Bayesian model-based dynamic flight time measurement method

The invention provides a Bayesian model-based dynamic flight time measurement method. The method comprises the following steps of S1, selecting a pulse ToF camera; S2, building a Bayesian flight timemodel; S3, building a multipath model; S4, building a state spatial model; and S5, building a steady temporal model. A generative model is built by using motions of the camera and an object as observation values; principle-based Bayesian inference is performed on the model; by shooting multiple frames of infrared images, depth estimation is obtained; a positive measurement mode is designed; supplementary information is collected as time goes on; and by a using a regression tree method, first real-time time integration is demonstrated in low operation and low memory operation. Through the Bayesian model-based dynamic flight time measurement method provided by the invention, the camera is improved in calculation efficiency and information capture; the depth precision is improved in a strongbaseline method; the timeliness and the transformation resistance are achieved; and the system precision is improved.