Pulp Fibre Recognition Expert System Based on Neural Network

The characteristics of pulp fibre influence the final qualities of paper production, but the phenomenon of intersectant, curled and bifurcate of the pulp fibre is very familiar, as a result the routine methods can hardly achieve the analysis and recognition of complex pulp fibre image preferably. The paper introduces the basic principle, structure and recognition tactics of the pulp fibre recognition artificial intelligence expert system based on neural network. According to the characters of pulp fibre, it distinguishes from different levels, The system start the fleet repository to analyze the input simple fibre character, then the system will start the deep repository of expert system to get the result of complex fiber character. From the parallel processing and oneself learning of neural network, it overcomes some insufficiency of traditional recognition methods, obtaining preferable recognition effect.