Research on edge defects image recognition technology based on artificial neural network

摘要

There are many edge defects such as cracks, joints and knots in the rough laminated wood surface, which make it difficult to detect edges of the rough laminated wood in edge cutting process. To fix edge defects mentioned above, an image recognition technology is developed on the artificial neural network(ANN). The Canny operator is used to detect edges of the rough laminated wood, and a classifier is designed and trained by ANN to distinguish defects of cracks, joints and knots. According to the classified results, the hough transform method is applied to reconstruct edges image of the rough laminated wood. Experimental results indicated that edge defect pattern recognition based on ANN performs well to classify the crack, joint and slipknot of the rough laminated wood surface.