Parameter estimation of Gaussian gray diffusion model of static image spot

In order to estimate the parameters in Gaussian gray diffusion model, experiments with collimator are carried out while the parameter calibration algorithm is proposed. Equations which take the Gaussian diffusion radius and the deviation values of the centroid coordinate as variables are established and solved, furthermore, the energy-gray coefficient is obtained. Then, series of photos are taken under the same condition in three different orientations, and parameters whose mean values are taken as their estimates are calculated respectively using these gray data of photos. Finally, all the estimated parameters are put into the simulating model, and a model-generated star image spot is therefore got, the similarity between which and real shot ones is calculated in three different windows under the condition that noise is bound to exist. The similarity is a value of more than 0.97 within the window of 3 pixel×3 pixel, more than 0.98 for a 5 pixel×5 pixel, and more than 0.98 for a 7 pixel×7 pixel. The experimental results show that the solution formulas about σ, A and the deviation Δx and Δy are right.