through the MURA coded hole,and then obtained the projection images on the detector.Used ? Decoded,the Fine Sampled Balanced Decoded,MLEM Iteration and modified MLEM Iterative algorithm to reconstruct the image and compare the quality of the image that different reconstruction algorithms reconstruct.By analyzing the results of the reconstructed point source image and the line source image,it could be seen that the signal to noise ratio of the ? Decoded reconstructed image is generally lower than the Finely Sampled Balanced Decoded.This phenomenon is especially noticeable wehen the statustucal noise in the reconstruced Through the Finely Sampled balanced decoded algorithm,them were analyzed that the intensity and position resolution which the point source reconstructed image at different positions change with the distance from the center of the field of view.The "front and back dual-mode measurement method"was used to eliminate artifacts generated by reconstructing images in near-field imaging.In the condition of the MURA coded aperture imaging in far field,the point source imaging at different positions was simulated.The Finely Sampled Balanced decoded algorithm and the MLEM Iterative algorithm were used to reconstruct the projected image.It is found that within one-half of the actual full field of view,the signal-to-noise ratio of the MLEM Iterative algorithm reconstructed image was higher than that of the Finely Sampled Balanced Decoded algorithm by about 50%.But after this range was exceeded,the MLEM Iterative algorithm was used to reconstruct the image.The central artifacts severely affected the quality of the reconstructed image.Therefore,the correction coefficient ? was introduced in the MLEM Iterative algorithm to optimize the algorithm itself.Finally,the central artifact generated in the iterative image was successfully eliminated,and the value of the correction coefficient ? was given.The position correction of the point source reconstructed by the modified MLEM Iterative algorithm was performed.The position correction function in the actual full field of view was given,and the square matrix distribution of the point source was verified.The results showed that the relative deviation of the point source position in the reconstructed image was reduced from 5.4%-27.7%to 0.025%-7%.Finally,the multi-point source MURA coded aperture imaging at different locations was simulated,and three decoding algorithms were used for image reconstruction.The results showed that the modified MLEM Iterative algorithm could reconstruct the multi-point source image better.