Wood Surface Defect Identification Method Based On Features Of Color And Texture

Wood surface defect detection technology is a multidisciplinary technology based on machine vision and computer image recognition.Through the extraction and analysis of wood image eigenvalues,defects can be recognized and classified.It is widely used in the selection of raw materials for wood products,and has very high practical value.The detection principle of wood defects is to use the image capture of optical equipment and the logical operation ability of computer to replace the human eyes and the brain to perceive wood and identify defects in a specific way.Firstly,the image of wood surface defect is obtained by optical equipment and stored in the form of digital image.Secondly,the obtained wood images are pre processed and divided into blocks,and then the features of the image sub blocks are extracted.Finally,a model is established by training the features extracted from the pattern recognition method.Through the model to achieve the recognition and classification of wood defect image requirements.This paper carries on the study of pattern recognition methods on three kinds of common defects,such as dead knot,wormhole and live knot.Through the observation of wood surface defects,it is found that the color and texture features are one of the biggest differences between wood defects and normal parts.Then,the wood surface image is divided into blocks to extract the color and texture features of the sub block,and the training model is established by using support vector machine.Then,a method of wood surface defect recognition based on color and texture features is proposed.The maximum entropy of gray level,color coherence vector,color moment and LBP texture feature are used to identify the defects.The following is the main work of this paper:(1)The color or texture is used as the feature,and the model is established,and the wood surface defect is identified correctly.However,the single feature of wood surface defect detection is prone to false detection,so that the identification results are not consistent with the real situation.(2)Because of the single use of color or texture as a feature,the effect of identifying wood surface defects is poor.In this paper,the author attempts to the two features together as the feature recognition,identification methods of wood surface defects in order to single feature can further be defect recognition.The experimental results show that the feature fusion is complementary,and the false detection rate of the recognition can be reduced under the premise of correctly identifying the wood surface defects.The accuracy of recognizing wood without apparent texture is much higher than that of measuring wood with apparent texture.