Statistical property analyses of signal and noise in photon imaging system

A novel method of choosing threshold range, which bases on the criterion that the mean value and the variance are equal in Poisson distribution, is proposed to eliminate the background noise in photon images. Point process analysis method is used to study the statistical properties of point target under the photon limited condition and verify the results experimentally. The results show that the detection statistics of point target under the photon limited condition subject to a stationary Poisson point process. Through the point process analysis method, the expectation occurrence-rate of ultra-weak target and the recurrence probability in the follow period can be obtained quickly, which lay the foundations for detection, tracking and location forecasting.