Control method of early education robot

The invention provides a control method of an early education robot, and relates to the technical field of artificial intelligence. Whether object recognition is executed or not is confirmed firstly;if yes, a target image is acquired, a standard image is obtained after preprocessing, an object recognition model is established according to the convolutional neural network, the standard image is input into the object recognition network to obtain a recognition result, and an action console of the robot is called to command the robot to make corresponding actions, sounds and 3D display by adopting a method of simulating mouse click actions; If not, obtaining a gesture image, establishing a gesture recognition model, inputting the gesture image into the gesture recognition model to obtain a recognition result, and calling an action console of the robot to command the robot to perform corresponding actions by adopting a method of simulating mouse click actions. According to the method, based on the cnn convolution