Research On Location-sensitive ? Imaging Positioning Algorithm Technology

Gamma camera is a device used to image gamma radiation emitting radioisotopes,and can intuitively and visually obtain the spatial distribution,isotopic constituency and radiation intensity of radioactive sources.Due to advantages of nondestructive analysis and non-contact detection,gamma cameras are widely used in astronomy,nuclear medicine imaging,radioactive environment monitoring,military and homeland security and so on.The hardware module and positioning algorithms of gamma-ray imaging system are two main factors of affecting the performance of gamma camera.It is very challenging and costly to reduce or eliminate the reconstructed problems in hardware modules by upgrading modules and optimizing design.Another way to improve the performance of the gamma camera is to establish reliable positioning algorithms by using the parameters of the hardware modules and the principle of readout circuit.Meantime,it is more economical,effective and feasible to enhance performance of gamma camera by using proper positioning algorithms.Positioning algorithms commonly used in gamma cameras include Center of Gravity(COG),look-up table(LUP),maximum likelihood estimation(MLE),least squares estimation(LSE),machine learning(ML)and neural network(NN)algorithms and so on.Due to superior capabilities in positioning,reducing compression and distortion effects and excellent properties of easy development and effortless optimization on the particular experimental setup,LSE is considered as an excellent positioning algorithm that can obviously improve the performance of gamma camera.This paper presents a PSO based LSE(LSE-PSO)positioning algorithm to estimate the position of a scintillation event detected in scintillation crystal by position sensitive photoelectric devices.According to the principle of readout circuit,objective functions for individual and Symmetric Charge Division Circuit(SCDC)are established,respectively.The main contents and results are listed as follows:First,a general overview of application field and research status of gamma-ray imager,especially gamma-cameras,are provided briefly.Meanwhile,several frequently-used positioning algorithms of gamma-ray imaging,especially the Least Square Estimation,are introduced.Second,we describe the basic principles of the most widely used gamma camera and discuss briefly some general aspects of gamma camera.The major components are a collimator,scintillator,a position sensitive detector and a readout circuit.In addition,the key techniques and quality control of gamma-ray imaging are introduced.Third,Particle Swarm Optimization is used to solve the problem of Least Squares Estimation,referred as PSO-LSE.The numerical test results show that PSO-LSE can correctly find out the optimal solutions of objective function within fewer iterations.Meanwhile,it is demonstrated that PSO-LSE has strong searching ability and superior computational efficiency.Two types of objective function for individual and Symmetric Charge Division Circuit(SCDC)are established based on their principles,respectively.Fourth,PSO-LSE is proposed to study the problem of positioning in individual and SCDC readout circuit based on a simulative gamma camera.The constant coefficient introduced to objective functions is demonstrated to effectively improve the position linearity response and expand the useful field of view(UFOV)up to almost the dimension of the crystal for both wide and narrow light distribution.A series of numerical experiments on the effect of noise show that the effect of noise on the reconstruction result is small when the SNR is higher than 20 dB.The addition of an appropriate constant term to the fitting equation may reduce the effect of noise.In order to reduce the influence of scattering of events on the reconstructed image,those abnormal light distribution caused by scattering of events can be identified and discarded by a screening process and will not be calculated by the positioning program.PSO-LSE is helpful to correct the non-uniformity of anode gain through looking up the gain table provided by the manufacturer.Finally,PSO-LSE and other three algorithms are implemented and evaluated on a gamma camera based on SCDC readout.The detector of gamma camera is built of a 22×22 array of CsI(T1)scintillation crystals and an 8×8 matrix of SiPM pixels.All the experimental data of scintillation events in this study are generated by 241 Am radioisotope source.PSO-LSE algorithm shows strong adaptability to the SCDC readout mode.It is noteworthy that PSO-LSE is a potential and promising reconstruction algorithm among the four implemented algorithms that allows achieving excellent pixel identification,peak-to-valley of line profile,UFOV and position linearity response at the same time.In conclusion,it can be reached that the performances of the scintillation camera are affected by the reconstruction algorithms and can be improved by the use of an appropriate reconstruction algorithm.The application of PSO-LSE algorithm offers the possibility to obtain high-quality reconstructed image in gamma camera based on scintillation crystal.