Bernoulli generalized likelihood ratio test for signal detection from photon counting images

Because exoplanets are extremely dim, an electron multiplying charge-coupled device operating in photon counting (PC) mode is necessary to reduce the detector noise level and enable their detection. Typically, PC images added together as a co-added image before processing. We present signal detection estimation technique that works directly with individual images. The method based on generalized likelihood ratio test (GLRT) uses Bernoulli distribution between derived from stochastic model for detector, which accurately represents its characteristics. show our outperforms previously used GLRT relies under Gaussian assumption two algorithms signal-to-noise ratio. Furthermore, provides maximum estimate of exoplanet intensity background while doing It can be applied online, so it possible stop observations once specified threshold reached, providing confidence existence (or absence) planets. As result, observation time efficiently used. In addition time, analysis performance introduced paper also gives quantitative guidance choice imaging parameters, such threshold. Lastly, though work focuses example detecting point source, framework widely applicable.