Grading flue-cured tobacco leaf based on artificial brain intelligent system

Measuring and grading flue-cured tobacco leaf is one of the important quality control means. It is imminent to be solved to investigate the reasonable characteristics and the way of automatically abstracting of flue-cured tobacco leaf quality and to design a type of assistant leaf grading system. The application of artificial intelligence and computer technique in automatic grading of flue-cored tobacco leaf was studied. A brain information processing structure was used for reference to design the intelligent system for grading flue-cured tobacco leaf. The system consists of three sub-systems: thinking model, perceived model and behavioral model, to simulate different kinds of intelligence of human grading experts. The system is in possession of multi ‘thinking’ functions such as learning and memory, judging and fuzzy reasoning, and grading decision making. The system also has the functions of coordinating and controlling, such as tobacco leaf image acquisition and the communication between the computer based artificial brain model and the SCM set in the image acquisition system. The grading judgments made by the system are highly consistent with the judgments by grading experts. The average consistent rate reaches 85%.