小样本光子图像的统计处理

讨论了一种对小样本光子图像的统计处理方法。在超微弱发光的研究中（例如细胞的超微弱荧光），由于发光强度极弱，需要用像增强器对超微弱发光图像进行增强得到可视图像，超微弱发光图像不可避免地受到像增强系统暗噪声及背景噪声的影响，使光子图像湮没在噪声中。为从原始图像中检验出信号，根据信号光子和噪声光子的不同统计分布，运用信号检测与的方法判断光子是否属于信号光子，并得到一简明的判据，由此判据剔除图像中的噪声光子，得到信噪比改善的光子图像。并用此方法处理了人掌的超微弱发光光子图像。

Thin Samples Photon Image Processing with Statistical Method

A statistical method has been developed to process thin samples photon image. In studying of ultraweak luminescence ( e.g. the cell ultraweak bioluminescence ), it is need to apply a image intensifier to get visible photon counting image. The photon images are unavoidably influenced by intensifier dark noise and background noise, which make the signal photons submerged in noise photons. According to the statistical difference of signal photons and noise photons, a signal test and evaluation method is given to test whether the photons belong to the signal photons and get a concise criterion. By this criterion the noise photons can be distinguished from original image and be discarded to get S/N improved photon image. Using this method a human palm ultraweak luminescence photon image is processed.