Image Detection Technology For Wood Defects

As for Chinese timber industry, the existence of the resource base is still weak, scattered layout and low level scale that seriously restrict the timber industry. The wood raw material utilization rate is very low and wasteful phenomenon is very serious in Chinese timber industries, less efficient of the wood defects detection is one of the most important reasons that cause this phenomenon. Therefore, in this paper, on the ground of the practical application of wood, we use wood defect images to carry on wood defects detection.In this paper, we firstly analyze the types of wood defects and mainly pay attention to the fish-scale and hard-muscle defects in wood. In the image pre-processing stage, through the gray-scale image processing, the color image are changed into gray-scale image; after histogram equalization, the shortcomings of the image features are highlighting and after image smoothing and denoising, the noise points exists in the image are removed.Secondly, by comparing the results obtained through using histogram threshold method, co-occurrence matrix method and k-means, we find the k-means has a better fish-scale defect detection results. Thirdly, As to dealing with the hard muscle types, after using the Gabor filters, a four-dimensional matrix is got, base on the computation of the four-dimensional matrix, the defects of the hard muscle types are found and the results show that this method is good.Lastly, this paper presents a preliminary design of timber detection system. We mainly give the introduction of software architecture composition the hardware composition. Through analyzing the results of the experiment which has used about 20 images about the scale and hard muscle, the algorithms used in this paper have good reorganized results for the wood.