Research On Recognition Algorithm Of The Wood Surface Defect Image

With the development of wood industry, the manufacture of wood products is increasing significantly. The demand of a consistent high-quality surface wood product introduces automatic inspection that cannot be easily satisfied by traditional manual inspection. Based on the theory of computer vision, a research on defect distinguish of the wood surface is made in the paper.Image preprocess, feature extraction and pattern recognition of wood surface defect images are also studied by means of digital image processing technique and pattern recognition technology based on SVM (Support Vector Machines). Image processing algorithms are studied and improved to orientate and recognize wood surface defect.Image preprocess is the first step for detection, which is vital to the correct extraction of the defection feature. In the fact of a traditional filtering algorithm can substantially damage the edges and details of the image and blur the image's edges and details, a weighted and directional smoothing algorithm is proposed in this paper. Merging several image segmentation method, a improved method of image fusion of multi-resolution analysis based on biorthogonal wavelet transform and a edge detection algorithm based on the fusion technology of wavelet transform and morphological edge detection are proposed in the paper. Thus segmentation result is optimized and laying the root for feature extraction of follow up.The defects are described from two aspects based on image characteristic, the texture features(five gray level co-occurrence matrix parameters) and color features (four color moment parameters) to identify the wood defects. According to the distribution of these parameters, the parameters which have small standard deviation are selected as the input eigenvector of the classifiers. And the features are extracted by the principal components analysis which can reduce the texture dimensions and eliminate the relevance between feature modes and highlight their difference to satisfy the input request of the recognition level. Using Support Vector Machines classifier to identify the defects, the correct rates of pattern recognition achieve better level.The experiment results show it is an effective way to solve the segmentation and identification of wood surface defects by texture features and color features of wood surface defect images according to the digital image processing technology,.