Research on feature extraction of veneer defects based on improved SURF algorithm

Abstract

In this paper, when the feature extraction algorithm extracts the veneer defect features, due to the veneer rotation and a large number of surface textures, the problem of poor feature matching and long time consumption occurs, and an improved SURF algorithm is proposed. The feature matching experiments of the three types of defects such as knots, dead knots, and holes in the veneer were carried out. The experiments verified that the algorithm had good rotation invariance and high efficiency when extracting veneer defect features.