Method, apparatus and device for recognizing three-dimensional gesture based on mark points

A method for recognizing a three-dimensional gesture based on mark points, applied to a server in a gesture recognition system which further comprises at least two cameras arranged in a designated space and at least one pair of gloves distributed with mark points corresponding to hand joints, the method comprising:

• acquiring an image collected by each of the cameras at a current time;
• determining, for each image, an identifier and a corresponding two-dimensional coordinate position of each mark point distributed on the glove or gloves in the image;
• determining, according to the two-dimensional coordinate position of each mark point and a calibration parameter of a corresponding camera, a three-dimensional coordinate position of each mark point in a coordinate system of the corresponding camera;
• converting the three-dimensional coordinate position of each mark point in the coordinate system of the corresponding camera to an initial three-dimensional coordinate position in a coordinate system of the designated space;
• and
• determining, according to the initial three-dimensional coordinate position of each mark point in each image and the identifier of the corresponding mark point, a target three-dimensional coordinate position of each mark point distributed on the glove or gloves in the coordinate system of the designated space.