Single stripe three-dimensional point cloud measurement method of end-to-end calibration deep learning network

The invention discloses a single fringe three-dimensional point cloud measurement method of an end-to-end calibration deep learning network, which can realize three-dimensional point cloud measurement of a single fringe pattern by using a deep learning technology only by using a checkerboard fringe pattern and a single object fringe pattern, and in order to solve the problems that the calibration process of a fringe projection measurement system is tedious and a data set is not universal under different calibration parameters when a deep learning technology is used, a checkerboard fringe pattern and a single fringe pattern are used for jointly training a deep learning network, the checkerboard fringe pattern can perform world coordinate calibration on the network, and through a calibrated deep learning network model, an object in the single fringe pattern is mapped to a world coordinate system corresponding to the checkerboard, so that the network directly outputs a three-dimensional point cloud of a real wor