Identification of wood defects based on LBP features

Feature extraction is an important concept in digital image processing, the extraction of the characteristic quantities directly affect the recognition rate of the image. Loeal Binary Pattem is a theory which is simple but useful ealculating method of texture algorithm. In this paper,taking wood defects as research objects,and extracting the LBP texture feature of dedect images. Using BP neural network to identify the defects, and the identification rates reached up to 93%.LBP texture algorithm can improve extraction accuracy of wood defects feature.