Robot rapid robust three-dimensional reconstruction method based on hierarchical Gaussian mixture model

The invention discloses a robot rapid robust three-dimensional reconstruction method based on a hierarchical Gaussian mixture model. The method comprises the following steps: enabling a robot to obtain point cloud data of a measurement object, enabling a GPU to accelerate to generate a layered Gaussian mixture model and a test set, constructing and updating a registration network, globally optimizing the registration network, updating the reconstructed Gaussian mixture model, and repeating the steps until the robot completes measurement at all measurement points; reconstructing a three-dimensional point cloud model of the measurement object, and analyzing and evaluating a reconstruction result. According to the method, a layered Gaussian mixture model is generated by accelerating point cloud data through GPU parallel computing, meanwhile, uncertainty of noise and measurement can be effectively processed, the speed and efficiency of three-dimensional reconstruction are improved, joint registration errors are r