Wood Defect Recognition Based on Affinity Propagation Clustering

A new wood defect detection method based on Affinity Propagation clustering was analyzed. By extracting the color characteristics of the wood image, multi-scanning the image, auto-adjusting the sliding window lattice, decreasing the data entry of the sample set after characteristics extracting, dimensions of distance matrix, Jacobi matrix, matching matrix among AP strategy was decreased, wood defect position was automatically identified and marked, and the clustering precision and speed was increased effectively. The experiment results showed that the method can identify wood defect effectively. The average accuracy is about 87.68%, the average recall is around 90.51% and the average identification time is around 2.44s.