Sublingual microcirculation video sequence physiological parameter estimation method and system

The invention discloses a sublingual microcirculation video sequence physiological parameter estimation method and system. The method comprises the following steps: S100, obtaining a sublingual microcirculation image sequence; S200, carrying out iteration by utilizing a correlation filtering algorithm to obtain stabilized sublingual microcirculation video data; S300, acquiring blood vessel video data by using a U-Net network model; S400, obtaining a blood flow rate grade by using the CNN-LSTM neural network model; and S500, calculating sublingual microcirculation physiological parameters according to a blood vessel segmentation result and a flow velocity grading result so as to represent the microcirculation perfusion state of the estimation target. According to the method and system, sublingual microcirculation video physiological parameter estimation is realized based on a deep learning technology, gray scale and time and space information of blood vessels in a microcirculation video can be fully utilized,