Seal following learning method

The invention belongs to the field of Internet of Things, and particularly relates to an imprint learning method which is particularly suitable for building a logistics freight system. A hardware layer of the seal learning method comprises a server side, a first terminal and a second terminal, and the second terminal learns from the first terminal to obtain independent action ability. In the method provided by the invention, a certain amount of data processing is executed by the terminal by adopting a printing-following learning mode, so that the computing power of the terminal can be fully called, the resource utilization maximization is realized, the pressure of a server side is reduced, and the convergence of behavior modes of each terminal and the machine learning effect are improved. The learning process of the second terminal is not only the establishment process of action logic, but also the multiplication process of transportation material resources, so that the efficiency improvement of logic and rea