Object surface three-dimensional reconstruction method based on fixed sensor

The invention relates to an object surface three-dimensional reconstruction method based on a fixed sensor, which is used for recovering a three-dimensional model of an object. Firstly, a fixed Kinectv2 depth camera is used for sampling an object on a rotating table in two directions, point cloud data of the target object at different angles are uniformly collected in each direction, and the point cloud data are divided into a front group and an upper group according to the collection directions. Secondly, all point clouds are preprocessed through RANSAC plane fitting, background and noise points are eliminated, and the quality of the point clouds is improved; then, FGR and ICP algorithms are used for conducting point cloud registration on the two sets of data respectively, and registration point clouds in the front direction and the upper direction are obtained; then, outliers are removed through DBSCAN, and the point cloud quality is further improved; reducing the number of point clouds by using a mixed dow