The success of many Super-resolution fluorescence microscopy methods lie in the exploitation of the inherent stochasticity of a light emitting molecule’s photon emission state, allowing sparse subsets of molecules to be spatially detected with high precision. This photo-switching behavior, however, induces multiple localizations from each molecule during an imaging experiment, which therefore gives rise to misleading representations of their true spatial locations. By formulating a state-space model relating true molecular positions with observation sets collected across time, we show that the full Bayes filter for this problem can be derived and positions recovered via a Markov Chain Monte Carlo sampler.