Research On Classification Method Of The Wooden Board Color Based On Computer Vision

With the development of wood industry, the manufacture of wood products is becoming significantly increasing. The demand of a consistently high-quality surface finish in products inspection requirements that cannot be easily met by manual inspection. Color determination exists in each link of the wood producing and application. It is very important to the industrial production. It can achieve defects automatic determination, and have a further signification.This paper discusses the development of a machine vision system for the various defect types on wood surface. The range of surface defects occurred during the paper includes those with both color and texture variations. The machine vision system is to detect undesirable defects that may appear on the surface of rough wood lumber. A neural network is used for a labeling verification step of the high-level recognition system. According to the characteristics of wood itself, inspecting defects of wooden board based on its color is achieved. The research contents in the study are as follows:In this paper, design scheme of inspecting defects system of wood surface color is established. Then the system components and principles are demonstrated, the hardware and software components of the system are introduced.Two segmentation methods of wood defects are studied. One method is application of HSI three independent characters in wood defect segmentation. In this method, change images into HSI model first, and then segment every component separately, fuse the operated components properly last. The other method is comprehensive use edge detection and regional growth. In the HSI color space, the method combines morphology-gradient operator with close operator to extract the crack edges and fill the defect effectively. Then the boundary lines representing potential region distribution are taken as region model, and seeds are got automatically by boundary information to realize segmentation by region growing.The method that BP neural network is used to wood defects recognition is studied. Choose and extract the features of wood defects first, constitute the characteristic matrix, and then train the BP neural network by a part of the defect samples for getting a mature network, realize classification and recognition to the other samples, and have a satisfactory result.The thesis is an important part of the project that aided by the Fund of Heilongjiang Nature Science. In the system, the machine vision system is used to automatically recognized the wood defects by the network, and has been researched the model patterns of defects.